Quarry School

College Algebra / Rates of change and behavior of graphs

Rates of change and behavior of graphs

Subject: Mathematics

All 10 lessons, in order.

The unit at a glance

  1. Rate of change and average rate of change

    A rate of change tells you how much one quantity changes for each single unit of another.

  2. On a graph, the average rate of change is a slope

    On a graph, each input and its output make a point: the input tells you how far across, the output tells you how high.

  3. Average rate of change from a formula

    When the function is given as a formula, you build a small table of your own.

  4. When an endpoint is a letter, the answer is an expression

    Sometimes one end of the interval is a letter, like [0, a] or [5, a].

  5. The difference quotient: average rate of change from x to x + h

    The difference quotient has two ordinary actions hidden in its name.

  6. Difference quotient of a quadratic

    A quadratic polynomial uses constants, multiples of the input, and a nonzero multiple of the squared input.

  7. Difference quotient of a fraction

    When the input sits in the bottom of a fraction, the output change becomes one fraction minus another with different bottoms.

  8. Increasing, decreasing, and constant

    Picture a hiking trail seen from the side.

  9. Local maxima and minima

    Think of one hill on a roller coaster.

  10. Toolkit behavior, then absolute maximum and minimum

    Think of a toolkit as the basic shapes you keep in a drawer.

Rules (10)
  • Rate of change and average rate of change

    Average rate of change = ΔyΔx = y2−y1x2−x1 = f(x2)−f(x1)x2−x1, with x2 ≠ x1 and both outputs defined. Use the same endpoint order on top and bottom. Units are output units per input unit.

  • On a graph, the average rate of change is a slope

    For [a, b] = [left end, right end], with a < b and both outputs defined, average rate = f(b)−f(a)b−a. This is the slope of the straight ruler through (a, f(a)) and (b, f(b)).

  • Average rate of change from a formula

    On [a, b] = [left end, right end], with a < b and both outputs defined, average rate = f(b)−f(a)b−a. Evaluate the two endpoint outputs first, then subtract and divide. Preserve all nonzero-bottom and square-root restrictions.

  • When an endpoint is a letter, the answer is an expression

    For a fixed input 5 and an endpoint a, rate = f(a)−f(5)a−5, with a ≠ 5 and both outputs defined. On [5, a], a > 5; if a < 5, use the sorted interval [a, 5]. Factor a polynomial numerator before canceling a matching whole factor.

  • The difference quotient: average rate of change from x to x + h

    Difference quotient = f(x+h)−f(x)h, with h ≠ 0 and both outputs defined. It is the average rate on [x, x + h] if h > 0, and on the sorted interval [x + h, x] if h < 0.

  • Difference quotient of a quadratic

    (x + h)2 = x2 + 2xh + h2. For f(x) = x2, the difference quotient is 2x + h, with h ≠ 0. Multiply any outside coefficient into every expanded term, subtract all of f(x), then factor the whole numerator before canceling h.

  • Difference quotient of a fraction

    For f(x) = 2x, combine the output difference over x(x + h), then divide by h to obtain −2x(x+h). Require h ≠ 0, x ≠ 0, and x + h ≠ 0. Other fraction bottoms need their own product as the common bottom.

  • Increasing, decreasing, and constant

    On an interval, whenever a < b: increasing means f(a) < f(b); decreasing means f(a) > f(b); constant means f(a) = f(b). This course reports intervals of behavior with open endpoints. These are comparisons between inputs, not a direction assigned to one point.

  • Local maxima and minima

    A local maximum at input b has f(b) ≥ f(x) for every x in some open interval around b contained in the domain. A local minimum has f(b) ≤ f(x) there. Its value is f(b); its location is b; its point is (b, f(b)). Equal nearby heights are allowed by these nonstrict inequalities. A strict local extremum uses > or < for every nearby x ≠ b.

  • Toolkit behavior, then absolute maximum and minimum

    The absolute maximum is f(c) when c is in the domain and f(c) ≥ f(x) for every domain input x. The absolute minimum is f(d) when d is in the domain and f(d) ≤ f(x) for every domain input x. The value must be attained, meaning reached at an accepted input.

Recipes (10)
  • Rate of change and average rate of change
    1. Name the input and output, including their units.
    2. For a left-to-right calculation, call the smaller input x1 and the larger input x2; keep each output with its input.
    3. Find Δy = y2 − y1 and Δx = x2 − x1, using the same order.
    4. Check Δx ≠ 0, then divide Δy by Δx.
    5. Attach output units per input unit. With inputs in increasing order, positive means a net rise and negative a net fall.
  • On a graph, the average rate of change is a slope
    1. Locate both interval endpoints on the horizontal axis.
    2. At each input, move vertically to the solid graph and read its height. Count grid steps; estimate if it lies between lines.
    3. Subtract right height minus left height, then right input minus left input.
    4. Divide rise by run and interpret the sign.
  • Average rate of change from a formula
    1. Check that both inputs are allowed: no bottom may be zero, and a real square root needs a nonnegative inside.
    2. Evaluate f(a), placing the whole input in parentheses at every variable position.
    3. Evaluate f(b) the same way.
    4. Subtract f(b) − f(a). If fractions occur, make their bottoms match first.
    5. Divide by b − a. Simplify exact fractions and attach units.
  • When an endpoint is a letter, the answer is an expression
    1. Evaluate the fixed numerical endpoint.
    2. Substitute the endpoint letter to write the other output.
    3. Subtract the outputs and write the matching input difference.
    4. For a polynomial, factor the numerator and verify by multiplying back.
    5. Cancel only a matching factor multiplying the whole numerator, preserving the original restrictions.
    6. If the rate is given, set the simplified rate expression equal to that number, solve, and check the endpoint in the original quotient.
  • The difference quotient: average rate of change from x to x + h
    1. 1. Plug in: replace every x in the formula by the whole input (x + h); multiply out the needed products.
    2. 2. Subtract: write −(f(x)) with parentheses, then let that minus reach every term.
    3. 3. Simplify: combine like terms; for a fully expanded polynomial, the terms independent of h cancel.
    4. 4. Cancel h: when h multiplies the entire numerator, divide that nonzero factor out; keep every original restriction.
  • Difference quotient of a quadratic
    1. 1. Plug in: replace every x by (x + h), multiply out the square, and apply each outside multiplier.
    2. 2. Subtract: keep all of f(x) in parentheses after a minus, then flip every term's sign.
    3. 3. Simplify: combine matching polynomial terms; the terms independent of h cancel.
    4. 4. Cancel h: write every remaining term as h times something, factor h from the whole numerator, and cancel only for h ≠ 0.
  • Difference quotient of a fraction
    1. 1. Plug in: replace every input occurrence, including those below a fraction bar, by the whole new input.
    2. 2. Subtract: use the product of the original bottoms as a common bottom, multiply each top by its missing factor, and keep the second top grouped.
    3. 3. Simplify: distribute the outside minus and combine the numerator terms.
    4. 4. Cancel a valid h factor: multiply by 1h, cancel only a whole nonzero factor, and keep all original restrictions.
  • Increasing, decreasing, and constant
    1. Begin at the left of the whole given graph. Read the horizontal-axis label: the input might be x or t.
    2. Mark turning inputs, flat boundaries, and breaks.
    3. Mark actual domain ends. A finite end dot gives a finite boundary; arrows or stated continuation can justify infinity.
    4. Move right within each separate stretch. Name it increasing, decreasing, or constant.
    5. Write open intervals of inputs. Keep pieces on opposite sides of a break separate.
    6. Give each requested behavior on its own line. Say 'and' between stretches; ∪ joins the listed stretches piece by piece.
  • Local maxima and minima
    1. Scan the given graph for changes from rising to falling or falling to rising.
    2. Rising then falling identifies an isolated local maximum; falling then rising identifies an isolated local minimum.
    3. Also inspect flat stretches and shoulders when applying the formal nonstrict rule. Compare both sides, including any equal heights.
    4. Require a domain neighborhood on both sides. An end of the graph is considered separately for absolute extrema.
    5. Read the horizontal input as the location and the vertical output as the value.
    6. Give value and location in a full sentence; give the ordered pair if the question asks for the point.
  • Toolkit behavior, then absolute maximum and minimum
    1. Review the nine toolkit shapes and their domain restrictions before using the summary table.
    2. Look at the whole stated domain, including included endpoint dots and any arrows or stated continuation.
    3. If outputs rise without bound, there is no absolute maximum; if they fall without bound, there is no absolute minimum.
    4. A height approached but never reached is not an absolute extremum. Check excluded hollow dots and unattained levels.
    5. Compare the remaining reached heights with every domain piece. Find the greatest and least reached outputs and every input producing each.
    6. Answer with the value and location on separate lines. A viewing-window edge does not establish a domain endpoint.
Traps (10)

Lesson 1 of 10

Rate of change and average rate of change

Subject: MathematicsLeans on: Dollars to cents and roundingLeans on: Signed subtractionLeans on: Division by zeroOpen this lesson
Leans on 20Comes back in 0

1 In plain words

A rate of change tells you how much one quantity changes for each single unit of another. A car going 68 miles per hour adds 68 miles for every hour that passes, and a rat population growing by 40 rats per week adds 40 rats each week. The word 'per' means 'for each' and signals a rate. Gas prices in the table below rose in some years and fell in others. If you use only the values at the start and the end of a stretch of years, you get the average rate of change. It works like splitting a restaurant bill evenly: people ate different amounts, but everyone pays the total divided by the number of people.

2 Picture it

input yearoutput price ($ per gallon)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68↓ evaluate: input given, read the output below it
A table of 8 values of price ($ per gallon), for year from 2005 to 2012.

Reminder

Reminder: Dollars to cents and rounding, Signed subtraction, Division by zero
  • Dollars to cents and rounding. Multiply dollars by 100: 0.196 dollars = 19.6 cents. In 0.195714..., the fourth decimal digit is 7, so the three-place rounded value is 0.196.
  • Signed subtraction. Smaller minus bigger is negative: 2.41 − 2.84 = −0.43. Subtracting a negative adds its opposite: 2 − (−1) = 3.
  • Division by zero. A quotient needs a nonzero denominator: 2012 − 2012 = 0, so two identical inputs cannot give this average rate.

3 Why

A total change hides how fast it happened: a $1.37 rise in gas prices over one year is very different from the same rise over seven years. Dividing by the change in input shares the total out equally, one share per unit of input, which is why the result is called an average: if the price had risen by exactly that share every year, it would land on the same final price. The units are output units per input unit because you divided dollars per gallon by years, as miles divided by hours gives miles per hour.

4 The rule

Average rate of change =ΔyΔx=y2−y1x2−x1=f(x2)−f(x1)x2−x1, with x2≠x1 and both outputs defined.

Use the same endpoint order on top and bottom. Units are output units per input unit.

The same idea, five ways

Say it

Say 'delta y over delta x', or 'change in output per unit change in input'.

Write it

The average rate of change compares the net output change with the input change between two distinct endpoints.

In math
  • ΔyΔx
  • ΔfΔx
  • f(x2)−f(x1)x2−x1,x2≠x1
Like

The difference between two odometer readings divided by the time between them.

See it
2004200720102013+7lands on 2012
The move from 2005 to 2012 is Δx = 7 years.

The same idea, other ways

Sharing one total

A restaurant bill divided by the number of people gives the equal share per person. In the same way, a price increase divided by the years gives the equal increase per year that would produce the same final price.

total change ÷ input change

= average change per unit

With the gas numbers

Δ is the Greek capital letter delta, pronounced DEL-tuh and read 'change in'. Δy = 3.68 − 2.31 = 1.37 dollars per gallon. Δx = 2012 − 2005 = 7 years. Dividing gives 137700 dollars per gallon per year. Δf is another name for the output change; it does not change the function into a different function.

input yearoutput price ($ per gallon)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68↓ evaluate: input given, read the output below it
The same picture again: A table of 8 values of price ($ per gallon), for year from 2005 to 2012.
The point name tags

x1, read 'x one' or 'x sub one', names the input of point 1. x2 names the input of point 2. A subscript is a name tag, not a power and not multiplication. y2 and f(x2) name the same output. On this table, x2 = 2012 and y2 = f(2012) = 3.68.

point 1: (2005, 2.31)

point 2: (2012, 3.68)

x2: input of point 2

x2: x multiplied by x

As rise divided by run

On f(x) = x2, going from x = 1 to x = 3 raises the output from 1 to 9. That is rise 8 over run 2, giving 82 = 4. A straight ruler through the two points climbs 4 for each unit right, even though the curve changes steepness.

(1, 1)(3, 9)24246810
The dashed ruler connects two curve points; it is not another piece of the function.
YearAverage price of a gallon of gas
2005$2.31
2006$2.62
2007$2.84
2008$3.30
2009$2.41
2010$2.84
2011$3.58
2012$3.68

Part 1: A negative average, even with a rise in the middle

The average compares the two ends. A price can rise in the middle and still end lower than it began. The negative sign records the net drop, rather than claiming the price fell every year.

  • From 2007 to 2008 the price rises by 3.30 − 2.84 = 0.46 dollars per gallon.
  • From 2007 to 2009 the net change is −0.43 dollars, so the average is negative.
input yearoutput price ($ per gallon)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68↓ evaluate: input given, read the output below it
A table of 8 values of price ($ per gallon), for year from 2005 to 2012, read at year = 2007 and 2009.

Worked example

Average rate of change of a used truck's resale value

A dealership tracks the typical resale value of one model of pickup truck as it ages. Let t be the truck's age in years and V(t) be its resale value in dollars. The table gives these values.

Age t (years): 1, 3, 6, 8
Resale value V(t) ($): 38,400, 31,200, 22,800, 18,600

(a) Find the average rate of change of V on the interval [1, 6]. Give the units and say what the sign means.
(b) Find the average rate of change of V on [6, 8], and compare it with part (a).

input age (years)output resale value ($)138400331200622800818600
A table of 4 values of resale value ($), for age (years) from 1 to 8.
Show the steps and the answer
input age (years)output resale value ($)138400331200622800818600↓ evaluate: input given, read the output below it
Table of the pickup truck's resale value at ages 1, 3, 6 and 8 years.
  1. Input: t, the age in years. Output: V(t), the resale value in dollars.The units of the rate are output units per input unit, so both units are needed first.
  2. For (a), let x1=1 with y1=V(1)=38,400, and x2=6 with y2=V(6)=22,800.Working left to right, the smaller input is x1. Each value stays paired with its own age from the table.
  3. Δy=22,800−38,400=−15,600 and Δx=6−1=5.Both differences use the same order, second endpoint minus first, so the sign is meaningful.
  4. Δx=5≠0, so the average rate of change =−15,6005=−3,120.The average rate of change is ΔyΔx, and it is defined only when Δx is not zero.
  5. The average rate of change on [1, 6] is −3,120 dollars per year. The value fell by an average of $3,120 per year from age 1 to age 6.The inputs increase, so a negative result means the output fell overall.
  6. For (b), let x1=6 with y1=22,800, and x2=8 with y2=18,600. Then Δy=18,600−22,800=−4,200 and Δx=8−6=2.This uses the same left-to-right order on the new interval.
  7. Δx=2≠0, so the average rate of change =−4,2002=−2,100 dollars per year.Divide Δy by Δx and attach dollars per year.
  8. Compare the two rates: −2,100 is closer to zero than −3,120. The value still falls on [6, 8], but by about $1,020 less per year than on [1, 6].Both rates are negative, so both are net falls. The one with the smaller size is the slower fall.
Answer
(a) −3,120 dollars per year on [1, 6]. The resale value fell an average of $3,120 each year. (b) −2,100 dollars per year on [6, 8]. The value is still falling, but more slowly in those later years.
Double-check Start at V(1) = 38,400 and apply the rate over 5 years: 38,400 + 5(−3,120) = 38,400 − 15,600 = 22,800 = V(6). For (b): 22,800 + 2(−2,100) = 18,600 = V(8). Reversing the order also agrees: 38,400−22,8001−6 = 15,600−5 = −3,120.

Work to write

  1. Input t in years, output V(t) in dollars
  2. (a)x1=1,y1=38,400;x2=6,y2=22,800
  3. Δy=22,800−38,400=−15,600;Δx=6−1=5
  4. −15,6005 = −3,120 dollars per year
  5. Negative: the value fell an average of $3,120 per year on [1, 6]
  6. (b) 18,600−22,8008−6 = −4,2002 = −2,100 dollars per year
  7. The value falls more slowly on [6, 8] than on [1, 6]

Part 2: Rats per week

Count rats added for each week that passes.

  • The source example is growth of 40 rats per week.
  • Output: population change in rats. Input: elapsed weeks.

80 rats in 2 weeks

40 rats per week

Worked example

A population rate

This asks how many rats are added for each week. A population grows by 80 rats over 2 weeks. Find its average rate.

Show the steps and the answer

80 rats in 2 weeks

40 rats per week

  1. Change in output is 80 rats; change in input is 2 weeks.The problem gives both changes, so there are no endpoint outputs to subtract.
  2. 802=40 rats per week.Dividing the total addition into two equal weekly shares gives the amount per week.
Answer
40 rats per week.
Double-check 2 · 40 = 80 rats.

Part 3: Miles per hour

A travel rate shares a distance over elapsed hours.

  • The source example is 68 miles per hour.
  • Output: miles traveled. Input: hours.

136 miles in 2 hours

68 miles per hour

Worked example

A travel rate

This asks how many miles are covered per hour. A car covers 136 miles in 2 hours. Find its average rate.

Show the steps and the answer

136 miles in 2 hours

68 miles per hour

  1. 1362=68 miles per hour.Miles are the output change and hours are the input change, so divide miles by hours.
Answer
68 miles per hour.
Double-check 2 · 68 = 136 miles.

Part 4: Miles per gallon

You can compare distance with fuel instead of time.

  • The source example is 27 miles per gallon.
  • Output: miles traveled. Input: gallons used.

54 miles using 2 gallons

27 miles per gallon

Worked example

A fuel-use rate

This asks how many miles are covered for each gallon, rather than for each hour. A car covers 54 miles using 2 gallons. Find its rate.

Show the steps and the answer

54 miles using 2 gallons

27 miles per gallon

  1. 542=27 miles per gallon.Distance is the output and fuel used is the input, so time does not belong in this denominator.
Answer
27 miles per gallon.
Double-check 2 gallons · 27 miles per gallon = 54 miles.

Part 5: Amperes per volt

You can compare two measured electrical quantities without learning a physics formula.

  • The source example is current increasing by 0.125 amperes for each extra volt.
  • Output: current change in amperes. Input: voltage change in volts.

0.25 amperes per 2 volts

0.125 amperes per volt

Worked example

An electrical rate

This asks how much the current increases for each extra volt. The current increases by 0.25 amperes when voltage increases by 2 volts. Find the rate.

Show the steps and the answer

0.25 amperes per 2 volts

0.125 amperes per volt

  1. 0.252=0.125 amperes per volt.An ampere measures electrical current and a volt measures voltage. Dividing the current change by the voltage change gives the requested comparison.
Answer
0.125 amperes per volt.
Double-check 2 · 0.125 = 0.25 amperes.

Part 6: Dollars per quarter

The output can fall while the input time increases.

  • The source example is a college account decreasing by $4,000 per quarter.
  • A quarter of a year is three months.

−$8,000 in 2 quarters

−$4,000 per quarter

Worked example

A falling account balance

This asks how much money changes per quarter. An account loses $8,000 over 2 quarters. Find the average rate.

Show the steps and the answer

−$8,000 in 2 quarters

−$4,000 per quarter

  1. Change in money =−8,000 dollars per gallon.A loss is a negative change in the output balance.
  2. −8,0002=−4,000 dollars per quarter.Divide the loss by the positive time span. Here a quarter means three months.
Answer
−$4,000 per quarter.
Double-check 2 · (−4,000) = −8,000 dollars per gallon.

5 How

  1. Name the input and output, including their units.
  2. For a left-to-right calculation, call the smaller input x1 and the larger input x2; keep each output with its input.
  3. Find Δy = y2 − y1 and Δx = x2 − x1, using the same order.
  4. Check Δx ≠ 0, then divide Δy by Δx.
  5. Attach output units per input unit. With inputs in increasing order, positive means a net rise and negative a net fall.

Find an average rate from the information given

1
Is the information a table?
YesRead the output directly beneath each chosen input.
NoLook for a graph or formula.
↓
2
Is it a graph?
YesRead the two heights at the chosen inputs; lesson 1 shows how.
NoIf neither graph nor table was given, evaluate the formula; otherwise keep the outputs already read.
↓
3
Are the two inputs equal?
YesStop: the denominator is 0, so this average rate is undefined.
NoSubtract in matching order and divide.
↓
4
Is an endpoint a letter or is the input step h?
YesKeep the endpoint symbolic; lessons 3 and 4 show those forms.
NoCompute the numerical rate.
  1. Find two distinct input values and their outputs.
  2. Read a table column, read a graph height, or evaluate a formula, depending on what is given.
  3. Subtract the two outputs and the two inputs in matching order.
  4. Divide and state units; keep exact fractions until a rounded answer is requested.

6 Worked example

Worked example

Average rate of change of a regional gas price, 2014 to 2021

A regional energy office records the average price of one gallon of regular gasoline each year. Let P(t) be the price in dollars per gallon in year t. The table gives these values.

Year t: 2014, 2016, 2019, 2021
Price P(t) ($/gal): 3.36, 2.14, 2.60, 3.01

(a) Find the average rate of change of the gas price from 2014 to 2021. Give units and say what the sign means.
(b) Find the average rate of change from 2016 to 2021. Explain why the signs in (a) and (b) differ, even though both intervals end in 2021.

input yearoutput price ($ per gallon)20143.3620162.1420192.620213.01
A table of 4 values of price ($ per gallon), for year from 2014 to 2021.
Show the steps and the answer
input yearoutput price ($ per gallon)20143.3620162.1420192.620213.01↓ evaluate: input given, read the output below it
Average regional gas price P(t) in dollars per gallon for the years 2014, 2016, 2019 and 2021.
  1. Input: t, the year, in years. Output: P(t), the price, in dollars per gallon ($/gal).The units of the answer are output units per input unit, so both units are needed first.
  2. (a) Let x1=2014 with y1=P(2014)=3.36, and x2=2021 with y2=P(2021)=3.01.For a left-to-right calculation the smaller input is x1. Each price stays paired with its own year.
  3. Δy=3.01−3.36=−0.35 and Δx=2021−2014=7.Both differences use the same order, the later value minus the earlier value.
  4. Δx=7≠0, so the average rate of change is −0.357=−0.05.Average rate of change = ΔyΔx. Division is allowed because Δx is not zero.
  5. (a) The rate is −0.05 dollars per gallon per year. The price fell by a net 5 cents per gallon per year from 2014 to 2021.The inputs are in increasing order, so a negative rate means a net fall. The price did not fall every year: it rose from 2016 to 2021. The rate describes only the overall change between the endpoints.
  6. (b) Let x1=2016 with y1=2.14, and x2=2021 with y2=3.01. Then Δy=3.01−2.14=0.87 and Δx=2021−2016=5.This is the same left-to-right method on the new interval, keeping each price with its year.
  7. Δx=5≠0, so the rate is 0.875=0.174 dollars per gallon per year, a net rise of about 17.4 cents per gallon per year.The rate is positive with the inputs in increasing order, so the price rose overall on [2016, 2021].
  8. The signs differ because the starting prices differ. The 2014 price of $3.36 is above the 2021 price of $3.01. The 2016 price of $2.14 is below it.An average rate of change depends only on the two endpoint values, not on the values in between.
Answer
(a) 3.01−3.362021−2014 = −0.357 = −0.05 dollars per gallon per year, a net fall of 5 cents per gallon per year from 2014 to 2021. (b) 3.01−2.142021−2016 = 0.875 = 0.174 dollars per gallon per year, a net rise of about 17.4 cents per gallon per year from 2016 to 2021.
Double-check Start at the first endpoint price and add the rate times the number of years. (a) 3.36 + 7(−0.05) = 3.36 − 0.35 = 3.01 = P(2021) ✓. (b) 2.14 + 5(0.174) = 2.14 + 0.87 = 3.01 = P(2021) ✓. Reversing the endpoint order in (a) gives 3.36−3.012014−2021 = 0.35−7 = −0.05, the same result, as it should be.

Work to write

  1. Input t in years; output P(t) in dollars per gallon
  2. (a)x1=2014,y1=3.36;x2=2021,y2=3.01
  3. Δy=3.01−3.36=−0.35,Δx=2021−2014=7
  4. −0.357 = −0.05 dollars per gallon per year
  5. Negative: net fall of 5 cents per gallon per year from 2014 to 2021
  6. (b) 3.01−2.142021−2016 = 0.875 = 0.174 dollars per gallon per year, a net rise
  7. The signs differ because P(2014) > P(2021) but P(2016) < P(2021)

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

A population rate

This asks how many rats are added for each week. A population grows by 80 rats over 2 weeks. Find its average rate.

Show the steps and the answer

80 rats in 2 weeks

40 rats per week

  1. Change in output is 80 rats; change in input is 2 weeks.The problem gives both changes, so there are no endpoint outputs to subtract.
  2. 802=40 rats per week.Dividing the total addition into two equal weekly shares gives the amount per week.
Answer
40 rats per week.
Double-check 2 · 40 = 80 rats.

Rung 2 of 5

Average yearly change in gas prices from a table, 2016 to 2023

A state consumer office reports the average price of one gallon of regular gasoline in selected years. Let P(t) be the price in dollars per gallon in year t. The table gives these values.

Year t: 2016, 2018, 2020, 2023
Price P(t) ($/gal): 2.27, 2.85, 2.24, 3.53

(a) Find the average rate of change of P from 2016 to 2023. Give units and say what the sign means.
(b) Find the average rate of change of P from 2018 to 2020. Give units and say what the sign means.

input yearoutput price ($ per gallon)20162.2720182.8520202.2420233.53
A table of 4 values of price ($ per gallon), for year from 2016 to 2023.
Show the steps and the answer
input yearoutput price ($ per gallon)20162.2720182.8520202.2420233.53↓ evaluate: input given, read the output below it
Average price of regular gasoline in 2016, 2018, 2020 and 2023.
  1. Input: t, the year. Output: P(t), the price in dollars per gallon. The rate will be in dollars per gallon per year.The units of an average rate of change are output units per input unit, so both quantities and their units must be named first.
  2. (a) Let x1=2016 with P(2016)=2.27, and x2=2023 with P(2023)=3.53.For a left-to-right calculation the smaller input is x1. Each price stays with its own year.
  3. Δy=P(2023)−P(2016)=3.53−2.27=1.26 and Δx=2023−2016=7.Both differences use the same order, later minus earlier, so the quotient has the correct sign.
  4. Δx=7≠0, so the average rate of change =1.267=0.18.Average rate of change = ΔyΔx, which is defined only when the inputs differ.
  5. The rate is 0.18 dollars per gallon per year. It is positive, so the price rose overall: on average it went up 18 cents per gallon each year from 2016 to 2023.The inputs are in increasing order, so a positive rate means a net rise in the output.
  6. (b) Let x1=2018 with P(2018)=2.85, and x2=2020 with P(2020)=2.24. Then Δy=2.24−2.85=−0.61 and Δx=2020−2018=2.These are the same steps with the new endpoints, again taking later minus earlier on both top and bottom.
  7. Δx=2≠0, so the average rate of change =−0.612=−0.305 dollars per gallon per year.The difference in price is divided by the difference in years.
  8. The rate is negative, so the price fell overall from 2018 to 2020: on average it dropped about 30.5 cents per gallon each year.With the inputs in increasing order, a negative rate means a net fall in the output.
Answer
(a) 0.18 dollars per gallon per year. The price rose by 18 cents per gallon per year on average from 2016 to 2023. (b) −0.305 dollars per gallon per year. The price fell by about 30.5 cents per gallon per year on average from 2018 to 2020.
Double-check Start at each first price and add the rate times the number of years. (a) 2.27 + 7(0.18) = 2.27 + 1.26 = 3.53, which matches P(2023). (b) 2.85 + 2(−0.305) = 2.85 − 0.61 = 2.24, which matches P(2020). The signs also agree with the table: the price went up from 2016 to 2023 and down from 2018 to 2020.

Work to write

  1. Input t = year; output P(t) in dollars per gallon
  2. (a)x1=2016,P=2.27;x2=2023,P=3.53
  3. 3.53−2.272023−2016=1.267=0.18
  4. 0.18 dollars per gallon per year; positive, so the price rose on average
  5. (b)2.24−2.852020−2018=−0.612=−0.305
  6. −0.305 dollars per gallon per year; negative, so the price fell on average

Rung 3 of 5

Average rate of change of a fuel price from 2015 to 2019

A small delivery company keeps a fuel log of the average price it paid for one gallon of regular gasoline each year. Let P(t) be the price in dollars per gallon in year t. The table gives these values.

Year t: 2015, 2017, 2018, 2019
Price P(t) ($/gal): 2.62, 2.48, 2.71, 3.02

Find the average rate of change of P over the four years from 2015 to 2019. Give units and say what the sign means.

input yearoutput price ($ per gallon)20152.6220172.4820182.7120193.02
A table of 4 values of price ($ per gallon), for year from 2015 to 2019.
Show the steps and the answer
input yearoutput price ($ per gallon)20152.6220172.4820182.7120193.02↓ evaluate: input given, read the output below it
The company's fuel-log prices for 2015, 2017, 2018 and 2019.
  1. Input: t, the year (years). Output: P(t), the price of gasoline (dollars per gallon).The units of the rate of change are output units per input unit, so both must be named first.
  2. Let x1=2015 with y1=P(2015)=2.62. Let x2=2019 with y2=P(2019)=3.02.For a left-to-right calculation the smaller input is x1. Each output stays with its own input. The 2017 and 2018 values are not endpoints of the interval, so they are not used.
  3. Δy=y2−y1=3.02−2.62=0.40 dollars per gallon. Δx=x2−x1=2019−2015=4 years.The same endpoint order (2019 first, then 2015) is used on top and bottom, so the signs are consistent.
  4. Δx=4≠0, so the average rate of change =ΔyΔx=0.404=0.10.The rule ΔyΔx is defined only when the inputs differ. Here they do.
  5. Average rate of change =0.10 dollars per gallon per year. It is positive.The inputs are in increasing order, so a positive rate means a net rise in price from 2015 to 2019. The dip in 2017 does not change the net result, because only the endpoints enter the calculation.
Answer
The average rate of change of P from 2015 to 2019 is 0.10 dollars per gallon per year. On average the price rose by 10 cents per gallon each year. The positive sign means a net increase over the four years.
Double-check Start at 2.62 and add 0.10 for each of the 4 years: 2.62 + 0.10 · 4 = 2.62 + 0.40 = 3.02, which matches P(2019).

Work to write

  1. Input t in years; output P(t) in dollars per gallon
  2. x1=2015,y1=2.62;x2=2019,y2=3.02
  3. Δy=3.02−2.62=0.40
  4. Δx=2019−2015=4
  5. ΔyΔx=0.404=0.10
  6. 0.10 dollars per gallon per year
  7. Positive: the price rose overall from 2015 to 2019

Rung 4 of 5

How fast a game console's price falls

A retailer tracks the price of a new handheld game console. Let t be the number of months since launch and P(t) be the price in dollars. The table gives these values.

Months since launch t: 2, 6, 14, 20
Price P(t) ($): 449, 425, 377, 335

(a) Find the average rate of change of P on the interval [2, 14]. Give units and say what the sign means.
(b) Find the average rate of change of P on [14, 20]. Over which interval did the price fall faster on average?

input months since launchoutput price ($)244964251437720335
months since launch: 2, 6, 14, 20; price ($): 449, 425, 377, 335.
Show the steps and the answer
input months since launchoutput price ($)244964251437720335↓ evaluate: input given, read the output below it
Table of the console's price at 2, 6, 14 and 20 months after launch.
  1. Input: t, the time since launch in months. Output: P(t), the price in dollars.The units of the rate are output units per input unit. Here that is dollars per month.
  2. For (a), let x1=2 and y1=P(2)=449. Let x2=14 and y2=P(14)=377.Going left to right, the smaller input is x1. Each price stays paired with its own month.
  3. Δy=377−449=−72 and Δx=14−2=12.Both differences subtract in the same order: second endpoint minus first.
  4. Δx=12≠0, so the average rate of change =−7212=−6.The average rate of change is ΔyΔx, and it is defined only when Δx ≠ 0.
  5. The rate on [2, 14] is −6 dollars per month. The sign is negative, so the price had a net fall. On average it dropped $6 each month.The inputs are in increasing order, so a negative rate means a net decrease.
  6. For (b), let x1=14,y1=377,x2=20 and y2=335. Then Δy=335−377=−42 and Δx=20−14=6.These are the same steps on the new interval, keeping the same endpoint order on top and bottom.
  7. Δx=6≠0, so the rate =−426=−7 dollars per month.Divide Δy by Δx and attach dollars per month.
  8. Compare: |−7|>|−6|, so the price fell faster on average over [14, 20].Both rates are negative. The rate with the larger size shows the steeper average drop.
Answer
(a) −6 dollars per month: on average the price fell $6 per month from month 2 to month 14. (b) −7 dollars per month. The price fell faster on average over [14, 20].
Double-check Start at the first price and add the rate times the length of the interval. Each result should match the table.
(a) 449 + (−6)(12) = 449 − 72 = 377 = P(14). ✓
(b) 377 + (−7)(6) = 377 − 42 = 335 = P(20). ✓

Work to write

  1. Input t in months, output P(t) in dollars
  2. [2,14]:x1=2,y1=449;x2=14,y2=377
  3. Δy=377−449=−72,Δx=14−2=12
  4. Average rate of change = −7212 = −6 dollars per month
  5. Negative, so the price had a net fall of $6 per month on average
  6. [14, 20]: 335−37720−14 = −426 = −7 dollars per month
  7. The price fell faster on average over [14, 20]

Rung 5 of 5

A rate from clock times across an hour

A faucet adds 30 gallons from 7:52 p.m. to 8:04 p.m. Find the average gallons added per minute; first find how many minutes actually passed.

Show the steps and the answer
04812+8+4lands on 12
The walk measures elapsed minutes: 8 minutes to the hour plus 4 afterward makes a 12-minute run.
  1. From 7:52 p.m. to 8:00 p.m. takes 60−52=8 minutes.An hour contains 60 minutes, so count forward to the next hour.
  2. From 8:00 p.m. to 8:04 p.m. takes 4 more minutes. Total elapsed time: 8+4=12 minutes.Both spans belong to the same continuous elapsed-time interval; add them rather than subtracting clock digits as decimals.
  3. Average rate =3012=52=2.5 gallons per minute.Gallons are the output change and elapsed minutes the input change; divide top and bottom by 6 to reduce.
Answer
2.5 gallons per minute.
Double-check 12 · 2.5 = 30 gallons, recovering the amount added. Count to the next hour, then add the minutes after it.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: Δx means Δ multiplied by x, and x2 means x squared.
Δx is one name for the input change. The small lower 2 identifies the second point; the raised 2 in x2 means x · x.
✓ Instead: For the gas example, Δx = 2012 − 2005 = 7; x2 = 2012.
✗ Not this: A negative average means the price fell at every intermediate year.
The 2007-to-2009 average is −0.215, but the price rises 0.46 dollars from 2007 to 2008.
✓ Instead: An average compares only the endpoint outputs; it can hide the path.
✗ Not this: Use ΔxΔy = 71.37 ≈ 5.1 as the price change per year.
That division has units years per dollar, so it answers a different question.
✓ Instead: Use ΔyΔx = 1.377 dollars per gallon per year.
✗ Not this: A rate must be measured per unit of time.
Miles per gallon and amperes per volt use gallons and volts as inputs.
✓ Instead: Always divide output units by input units, whatever the input measures.

Tips and tricks

  • Remember 'output over input' and 'same trip, same order'. Write the units as dollarsyear and read the bar as 'per'.
  • A rate is also negative if output rises while input falls: reversing both endpoint orders changes both signs and leaves the ratio unchanged.
  • A subscript names a point. A superscript tells you a power.
  • If both outputs match and the inputs differ, the average rate is 0; the function may still move up and down between those endpoints.

7 The trap

Subtracting in different orders on top and bottom flips the sign. Point 2 minus point 1 must be used for both output and input changes.

Lesson 2 of 10

On a graph, the average rate of change is a slope

Subject: MathematicsLeans on: Signed divisionLeans on: Subtracting a negativeLeans on: SlopeOpen this lesson
Leans on 0Comes back in 0

1 In plain words

On a graph, each input and its output make a point: the input tells you how far across, the output tells you how high. To find the average rate of change between two inputs, locate the points on the graph at those two inputs and connect them with a straight ruler line. The average rate of change is the slope of that line. Picture a hiking trail between two signposts: the trail may wiggle up and down, but the ruler line ignores the wiggles and measures only how much height you gained overall for each step east.

2 Picture it

(−1, 4)(2, 1)top 5−22−4−2246(−1, 4)(2, 1)top 5−2−1123−4−2246
y = −x2 + 5, solid, beside y = −x + 3, dashed.

Reminder

Reminder: Signed division, Subtracting a negative, Slope
  • Signed division. Same signs divide to a positive; different signs divide to a negative. −3−3 = 1 and 3−3 = −1.
  • Subtracting a negative. 2 − (−1) = 2 + 1 = 3. Parentheses keep the negative input together.
  • Slope. Rise over run compares two points: a rise of 9 over a run of 3 gives 93 = 3.

3 Why

The average rate of change formula f(x2)−f(x1)x2−x1 is the slope formula y2−y1x2−x1 in disguise, because on a graph the output f(x) is the height y. So a connecting line rising to the right has a positive slope, a falling one a negative slope, and a flat one zero. What happens between the two points never enters the formula, which is why only the endpoints matter. A curved graph can rise near one endpoint yet have a negative average over the whole interval.

4 The rule

For [a,b]= [left end, right end], with a < b and both outputs defined, average rate =f(b)−f(a)b−a.

This is the slope of the straight ruler through (a, f(a)) and (b, f(b)).

The same idea, five ways

Say it

Say 'the slope of the line joining the two graph points'.

Write it

The average rate compares the ending height with the starting height, per horizontal unit.

In math
  • f(b)−f(a)b−a,a<b
  • riserun
  • Graph words: connect the two endpoint points with a straight ruler.
Like

A ruler laid between two trail signposts measures the overall climb per step east.

See it
(−1, 4)(2, 1)top 5−22−4−2246
The same picture again: y = −x2 + 5, solid, beside y = −x + 3, dashed.

The same idea, other ways

Follow the graph to a height

To read g(−1), begin at −1 on the horizontal axis, go vertically to the solid curve, and read height 4. The point (−1, 4) and the notation g(−1) = 4 say the same thing.

(−1, 4)(2, 1)top 5−22−4−2246
The same picture a third time: y = −x2 + 5, solid, beside y = −x + 3, dashed.
Connect only the two ends

A hiker can climb to a high point in the middle and finish below the starting point. On this curve, the endpoints are at heights 4 and 1. The ruler drops 3 over a horizontal move of 3, so the average rate is −1.

(−1, 4)(2, 1)top 5−22−4−2246
The same picture, shown 4 times now: y = −x2 + 5, solid, beside y = −x + 3, dashed.
Explain it in a sentence

For a straight line the average rate stays the same on every interval because the line has the same steepness throughout. A curve can give different averages: x2 gives 4 on [1, 3] and 8 on [3, 5].

x2 on [1, 3]: 9−12 = 4

x2 on [3, 5]: 25−92 = 8

Part 1: Reading a value off a graph

A graph is another way to store input-output pairs. Go from the input axis to the curve before reading the output axis. Count grid squares using the printed scale; a square can represent more than one unit. A height between grid lines is an estimate.

  • g(t) is the output at input t, rather than g multiplied by t.
  • A point (t, y) on the graph means g(t) = y.
(−1, 4)(2, 1)top 5−22−4−2246
The same picture, shown 5 times now: y = −x2 + 5, solid, beside y = −x + 3, dashed.

Worked example

Reading two heights on a downward parabola, then finding the slope

The solid graph of f(x) = −(x + 1)2 + 4 is drawn on a grid where each square is 1 unit in both directions. Its highest point is (−1, 4). The graph passes exactly through the grid points (−2, 3) and (2, −5). Read the graph to find the average rate of change of f over [−2, 2]. Say what the sign means.

(−2, 3)(2, −5)(−1, 4)−4−224−44
Graph of f(x) = −(x + 1)2 + 4 with its highest point at (−1, 4) and the interval endpoints (−2, 3) and…
Show the steps and the answer
  1. Locate x=−2 (left end) and x=2 (right end) on the horizontal axis.The interval [a, b] = [−2, 2] has a = −2 < b = 2. These two inputs are where the heights must be read.
  2. From x=−2, move vertically to the solid curve. It meets the curve 3 squares above the x-axis, so f(−2)=3.The height of the graph at an input is the output. The point (−2, 3) lies exactly on a grid point, so no estimate is needed.
  3. From x=2, move vertically to the solid curve. It meets the curve 5 squares below the x-axis, so f(2)=−5.The point (2, −5) lies exactly on a grid point. It is below the axis, so the height is negative.
  4. Rise =f(2)−f(−2)=−5−3=−8. Run =2−(−2)=4.Subtract right minus left for both outputs and inputs, in the same order.
  5. Average rate =−84=−2.The average rate of change is rise divided by run. This is the slope of the ruler through (−2, 3) and (2, −5).
  6. Interpret: the sign is negative, so over [−2,2] the function falls overall. On average, f drops 2 units for each 1 unit increase in x.A negative slope means the right endpoint is lower than the left endpoint. The graph rises briefly to its top at x = −1, but the overall change across the interval is a decrease.
Answer
The average rate of change of f over [−2, 2] is −2. Overall f decreases, on average 2 units per 1 unit of x.
Double-check Use the formula. f(−2) = −(−2 + 1)2 + 4 = −1 + 4 = 3, and f(2) = −(2 + 1)2 + 4 = −9 + 4 = −5. These match the graph readings. Then −5−32−(−2) = −84 = −2. Also, the ruler from (−2, 3) drops 8 units over 4 units to the right, which is a slope of −2.

Work to write

  1. f(−2) = 3 (read from graph)
  2. f(2) = −5 (read from graph)
  3. Average rate = f(2)−f(−2)2−(−2) = −5−34
  4. =−84=−2
  5. Negative: f decreases on average by 2 units per unit of x on [−2, 2]

Part 2: Average speed from Anna's table

The table stores distance from home, so the starting 10 miles must be subtracted. In this outward-driving example, the change is also the distance driven during the chosen hours. If a trip turns back, distance from home alone does not measure all miles driven.

  • Average speed over [0, 6] is 47 miles per hour.
  • Over [2, 3], it is 63 miles per hour.
  • The chosen interval can change a travel average.
input t (hours)output D(t) (miles)01015529031534214524062927300↓ evaluate: input given, read the output below it
The highlighted endpoint columns show distance from home at the start and after six hours.

Worked example

Average speed of a cyclist between two times in a distance table

A cyclist's total distance from the trailhead is recorded at several times.

Time t (minutes): 10, 25, 40, 55
Distance d(t) (km): 3.2, 8.0, 12.5, 16.1

Find the cyclist's average speed on the interval [10, 40]. Give it in km per minute and in km per hour, and say what the sign means.

Show the steps and the answer
input time (minutes)output distance (km)103.22584012.55516.1↓ evaluate: input given, read the output below it
Table of the cyclist's distance from the trailhead at 10, 25, 40 and 55 minutes.
  1. Find the interval endpoints in the time row. The left end is a=10 and the right end is b=40. Both values appear in the table, and 10 < 40.The rule needs a left input and a right input with a < b, and both outputs must be known.
  2. Read the distance under each time: d(10)=3.2 km and d(40)=12.5 km. The points are (10, 3.2) and (40, 12.5).These are the heights of the two points that the straight ruler would join on a distance–time graph.
  3. Find the rise as the right output minus the left output: 12.5−3.2=9.3 km. Find the run as the right input minus the left input: 40−10=30 minutes.The numerator of d(b)−d(a)b−a is the change in distance and the denominator is the change in time. Both are taken right minus left.
  4. Divide the rise by the run: 9.330=0.31 km per minute.Average rate of change is rise over run, which is the slope of the line through (10, 3.2) and (40, 12.5). The units are km per minute.
  5. Convert to km per hour: 0.31×60=18.6 km/h.There are 60 minutes in an hour, so the number of km per hour is 60 times the number of km per minute.
  6. Interpret the sign. The rate is positive, so the cyclist's distance from the trailhead increased over the interval.A positive slope means the output rises as the input increases.
Answer
Average speed on [10, 40] = 12.5−3.240−10 = 9.330 = 0.31 km/min, which is 18.6 km/h. The rate is positive, so the cyclist moved farther from the trailhead.
Double-check Start at 3.2 km and ride at 0.31 km/min for 30 minutes. That adds 0.31 × 30 = 9.3 km, and 3.2 + 9.3 = 12.5 km, which matches d(40). Also, 18.6 km/h is a reasonable cycling speed.

Work to write

  1. a = 10, b = 40
  2. d(10) = 3.2, d(40) = 12.5
  3. average speed = d(40)−d(10)40−10 = 12.5−3.240−10
  4. = 9.330 = 0.31 km/min
  5. 0.31 × 60 = 18.6 km/h
  6. Positive rate: the distance from the trailhead increased.

Part 3: A line compared with a curve

A straight line rises by the same amount for each equal step right. It therefore gives the same average rate on every interval. A curve can have different steepness in different places, so its average may depend on the interval.

  • For f(x) = 2x + 5 the average rate is 2 on every pair of distinct inputs.
  • For f(x) = x2, the averages on [1, 3] and [3, 5] are 4 and 8.
(0, 5)(1, 7)(2, 9)(7, 19)run 1rise 224626101418
Every unit right adds 2 to the height of this straight line.

Worked example

Average rate of change of a straight-line graph

The solid graph of y = f(x) is a straight line. Reading the grid, it passes through the points (−2, 7), (2, 1) and (4, −2). Find the average rate of change of f on the interval [−2, 4], and then on [2, 4]. Interpret the sign, and explain why the two answers agree.

(−2, 7)(2, 1)(4, −2)−4−2246−4−22468
The straight line y = f(x) through (−2, 7), (2, 1) and (4, −2).
Show the steps and the answer
  1. Locate the endpoints of the first interval, x=−2 and x=4, on the horizontal axis.The interval [a, b] = [−2, 4] has left end a = −2 and right end b = 4, with a < b.
  2. Read the heights above those inputs. At x=−2 the graph is at height 7, so f(−2)=7. At x=4 it is at height −2, so f(4)=−2.The average rate uses the outputs at the two ends of the interval. Both points sit on grid intersections, so no estimating is needed.
  3. Find the rise and the run. Rise =f(4)−f(−2)=−2−7=−9. Run =4−(−2)=6.The rule subtracts right minus left in both places, so the order matches.
  4. Divide: average rate =−96=−32=−1.5.Average rate = f(b)−f(a)b−a is the slope of the segment joining (−2, 7) and (4, −2).
  5. Repeat for [2, 4]. f(2)=1 and f(4)=−2. Rise =−2−1=−3. Run =4−2=2. Average rate =−32=−1.5.This is the same method on a different interval of the same line.
  6. Interpret the result. The rate is negative, so f falls 1.5 units for every 1 unit increase in x. Both intervals give −1.5.A straight line has one constant slope. A ruler through any two of its points lies along the line itself, so every interval gives the same average rate.
Answer
The average rate of change is −32 = −1.5 on [−2, 4] and also −1.5 on [2, 4]. The function decreases by 1.5 units per unit of x. The answers agree because the graph is a straight line with slope −1.5.
Double-check Use the third pair of points, (−2, 7) and (2, 1): 1−72−(−2) = −64 = −1.5. This matches. The equation y = −1.5x + 4 also checks: at x = −2, y = 3 + 4 = 7; at x = 2, y = −3 + 4 = 1; at x = 4, y = −6 + 4 = −2.

Work to write

  1. f(−2)=7,f(4)=−2
  2. average rate on [−2, 4] = f(4)−f(−2)4−(−2) = −2−76 = −96 = −1.5
  3. f(2) = 1, so average rate on [2, 4] = −2−14−2 = −32 = −1.5
  4. negative: f decreases 1.5 units per unit of x
  5. same answer on both intervals because a straight line has constant slope

5 How

  1. Locate both interval endpoints on the horizontal axis.
  2. At each input, move vertically to the solid graph and read its height. Count grid steps; estimate if it lies between lines.
  3. Subtract right height minus left height, then right input minus left input.
  4. Divide rise by run and interpret the sign.

Read an average rate from a graph

1
Is an endpoint output missing, shown only by a hollow dot?
YesThat output is not attained there; this endpoint quotient is undefined unless another filled point gives the output.
NoRead its solid-graph height.
↓
2
Does a height fall between grid lines?
YesEstimate the height and mark the final rate approximate.
NoUse the exact marked height.
↓
3
Does the curve change direction in the middle?
YesKeep only the two endpoint heights for the average.
NoUse the same endpoint method.
  1. Read two curve heights before doing arithmetic.
  2. Write the paired points or function values so the endpoint order is visible.
  3. Compute vertical change divided by horizontal change.
  4. Describe the average net rise or fall per unit.

6 Worked example

Worked example

Average rate of change of a parabola from its graph on [−1, 3]

The graph of f(x) = x2 − 1 is drawn on a grid where each square is 1 unit. Read the graph to find the average rate of change of f over [−1, 3]. Say what the sign means.

(−1, 0)(3, 8)−224−22468
Graph of f(x) = x2 − 1 with the interval endpoints (−1, 0) and (3, 8) marked.
Show the steps and the answer
  1. Locate the endpoints on the horizontal axis: a=−1 on the left and b=3 on the right.The interval [a, b] runs from the left end to the right end, so a = −1 and b = 3 with a < b.
  2. From x=−1, move vertically to the graph. It meets the curve on the x-axis, so f(−1)=0. The point is (−1,0).The output is the height of the graph above the input. Here the height lands exactly on a grid line, so no estimate is needed.
  3. From x=3, move vertically to the graph. Count up 8 grid steps, so f(3)=8. The point is (3, 8).Counting whole grid squares gives the exact height. Check with the formula: 32 − 1 = 8.
  4. Rise =f(3)−f(−1)=8−0=8. Run =3−(−1)=4.Use right minus left in both the numerator and the denominator. Subtracting −1 is the same as adding 1.
  5. Average rate =84=2. The sign is positive.The rate is rise divided by run. A positive slope means the ruler through the two points goes up from left to right.
Answer
Average rate of change = f(3)−f(−1)3−(−1) = 84 = 2. On average, f increases by 2 units for each 1-unit increase in x over [−1, 3].
Double-check The line through (−1, 0) with slope 2 is y = 2x + 2. At x = 3 it gives 2(3) + 2 = 8, which is f(3). So the ruler joins both endpoints and its slope is 2. The graph also dips below 0 near x = 0, which shows that the average rate describes only the ruler between the two endpoints, not the shape of the curve between them.

Work to write

  1. a=−1,b=3
  2. f(−1)=0,f(3)=8
  3. rise = 8 − 0 = 8, run = 3 − (−1) = 4
  4. average rate = 84 = 2
  5. Positive: f rises 2 units per unit of x on average over [−1, 3]

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 4

Reading two heights on a shifted parabola to find a negative average rate

The solid graph of f(x) = (x − 1)2 − 3 is drawn on a grid where each square is 1 unit in both directions. Its lowest point is (1, −3). The graph passes exactly through the grid points (−1, 1) and (2, −2). Read the graph to find the average rate of change of f over [−1, 2]. Say what the sign means.

(−1, 1)(2, −2)−224−4−2246
The parabola f(x) = (x − 1)2 − 3 on a 1-unit grid, with its lowest point at (1, −3).
Show the steps and the answer
  1. Locate the endpoints x=−1 and x=2 on the horizontal axis. The left end is a=−1 and the right end is b=2.The interval [a, b] tells us which two inputs to use, and a < b.
  2. From x=−1, move vertically to the solid curve. It meets the curve 1 square above the x-axis, so f(−1)=1.The height of the graph above an input is the output at that input. Here the curve crosses exactly at a grid point, so no estimate is needed.
  3. From x=2, move vertically to the solid curve. It meets the curve 2 squares below the x-axis, so f(2)=−2.A point below the x-axis has a negative height. Counting squares downward gives a negative output.
  4. Rise =f(2)−f(−1)=−2−1=−3. Run =2−(−1)=3.Both differences are taken right end minus left end, in the same order.
  5. Average rate =−33=−1.The average rate of change is rise divided by run. It is the slope of the straight ruler through (−1, 1) and (2, −2).
  6. Interpret the result: the rate is negative, so f decreased overall on [−1,2]. On average it dropped 1 unit for each 1 unit of x.A negative slope means the right end is lower than the left end. The curve does fall and then rise inside the interval, but the average rate uses only the two endpoints.
Answer
The average rate of change of f over [−1, 2] is −1. The negative sign means f decreased overall on this interval, by 1 unit of output per unit of input on average.
Double-check Compute from the formula. f(−1) = (−1 − 1)2 − 3 = 4 − 3 = 1, and f(2) = (2 − 1)2 − 3 = 1 − 3 = −2. Then −2−12−(−1) = −33 = −1. This matches the graph reading.

Work to write

  1. f(−1) = 1 and f(2) = −2 (read from the graph)
  2. Average rate = f(2)−f(−1)2−(−1)
  3. =−2−13=−33=−1
  4. Negative, so f decreased overall on [−1, 2], by 1 unit per unit of x on average

Rung 2 of 4

A straight line has the same average rate on any interval

The graph of the straight line f(x) = −12x + 2 is drawn on a grid where each square is 1 unit in both directions. The line passes exactly through the grid points (−4, 4), (0, 2) and (2, 1). Read the graph to find the average rate of change of f over [−4, 2]. Then find it over [0, 2] and compare the two results. Say what the sign means.

(−4, 4)(0, 2)(2, 1)−6−4−2246−2246
The line f(x) = −12x + 2 on a 1-unit grid.
Show the steps and the answer
  1. For [−4,2], locate x=−4 (left end) and x=2 (right end) on the horizontal axis.The rule uses a = −4 as the left input and b = 2 as the right input, with a < b.
  2. From x=−4, move vertically to the line. It meets the line 4 squares above the axis, so f(−4)=4. From x=2, the line is 1 square above the axis, so f(2)=1.The height of the graph at each input is the output we need. Both points sit exactly on gridlines, so no estimate is needed.
  3. Rise =f(2)−f(−4)=1−4=−3. Run =2−(−4)=6.Subtract right minus left in both the outputs and the inputs, so the signs stay consistent.
  4. Average rate =−36=−12.The average rate is rise divided by run. This is the slope of the ruler through (−4, 4) and (2, 1).
  5. For [0, 2], read f(0)=2 and f(2)=1. Then rise =1−2=−1, run =2−0=2, and the average rate =−12=−12.Use the same steps on a different interval. This tests whether the rate depends on where we measure.
  6. Compare: both intervals give −12. The rate is negative, so f decreases. On average, it falls 12 unit for every 1 unit increase in x.A straight line has a single, constant slope. Every ruler through two of its points lies along the line itself, so every interval gives the same average rate.
Answer
The average rate of change over [−4, 2] is −12, and over [0, 2] it is also −12. The negative sign means f decreases by 12 unit per unit increase in x on each interval.
Double-check From the formula, f(−4) = −12(−4) + 2 = 4, f(0) = 2 and f(2) = −12(2) + 2 = 1. These match the graph readings. The slope m in f(x) = −12x + 2 is −12, which agrees with both average rates.

Work to write

  1. f(−4)=4,f(2)=1
  2. f(2)−f(−4)2−(−4)=1−46=−36=−12
  3. f(0) = 2, so f(2)−f(0)2−0 = −12 = −12
  4. Same rate on both intervals because the graph is a straight line; negative means f is decreasing

Rung 3 of 4

Average rate of change read from a graph with a scaled vertical axis

A solid curve y = f(x) is drawn on a grid. Each horizontal square is 1 unit and each vertical square is 2 units. At x = −2 the curve crosses exactly the gridline 5 squares above the x-axis. At x = 4 the curve passes halfway between the gridlines 1 and 2 squares below the x-axis. Find the average rate of change of f on [−2, 4] and say what its sign means.

Show the steps and the answer
  1. Mark the interval endpoints on the horizontal axis: a=−2 (left) and b=4 (right).The rule uses a < b, so a is the left end and b is the right end of [−2, 4].
  2. From x=−2, move vertically to the solid curve. It sits on the gridline 5 squares up, so f(−2)=5·2=10.Each vertical square is worth 2 units. The height must be converted from squares to units.
  3. From x=4, move vertically to the solid curve. It lies halfway between 1 and 2 squares below the axis, which is −1.5 squares. So f(4)=−1.5·2=−3.When the point falls between gridlines, estimate the fraction of a square. Below the axis means a negative height.
  4. Rise =f(4)−f(−2)=−3−10=−13. Run =4−(−2)=6.Subtract right minus left in both the numerator and the denominator so the order matches.
  5. Average rate =−136≈−2.17.Rise divided by run gives the slope of the ruler through (−2, 10) and (4, −3).
  6. Interpret the sign: the value is negative, so on average f decreases by about 2.17 units for each 1 unit increase in x over [−2,4].A negative slope means the ruler line falls from left to right.
Answer
Average rate of change = −136 ≈ −2.17. The function decreases on average over [−2, 4].
Double-check Start at (−2, 10) and follow a line of slope −136 for a run of 6. The height changes by 6 · −136 = −13, which gives 10 − 13 = −3. This matches the reading f(4) = −3. The sign is negative, which agrees with the curve ending lower than it started.

Work to write

  1. a=−2,b=4
  2. f(−2) = 10 (5 squares × 2)
  3. f(4) = −3 (−1.5 squares × 2)
  4. f(4)−f(−2)4−(−2)=−3−106=−136
  5. ≈ −2.17; negative, so f decreases on average over [−2, 4]

Rung 4 of 4

Average speed of a road trip from a distance table

A car's distance from home, d(t) in miles, was recorded t hours after it left. The table gives t = 0, 1.5, 4 and 5 hours with distances 0, 84, 220 and 280 miles. Find the average speed of the car on the interval [1.5, 5] and say what the sign means.

input time (hours)output distance from home (miles)001.58442205280
A table of 4 values of distance from home (miles), for time (hours) from 0 to 5.
Show the steps and the answer
input time (hours)output distance from home (miles)001.58442205280↓ evaluate: input given, read the output below it
Table of the car's distance from home at four times.
  1. Locate the interval endpoints in the input row: a=1.5 (left end) and b=5 (right end).The average rate on [a, b] uses only the two endpoint inputs, with a < b. The entry at t = 4 lies inside the interval and is not needed.
  2. Read the output under each endpoint: d(1.5)=84 and d(5)=280.Each table column pairs an input with its output. These two pairs are the points (1.5, 84) and (5, 280) that the ruler would pass through on a graph.
  3. Subtract right minus left in both rows. Rise: d(5)−d(1.5)=280−84=196 miles. Run: 5−1.5=3.5 hours.The formula f(b)−f(a)b−a subtracts in the same order, right end minus left end, on top and on the bottom.
  4. Divide rise by run: 1963.5=56 miles per hour.Average rate of change is change in output per unit change in input. Its units are miles per hour, so the result is an average speed.
  5. Interpret the sign. The rate +56 is positive, so the distance from home increased.A positive slope means the output rises as the input increases. The car was moving away from home at an average of 56 miles each hour.
Answer
Average speed on [1.5, 5] = 280−845−1.5 = 1963.5 = 56 miles per hour. The positive sign means the car's distance from home grew.
Double-check Start at 84 miles at t = 1.5 and add 56 miles per hour for 3.5 hours: 84 + 56 × 3.5 = 84 + 196 = 280. That matches d(5) = 280.

Work to write

  1. a = 1.5, b = 5
  2. d(1.5) = 84, d(5) = 280
  3. average speed = d(5)−d(1.5)5−1.5 = 280−845−1.5
  4. = 1963.5 = 56 miles per hour
  5. Positive: the distance from home increased by 56 miles per hour on average

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: The average rate is the steepness of the curve at the starting point.
The curve rises near −1, but on [−1, 2] it finishes lower than it starts.
✓ Instead: Use the dashed connecting ruler; its average slope is −1.
✗ Not this: The average rate always changes if the interval changes.
The straight line 2x + 5 gives 2 on both [0, 1] and [2, 7].
✓ Instead: A straight line has one slope; a curve's interval averages may differ.
✗ Not this: Read the height of the dashed ruler at an intermediate input as the function output.
The dashed line joins the endpoints but is not the solid curve.
✓ Instead: Read each function value from the solid graph; only the endpoint points are shared.

Tips and tricks

  • Letters are name tags. In [a, b], a is the left input and b the right input; x1 and x2 can name those same jobs.
  • Write what the rate means: 'drops 1 output unit per input unit on [−1, 2]' explains −1.
  • The endpoints must give actual function outputs; a hollow dot by itself leaves that input undefined.

7 The trap

Putting right-minus-left on top and left-minus-right on the bottom reverses the sign. Keep each input paired with its output and use one order twice.

Lesson 3 of 10

Average rate of change from a formula

Subject: MathematicsLeans on: Common denominatorsLeans on: Dividing by a whole numberLeans on: Simplifying a fractionLeans on: Whole numbers as fractionsOpen this lesson
Leans on 0Comes back in 0

1 In plain words

When the function is given as a formula, you build a small table of your own. Plug the left endpoint into the formula to get its output, plug the right endpoint in to get its output, and then subtract and divide exactly as before. A formula is like a recipe: feed it an input and it cooks up the output. You only have to run the recipe twice, once for each end of the interval. Write the whole input in parentheses, especially if it is negative. Fractions show up often, so keep them as fractions until the end.

2 Picture it

input d (cm)output F(d) (newtons)21/261/18↓ evaluate: input given, read the output below it
d (cm): 2, 6; F(d) (newtons): 12, 118.

Reminder

Reminder: Common denominators, Dividing by a whole number, Simplifying a fraction, Whole numbers as fractions
  • Common denominators. When the smaller bottom divides the bigger one, use the bigger bottom: 12 = 918. Otherwise the product of the bottoms works.
  • Dividing by a whole number. −49 ÷ 4 = −49 · 14. Division by a nonzero number means multiplication by its reciprocal.
  • Simplifying a fraction. Divide top and bottom by the same nonzero number: −818 = −49, using ÷ 2 in both.
  • Whole numbers as fractions. 4 = 41 = 82, so a whole number can be renamed to match a fraction's bottom.

3 Why

The formula supplies the output at each allowed input, so evaluating it at the two endpoints gives the points you would read off a perfect graph. Exact fractions avoid early rounding errors. For example, −19 is exact, while −0.11 differs by about 0.001. Evaluate first, subtract next, and divide last, so the output change and the rate remain separate quantities. The units come from the comparison: force in newtons divided by distance in centimeters gives newtons per centimeter.

4 The rule

On [a,b]= [left end, right end], with a < b and both outputs defined, average rate =f(b)−f(a)b−a.

Evaluate the two endpoint outputs first, then subtract and divide. Preserve all nonzero-bottom and square-root restrictions.

The same idea, five ways

Say it

Say 'evaluate both ends, subtract the outputs, divide by the input span'.

Write it

A formula gives the two endpoint outputs needed for the same average-rate calculation.

In math
  • f(b)−f(a)b−a,a<b
  • f(a): output at the left input; f(b): output at the right input
Like

Run one recipe with two different ingredients, then compare the results per change in the ingredient.

See it
input d (cm)output F(d) (newtons)21/261/18↓ evaluate: input given, read the output below it
The same picture again: d (cm): 2, 6; F(d) (newtons): 12, 118.

The same idea, other ways

Run the recipe twice

The force recipe gives F(2) = 12 and F(6) = 118. Those are two outputs, not two rates. Put them in the endpoint table before computing the changes.

input d (cm)output F(d) (newtons)21/261/18↓ evaluate: input given, read the output below it
The same picture a third time: d (cm): 2, 6; F(d) (newtons): 12, 118.
Separate change from rate

The push changes by −49 newton over 4 centimeters. Dividing by 4 turns the whole-interval change into −19 newton per centimeter.

(2, 1/2)(6, 1/18)2461
The dashed ruler through the two curve points has slope −19; it is not part of F.
Parentheses hold an input together

For f(x) = 2 − x2, the input −3 makes f(−3) = 2 − (−3)2 = −7. Squaring the input happens before the outside subtraction. This prevents a sign mistake before the rate calculation begins.

(−3, −7)(1, 1)−2−8−6−4−22
y = −x2 + 2, solid, beside y = 2x − 1, dashed.

Part 1: Negative inputs in a square formula

A negative input is one complete number. Keep it in parentheses so a square acts on the whole number. The minus outside x2 then subtracts that positive square.

  • (−3)2 = 9, while −32 means −(32) = −9.
  • For 2 − x2, the outside subtraction remains outside after substitution.
(−3, −7)(1, 1)−2−8−6−4−22
The same picture again: y = −x2 + 2, solid, beside y = 2x − 1, dashed.

Worked example

A negative endpoint needs parentheses

This asks for the average output change per input unit, using one negative input. Find the average rate of f(x) = 2 − x2 on [−3, 1].

Show the steps and the answer
(−3, −7)(1, 1)−2−8−6−4−22
The same picture a third time: y = −x2 + 2, solid, beside y = 2x − 1, dashed.
  1. f(−3)=2−(−3)2=2−9=−7.The whole input is −3. Squaring (−3) gives (−3)(−3) = 9; the outside subtraction stays outside.
  2. f(1)=2−12=2−1=1.Evaluate the other endpoint in the same formula.
  3. Output change: 1−(−7)=1+7=8.Subtract the whole starting output; taking away a negative adds its opposite.
  4. Input change: 1−(−3)=1+3=4.The endpoint order must match the output subtraction.
  5. 84=2.Dividing the net change 8 by the span 4 gives the change per input unit.
Answer
2.
Double-check The starting output −7 plus 4 · 2 reaches −7 + 8 = 1, the ending output.

Part 2: Negative inputs in a cube formula

A cube multiplies the same number three times. Three negative factors give a negative result. Keep the multiplication term separate and handle its sign too.

  • (−2)3=(−2)(−2)(−2)=−8.
  • −2(−2) = +4, so these two contributions have opposite signs.
(1, −1)−22−16−8816
y = 100, solid, beside y = x − 2, dashed.

Worked example

A harder negative endpoint with a cube

This asks for the average output change per input unit when the formula has a cube. Find the average rate of f(x) = x3 − 2x on [−2, 1].

Show the steps and the answer
(1, −1)−22−16−8816
The same picture again: y = 100, solid, beside y = x − 2, dashed.
  1. f(−2)=(−2)3−2(−2)=−8+4=−4.Three factors of −2 give −8; multiplying −2 by −2 gives +4.
  2. f(1)=13−2(1)=1−2=−1.Do the power and multiplication before subtracting.
  3. Output change: −1−(−4)=−1+4=3.Subtracting the negative starting output adds 4.
  4. Input change: 1−(−2)=1+2=3.The input subtraction uses the same endpoint order.
  5. 33=1.A net rise of 3 over an input span of 3 averages 1 per unit.
Answer
1.
Double-check −4 + 3 · 1 = −1, so the average change reaches the ending output.

5 How

  1. Check that both inputs are allowed: no bottom may be zero, and a real square root needs a nonnegative inside.
  2. Evaluate f(a), placing the whole input in parentheses at every variable position.
  3. Evaluate f(b) the same way.
  4. Subtract f(b) − f(a). If fractions occur, make their bottoms match first.
  5. Divide by b − a. Simplify exact fractions and attach units.

Build the endpoint table from a formula

1
Does the formula have an input in a bottom?
YesExclude inputs that make that bottom 0; verify both endpoints are allowed.
NoContinue.
↓
2
Does a real square root occur?
YesCheck its whole inside is at least 0 at each endpoint; if it is in a bottom, it must be greater than 0.
NoContinue.
↓
3
Is an endpoint negative?
YesReplace the variable by the entire negative number in parentheses.
NoUse parentheses for the input anyway.
↓
4
Are the output fractions over different bottoms?
YesChoose a common denominator, rename both fractions, and subtract only the tops.
NoSubtract the two outputs directly.
  1. Check endpoint restrictions.
  2. Evaluate the formula twice using powers before multiplication, then addition and subtraction.
  3. Compute the output difference with parentheses around a negative starting output.
  4. Divide by the matching input difference and keep the result exact.

6 Worked example

Worked example

Average rate of change of the electric force as two charges move apart

Two small charged spheres repel each other. The size of the force between them is F(r) = 36r2 newtons, where r is the distance between the spheres in meters. Find the average rate of change of F on the interval [2, 4]. Give an exact answer with units.

Show the steps and the answer
input distance r (m)output force F (N)293442.25↓ evaluate: input given, read the output below it
Values of F(r) = 36/r2 at r = 2, 3 and 4 m.
  1. Check the inputs. For r=2,r2=4≠0. For r=4,r2=16≠0. There is no square root, and 2 < 4.Both endpoint outputs must be defined, and the bottom of a fraction may not be zero.
  2. Evaluate F(2)=36(2)2=364=9.Putting the input in parentheses gives the left-end output.
  3. Evaluate F(4)=36(4)2=3616=94.The right-end output is found the same way and simplified.
  4. Subtract: F(4)−F(2)=94−9=94−364=−274.Fractions can be subtracted only after their bottoms match. Here 9 is rewritten as 364.
  5. Divide by 4−2=2: −27/42=−278 N per m.The average rate is the change in output divided by the change in input. Its units are newtons per meter.
Answer
The average rate of change is −278 N per m, which is −3.375 N/m. On average, the force drops by 278 newtons for each meter of extra separation.
Double-check Start at F(2) = 9. Add the rate times the change in input: 9 + (−278)(2) = 9 − 274 = 94. This equals F(4). The negative sign also makes sense, because the force weakens as the spheres move apart.

Work to write

  1. r = 2 and r = 4 both give a nonzero bottom
  2. F(2)=36(2)2=9
  3. F(4)=36(4)2=94
  4. F(4)−F(2)=94−364=−274
  5. Average rate = −27/44−2 = −278 N per m

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

A line from a formula

This asks for the average output change per input unit using a formula. Find the average rate of f(x) = 6x − 7 on [1, 4].

Show the steps and the answer
input xoutput f(x)1−1417↓ evaluate: input given, read the output below it
The two formula evaluations become the endpoint table.
  1. f(1)=6(1)−7=−1; f(4)=6(4)−7=17.Run the formula once for each endpoint input.
  2. Output change: 17−(−1)=18.Subtract the whole starting output; subtracting −1 adds 1.
  3. Input change: 4−1=3.Use the same endpoint order as the outputs.
  4. 183=6.Divide the output change by the nonzero input span.
Answer
6.
Double-check The starting output −1 plus 3 · 6 reaches 17, the ending output.

Rung 2 of 5

Average rate of change of a ball's height on [1, 4]

A ball is thrown upward. Its height above the ground is h(t) = −5(t − 2)2 + 25 meters, where t is the time in seconds after the throw, for 0 ≤ t ≤ 4. Find the average rate of change of h on the interval [1, 4]. Give an exact value with units.

Show the steps and the answer
(1, 20)(4, 5)2439152127
Graph of h(t) = −5(t − 2)2 + 25 with the endpoints (1, 20) and (4, 5) marked.
  1. Check the inputs t=1 and t=4. The formula has no fraction bottom and no square root. Both values lie in 0≤t≤4, so both outputs are defined.The average rate needs both endpoint outputs to exist. A square formula has no zero-bottom or root restriction, so only the stated time range matters.
  2. Evaluate h(1)=−5((1)−2)2+25=−5(−1)2+25=−5(1)+25=20.Putting the input in parentheses keeps the order right. We square (−1) first, then multiply by −5.
  3. Evaluate h(4)=−5((4)−2)2+25=−5(2)2+25=−5(4)+25=5.The right-end output is found the same way as the left-end output.
  4. Subtract: h(4)−h(1)=5−20=−15.The top of the rate is the change in output, right end minus left end. Here it is a change in height of −15 meters.
  5. Divide by 4−1=3: −153=−5. Attach units: meters per second.Average rate = h(b)−h(a)b−a. The output is in meters and the input is in seconds.
Answer
The average rate of change is −5 meters per second. On average, the ball's height drops 5 meters each second between t = 1 and t = 4.
Double-check Expand: h(t) = −5(t2 − 4t + 4) + 25 = −5t2 + 20t + 5. This gives h(1) = −5 + 20 + 5 = 20 and h(4) = −80 + 80 + 5 = 5, which match. For f(t) = At2 + Bt + C, the average rate on [a, b] is A(a + b) + B, so −5(1 + 4) + 20 = −5. ✓

Work to write

  1. Both inputs t = 1 and t = 4 are allowed (0 ≤ t ≤ 4; no bottoms or roots)
  2. h(1)=−5((1)−2)2+25=20
  3. h(4)=−5((4)−2)2+25=5
  4. h(4)−h(1)=5−20=−15
  5. h(4)−h(1)4−1=−153=−5
  6. Average rate of change = −5 meters per second

Rung 3 of 5

Parentheses for a negative endpoint

This asks for the average output change per input unit, using one negative input. Find the average rate of f(x) = 2 − x2 on [−3, 1].

Show the steps and the answer
(−3, −7)(1, 1)−2−8−6−4−22
The same picture, shown 4 times now: y = −x2 + 2, solid, beside y = 2x − 1, dashed.
  1. f(−3)=2−(−3)2=2−9=−7.The whole input is −3. Squaring (−3) gives (−3)(−3) = 9; the outside subtraction stays outside.
  2. f(1)=2−12=2−1=1.Evaluate the other endpoint in the same formula.
  3. Output change: 1−(−7)=1+7=8.Subtract the whole starting output; taking away a negative adds its opposite.
  4. Input change: 1−(−3)=1+3=4.The endpoint order must match the output subtraction.
  5. 84=2.Dividing the net change 8 by the span 4 gives the change per input unit.
Answer
2.
Double-check The starting output −7 plus 4 · 2 reaches −7 + 8 = 1, the ending output.

Rung 4 of 5

Average rate of change of an inverse-square force on [2, 5]

The magnetic force between two magnets r centimeters apart is modeled by F(r) = 400r2 newtons, for r > 0. Find the average rate of change of F on the interval [2, 5]. Give an exact value with units.

Show the steps and the answer
input distance r (cm)output force F (N)2100516↓ evaluate: input given, read the output below it
Endpoint values of F(r) = 400r2: F(2) = 100 N and F(5) = 16 N.
  1. Check the inputs: r=2 gives bottom 22=4 and r=5 gives bottom 52=25. Neither bottom is zero, and both inputs satisfy r > 0.Both endpoint outputs must be defined before the formula f(b)−f(a)b−a can be used, and F has a variable in the bottom.
  2. Evaluate the left endpoint: F(2)=400(2)2=4004=100 newtons.Putting the whole input in parentheses at the variable position makes sure the input is squared correctly.
  3. Evaluate the right endpoint: F(5)=400(5)2=40025=16 newtons.The right endpoint b = 5 is evaluated the same way as the left endpoint.
  4. Subtract in the order right minus left: F(5)−F(2)=16−100=−84 newtons.The numerator is f(b) − f(a). Both outputs are whole numbers, so no common bottom is needed.
  5. Divide by the change in input: b−a=5−2=3, so the average rate =−843=−28 newtons per centimeter.The average rate is the change in output divided by the change in input. The units are output units per input unit.
Answer
The average rate of change of F on [2, 5] is −28 newtons per centimeter. On average, the force drops 28 N for each extra centimeter of separation.
Double-check Start at F(2) = 100 and apply the rate over 3 cm: 100 + (−28)(3) = 100 − 84 = 16, which equals F(5). The negative sign also fits the model, because an inverse-square force weakens as r grows.

Work to write

  1. Both inputs are allowed: (2)2 = 4 ≠ 0 and (5)2 = 25 ≠ 0
  2. F(2)=400(2)2=100
  3. F(5)=400(5)2=16
  4. F(5)−F(2)=16−100=−84
  5. b−a=5−2=3
  6. Average rate = −843 = −28 newtons per centimeter

Rung 5 of 5

A negative endpoint with a cube

This asks for the average output change per input unit when the formula has a cube. Find the average rate of f(x) = x3 − 2x on [−2, 1].

Show the steps and the answer
(1, −1)−22−16−8816
The same picture a third time: y = 100, solid, beside y = x − 2, dashed.
  1. f(−2)=(−2)3−2(−2)=−8+4=−4.Three factors of −2 give −8; multiplying −2 by −2 gives +4.
  2. f(1)=13−2(1)=1−2=−1.Do the power and multiplication before subtracting.
  3. Output change: −1−(−4)=−1+4=3.Subtracting the negative starting output adds 4.
  4. Input change: 1−(−2)=1+2=3.The input subtraction uses the same endpoint order.
  5. 33=1.A net rise of 3 over an input span of 3 averages 1 per unit.
Answer
1.
Double-check −4 + 3 · 1 = −1, so the average change reaches the ending output.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: For f(x) = 2 − x2, f(−3) = 2 + 9 = 11.
The square acts on the entire input (−3), giving +9. The outside minus then subtracts 9.
✓ Instead: f(−3)=2−(−3)2=2−9=−7.
✗ Not this: The force rate is −49 because that is the force change.
That change occurs over 4 centimeters. The rate still needs division by the input change.
✓ Instead: −49 ÷ 4 = −19 newton per centimeter.
✗ Not this: Round F(6) to 0.06 before subtracting: (0.06 − 0.5) ÷ 4 = −0.11 exactly.
0.06 is a rounded replacement for 118, so the rate −0.11 is approximate rather than exact.
✓ Instead: Keep 118 through the calculation; the exact rate is −19.

Tips and tricks

  • Before using the rate formula, write a tiny endpoint table so the two inputs stay paired with their outputs.
  • Use parentheses as a seat belt around every substituted input and every negative output you subtract.
  • A fraction bottom is also called a denominator. To simplify, divide both its top and bottom by the same nonzero number.

7 The trap

Stopping after subtracting gives only the output change. Divide by the input change to get the average rate.

Lesson 4 of 10

When an endpoint is a letter, the answer is an expression

Subject: MathematicsLeans on: Sum-and-product factoringLeans on: Difference of squaresLeans on: Solving a one-step equationLeans on: Canceling a common factorOpen this lesson
Leans on 0Comes back in 0

1 In plain words

Sometimes one end of the interval is a letter, like [0, a] or [5, a]. You follow exactly the same steps, but the letter stays a letter, so the answer is an expression such as a + 7 instead of a single number. That expression is a reusable shortcut: put in any allowed endpoint for a and it gives the average rate of change, with no new work. Think of a taxi fare formula: once you have it, you can price different trip lengths. Here you are building the rate formula rather than the fare formula. Factoring means rewriting a sum as multiplication; that can reveal a whole factor to cancel.

2 Picture it

(5, 27)(7, 55)−44816324864(5, 27)(7, 55)−4−2246816324864
y = (x + 1)2 − 9, solid, beside y = 14x − 43, dashed.

Reminder

Reminder: Sum-and-product factoring, Difference of squares, Solving a one-step equation, Canceling a common factor
  • Sum-and-product factoring. (a + p)(a + q) = a2 + (p + q)a + pq. For a2 + 2a − 35, choose p = 7 and q = −5.
  • Difference of squares. b2 − 4 = (b − 2)(b + 2), since the +2b and −2b terms cancel when multiplied out.
  • Solving a one-step equation. In b + 3 = 11, subtract 3 from both sides to get b = 8, then substitute 8 back into the equation.
  • Canceling a common factor. a(a+3)a = a + 3 requires a ≠ 0. Canceling reverses multiplication, not addition.

3 Why

A letter stands for a number you have not picked, so arithmetic still applies. For the harder example, (a + 7)(a − 5) multiplies out to a2 − 5a + 7a − 35 = a2 + 2a − 35. This directly proves the factorization. The denominator a − 5 is then a factor of the whole numerator, so it cancels when a ≠ 5. At a = 5 the original rate still divides by zero. The matching-factor method works for these polynomial examples; other formulas can require fraction arithmetic or a different simplification.

4 The rule

For a fixed input 5 and an endpoint a, rate =f(a)−f(5)a−5, with a≠5 and both outputs defined.

On [5, a], a > 5; if a < 5, use the sorted interval [a, 5]. Factor a polynomial numerator before canceling a matching whole factor.

The same idea, five ways

Say it

Say 'the endpoint is a number I have not chosen yet'.

Write it

The average rate can be an expression in the endpoint letter; choosing an endpoint then gives a numerical rate.

In math
  • g(a)−g(0)a,a≠0
  • f(a)−f(5)a−5,a≠5
  • [5, a] for a > 5; [a, 5] for a < 5
Like

A reusable taxi price formula works for many trip lengths; a reusable rate formula works for many endpoints.

See it
input aoutput rate on [5, a]613714815916↓ evaluate: input given, read the output below it
a: 6, 7, 8, 9; rate on [5, a]: 13, 14, 15, 16.

The same idea, other ways

Numbers before a formula

For f(x) = x2 + 2x − 8, start at 5. Read the drawn columns as input endpoint above average rate. At a = 7, the column gives 14 because 55−277−5 = 14. Each column follows the expression a + 7; the worked algebra below explains why the same shortcut serves every allowed endpoint.

input aoutput rate on [5, a]613714815916↓ evaluate: input given, read the output below it
The same picture again: a: 6, 7, 8, 9; rate on [5, a]: 13, 14, 15, 16.
A reusable price card

A taxi price card accepts a chosen trip length. In the same way, the rate expression a + 7 accepts a chosen endpoint. Choosing a = 6 gives 13; choosing a = 8 gives 15. The endpoint letter names a number, rather than an instruction to solve yet.

a = 8a + 7inputoutput15
Once built, the rate expression can be evaluated for a chosen endpoint.
Factoring runs multiplication backward

The product (a + 7)(a − 5) gives a2 + 2a − 35 when multiplied out. This makes the matching whole factor a − 5 visible. Canceling then means dividing that factor by itself, which gives 1 when it is nonzero.

a2+2a−35

=(a+7)(a−5)

a ≠ 5

Part 1: A letter becomes a number when you choose it

A letter endpoint allows many numerical intervals to be described at once. A rate table can show the pattern before you rebuild it algebraically.

  • For f(x) = x2 + 2x − 8 starting at 5, rate = a + 7.
  • x, t, a, and b are names; their role is determined by the formula and the question.
input aoutput rate on [5, a]613714815916↓ evaluate: input given, read the output below it
The same picture a third time: a: 6, 7, 8, 9; rate on [5, a]: 13, 14, 15, 16.

Worked example

A letter names an endpoint number

This asks for one numerical average before rebuilding all endpoints at once. Take a = 7 in f(x) = x2 + 2x − 8 on [5, a].

Show the steps and the answer
input aoutput rate on [5, a]613714815916↓ evaluate: input given, read the output below it
The same picture, shown 4 times now: a: 6, 7, 8, 9; rate on [5, a]: 13, 14, 15, 16.
  1. The endpoint interval becomes [5, 7], with f(5)=25+10−8=27 and f(7)=49+14−8=55.Choosing a = 7 gives a numerical value to the endpoint letter.
  2. 55−277−5=282=14.The ordinary endpoint formula still applies; the rate table highlights this result under input 7.
Answer
14.
Double-check The reusable answer a + 7 gives 7 + 7 = 14, agreeing with the direct calculation.

Part 2: A three-term numerator needs a factoring puzzle

After evaluating and subtracting, a2 + 2a − 35 needs two factors. Find numbers that multiply to −35 and add to 2. Check the product before canceling.

  • 7 · (−5) = −35 and 7 + (−5) = 2.
  • (a+7)(a−5)=a2+2a−35.
  • The original rate excludes a = 5.
(5, 27)(7, 55)−44816324864
The same picture again: y = (x + 1)2 − 9, solid, beside y = 14x − 43, dashed.

Worked example

Average rate of change on [5, a] for a quadratic, then solving for a

Let f(x) = x2 − 4x + 1, which can also be written f(x) = (x − 2)2 − 3. Consider the interval [5, a], where a > 5.
(1) Write the average rate of change of f on [5, a] as a simplified expression in a.
(2) Find the value of a for which the average rate of change on [5, a] equals 9.

Show the steps and the answer
(5, 6)(8, 33)−22468108162432
The parabola f(x) = (x − 2)2 − 3 with the points (5, 6) and (8, 33) marked.
  1. Evaluate the fixed endpoint: f(5)=52−4(5)+1=25−20+1=6.The input 5 is a number, so its output is a number. It is one of the two outputs in the difference quotient.
  2. Substitute the letter: f(a)=a2−4a+1.The other endpoint is the letter a. Replacing x with a gives the second output as an expression.
  3. Write the quotient: rate =f(a)−f(5)a−5=(a2−4a+1)−6a−5=a2−4a−5a−5, with a≠5.Average rate of change is the change in output over the change in input. Both differences take the same order: a first, then 5. The denominator cannot be zero, so a ≠ 5. This also follows from the condition a > 5.
  4. Factor the numerator: a2−4a−5=(a−5)(a+1). Check by multiplying back: (a−5)(a+1)=a2+a−5a−5=a2−4a−5. ✓Two numbers that multiply to −5 and add to −4 are −5 and +1. Multiplying back confirms the factorization before anything is canceled.
  5. Cancel the common factor: (a−5)(a+1)a−5=a+1, for a≠5.The factor (a − 5) multiplies the whole numerator, so it may be canceled. The restriction a ≠ 5 comes from the original quotient and stays in place.
  6. Set the rate equal to 9 and solve: a+1=9, so a=8.The simplified expression gives the rate for every allowed a. Setting it equal to the given rate leaves a linear equation.
  7. Check the endpoint. The value a=8 satisfies a > 5 and a≠5. In the original quotient, f(8)=64−32+1=33, and 33−68−5=273=9. ✓Any solution must be checked against the original restrictions and the original unsimplified quotient, not only the simplified expression.
Answer
(1) The average rate of change on [5, a] is a + 1, for a ≠ 5 (here a > 5). (2) a = 8, so the interval is [5, 8].
Double-check f(5) = 6 and f(8) = 33. The rate is 33−68−5 = 273 = 9. The expression a + 1 at a = 8 also gives 9, so the two agree.

Work to write

  1. f(5) = 6
  2. f(a)=a2−4a+1
  3. rate = a2−4a−5a−5, a ≠ 5
  4. a2−4a−5=(a−5)(a+1)
  5. rate = a + 1, a ≠ 5
  6. a + 1 = 9 ⇒ a = 8
  7. Check: 33−68−5 = 9 ✓

Part 3: Working backward from a given rate

Evaluate means an input is given and you find the output. Solve means a target is given and you find the input that produces it. First build the rate expression, then solve for the endpoint that gives the target rate.

  • A target rate produces an equation for the endpoint.
  • Check the candidate in the original quotient, not only the shortened expression.
  • A canceled denominator still leaves an excluded input.
=b + 311do the same thing to both sides
The average rate is known, so solve for the endpoint that balances this equation.

Worked example

A given rate of change that leads to an endpoint left of 5

Let f(x) = x2 + 2x − 6, which can also be written f(x) = (x + 1)2 − 7. Consider the interval whose endpoints are 5 and a, where a ≠ 5.
(1) Write the average rate of change of f between 5 and a as a simplified expression in a.
(2) Find the value of a for which this average rate of change equals 4. Write the interval in sorted order and check your answer in the original quotient.

Show the steps and the answer
(5, 29)(−3, −3)−44−88162432
The parabola f(x) = (x + 1)2 − 7 with the points (−3, −3) and (5, 29).
  1. f(5)=52+2(5)−6=25+10−6=29The fixed numerical endpoint is evaluated first, so one output is a plain number.
  2. f(a)=a2+2a−6Substituting the letter a for x gives the other output as an expression. It is defined for every real a.
  3. rate =f(a)−f(5)a−5=(a2+2a−6)−29a−5=a2+2a−35a−5, with a≠5Average rate of change is the change in output over the matching change in input. Both differences take the same order, a first and 5 second. The denominator cannot be zero, so a ≠ 5.
  4. Factor: a2+2a−35=(a−5)(a+7). Multiply back to verify: (a−5)(a+7)=a2+7a−5a−35=a2+2a−35✓We need two numbers with product −35 and sum 2. These are 7 and −5. Multiplying back confirms the factoring.
  5. rate =(a−5)(a+7)a−5=a+7, for a≠5(a − 5) multiplies the whole numerator, so it can be canceled with the denominator. The restriction a ≠ 5 comes from the original quotient and is kept.
  6. Set a+7=4, so a=−3.The rate is given as 4, so set the simplified expression equal to 4 and solve. The result a = −3 satisfies a ≠ 5.
  7. Since −3<5, the sorted interval is [−3,5].When the letter endpoint is less than 5, the interval is written from smaller to larger. The quotient f(a)−f(5)a−5 gives the same value in either order.
  8. Check in the original quotient: f(−3)=9−6−6=−3, so f(−3)−f(5)−3−5=−3−29−8=−32−8=4✓Checking the endpoint in the unsimplified quotient confirms that the cancellation and the solving are correct.
Answer
(1) The average rate of change is a + 7, for a ≠ 5. (2) a = −3, so the interval is [−3, 5].
Double-check f(5) = 29 and f(−3) = −3. Over [−3, 5] the output rises by 29 − (−3) = 32 while the input rises by 5 − (−3) = 8. The rate is 328 = 4, which matches the given rate. The formula agrees: a + 7 = −3 + 7 = 4.

Work to write

  1. f(5) = 29
  2. rate = a2+2a−6−29a−5 = a2+2a−35a−5, a ≠ 5
  3. a2+2a−35=(a−5)(a+7)
  4. rate = a + 7, a ≠ 5
  5. a+7=4⇒a=−3
  6. interval [−3, 5]
  7. check: −3−29−3−5 = −32−8 = 4

Part 4: Working backward when the rate contains a fraction

First combine the function outputs into one fraction. Dividing by the input difference can cancel a whole factor, leaving a rate equation you solve by doing the same multiplication to both sides.

  • For g(x) = 12x starting at 3, rate = −4b, with b ≠ 0 and b ≠ 3.
  • Target rate −1 gives b = 4.
=−4 ÷ b−1do the same thing to both sides
A balance: −4 ÷ b on the left, −1 on the right.

Worked example

Work backward with a fraction

This asks for an endpoint giving a specified average rate. Let g(x) = 12x. Starting at 3, find b > 3 so the average rate on [3, b] is −1.

Show the steps and the answer
=−4 ÷ b−1do the same thing to both sides
The same picture again: A balance: −4 ÷ b on the left, −1 on the right.
  1. Require b≠0 for g(b), and b≠3 for the input change. g(3)=123=4; g(b)=12b.Both the function output and the rate quotient need nonzero bottoms.
  2. g(b)−g(3)=12b−4=12b−4bb=12−4bb.Write the whole number 4 over bottom b by multiplying it by bb = 1.
  3. 12−4b=−4(b−3).Multiplying −4 through the parentheses gives −4b + 12, the same numerator.
  4. The rate is −4(b−3)b÷(b−3)=−4(b−3)b·1b−3=−4b.Dividing by b − 3 means multiplying by its reciprocal; the factor cancels only because b ≠ 3.
  5. Set −4b=−1, then multiply both sides by b: −4=−b.This finds the input that gives rate −1; multiplying by b is allowed because b ≠ 0.
  6. Multiply both sides by −1: b=4.The same multiplication preserves the equality and makes b positive.
  7. 4 > 3, 4≠0, and 4≠3.The candidate satisfies the requested interval and both original restrictions.
Answer
b = 4.
Double-check In the original rate, g(4) = 3 and g(3) = 4, so 3−44−3 = −11 = −1.

5 How

  1. Evaluate the fixed numerical endpoint.
  2. Substitute the endpoint letter to write the other output.
  3. Subtract the outputs and write the matching input difference.
  4. For a polynomial, factor the numerator and verify by multiplying back.
  5. Cancel only a matching factor multiplying the whole numerator, preserving the original restrictions.
  6. If the rate is given, set the simplified rate expression equal to that number, solve, and check the endpoint in the original quotient.

A letter endpoint, forward or backward

1
Does the fixed endpoint equal 0?
YesCompute f(0); the input difference is the endpoint letter.
NoSubtract the actual fixed input in the denominator.
↓
2
Is the numerator a polynomial with a common factor?
YesWrite each term as that common factor times something and factor it out.
NoLook for a difference of squares or the sum-and-product pattern; for fractions use a common bottom.
↓
3
Is the rate given and the endpoint requested?
YesSet the simplified expression equal to the given rate, solve, and check the original restrictions.
NoLeave the result as an expression with restrictions.
↓
4
Does the proposed endpoint make any original denominator 0?
YesReject it even if it satisfies the simplified expression.
NoCheck it in the original rate formula.
  1. Build the endpoint quotient before simplifying.
  2. Record all excluded inputs from function bottoms and the input difference.
  3. Use common factors, a difference of squares, or the sum-and-product factor puzzle when applicable.
  4. If requested, solve the rate equation and check the candidate in the original quotient.

6 Worked example

Worked example

Average rate of change from 5 to a letter endpoint, then solving for the endpoint

Let f(x) = x2 − 2x − 2. (a) Write the average rate of change of f between the inputs 5 and a, where a ≠ 5, as a simplified expression in a. (b) The average rate of change of f between 5 and a equals 7. Find a and state the interval.

Show the steps and the answer
(4, 6)(5, 13)−2246−44812
The graph of f(x) = (x − 1)2 − 3 = x2 − 2x − 2, marking the points (4, 6) and (5, 13).
  1. f(5)=52−2(5)−2=25−10−2=13.The fixed numerical endpoint gives a number, so evaluate it first.
  2. f(a)=a2−2a−2.Replace every x with the endpoint letter a to get the other output.
  3. rate =f(a)−f(5)a−5=(a2−2a−2)−13a−5=a2−2a−15a−5, with a≠5.The average rate is the change in output over the matching change in input, in the same order. a = 5 would make the denominator zero.
  4. a2 − 2a − 15 = (a − 5)(a + 3). Check: (a−5)(a+3)=a2+3a−5a−15=a2−2a−15.Factor the polynomial numerator so that any factor matching the denominator becomes visible. Multiplying back confirms the factoring.
  5. rate =(a−5)(a+3)a−5=a+3, for a≠5.(a − 5) multiplies the whole numerator and equals the denominator, so it cancels. The restriction a ≠ 5 from the original quotient still holds.
  6. Set a+3=7, so a=4.The given rate equals the simplified expression, so solve that equation for the endpoint.
  7. Check a=4 in the original quotient: f(4)=16−8−2=6, and 6−134−5=−7−1=7. Since 4 < 5, the sorted interval is [4, 5].The endpoint must satisfy a ≠ 5 and give the stated rate in the unsimplified quotient. An interval is written with its smaller endpoint first.
Answer
(a) The rate is a + 3, for a ≠ 5. (b) a = 4, and the interval is [4, 5].
Double-check Use a = 6 in part (a): f(6) = 36 − 12 − 2 = 22, and 22−136−5 = 9. The formula gives 6 + 3 = 9, so the expression agrees. For part (b), the points (4, 6) and (5, 13) give a slope of 13−65−4 = 7.

Work to write

  1. f(5) = 13
  2. f(a)=a2−2a−2
  3. a2−2a−15a−5,a≠5
  4. a2−2a−15=(a−5)(a+3)
  5. rate = a + 3, a ≠ 5
  6. a + 3 = 7 ⇒ a = 4
  7. Check: 6−134−5 = 7; interval [4, 5]

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 7

A linear formula needs no factoring puzzle

This asks for the average rate from a fixed input to any different input b. Find it for f(x) = 4x + 9, starting at −2.

Show the steps and the answer
(−2, 1)(6, 33)run 1rise 4−22468162432
A straight line adds 4 output units per input unit.
  1. f(−2)=4(−2)+9=−8+9=1; f(b)=4b+9.The first output is numerical; the second keeps the endpoint letter.
  2. Output change: f(b)−f(−2)=(4b+9)−1=4b+8.Subtract the entire starting output.
  3. Input change: b−(−2)=b+2.The input change follows the same endpoint order.
  4. 4b+8=4(b+2), so 4(b+2)b+2=4, with b≠−2.The factor b + 2 multiplies the whole top. The original bottom excludes b = −2.
Answer
4, for b ≠ −2. Use [−2, b] if b > −2, and [b, −2] if b < −2.
Double-check Choose b = 6: f(6) = 33 and 33−16−(−2) = 328 = 4.

Rung 2 of 7

Average rate of change on [5, a] when f(5) = 0

Let f(x) = x2 − 3x − 10. (a) Write the average rate of change of f from x = 5 to x = a, where a ≠ 5, as a simplified expression in a. (b) The average rate of change on [5, a] is 9. Find a.

Show the steps and the answer
(5, 0)(7, 18)−448−12−661218
The parabola f(x) = x2 − 3x − 10 crosses the x-axis at (5, 0) and passes through (7, 18).
  1. f(5)=52−3(5)−10=25−15−10=0.The fixed numerical endpoint is evaluated first. Its output is 0, so subtracting it will not change the numerator.
  2. f(a)=a2−3a−10.Replacing x with the endpoint letter a gives the other output.
  3. Rate =f(a)−f(5)a−5=(a2−3a−10)−0a−5=a2−3a−10a−5, with a≠5.Average rate of change is the change in output divided by the matching change in input, taken in the same order. The denominator cannot be 0.
  4. Factor the numerator: a2−3a−10=(a−5)(a+2). Check by multiplying back: (a−5)(a+2)=a2+2a−5a−10=a2−3a−10.We need two numbers with product −10 and sum −3. Those are −5 and 2. Multiplying back confirms the factoring.
  5. Rate =(a−5)(a+2)a−5=a+2, for a≠5.(a − 5) multiplies the whole numerator and matches the denominator, so it cancels. The restriction a ≠ 5 from the original quotient still applies.
  6. Set a+2=9, so a=7. Since 7 > 5, the interval is [5, 7].The given rate equals the simplified expression. The solution a = 7 is not 5, so it is allowed, and it gives the sorted interval [5, a].
  7. Check in the original quotient: f(7)=49−21−10=18, so 18−07−5=182=9.The endpoint must be checked in the original, unsimplified quotient to confirm the rate is 9.
Answer
(a) Rate = a + 2, for a ≠ 5. (b) a = 7.
Double-check Test a = 6 in part (a): f(6) = 36 − 18 − 10 = 8, so 8−06−5 = 8, and a + 2 = 8. For part (b), the secant line from (5, 0) to (7, 18) has slope 182 = 9.

Work to write

  1. f(5) = 0
  2. f(a)=a2−3a−10
  3. Rate = a2−3a−10a−5, a ≠ 5
  4. a2−3a−10=(a−5)(a+2)
  5. Rate = a + 2, a ≠ 5
  6. a + 2 = 9, so a = 7
  7. Check: 18−07−5 = 9

Rung 3 of 7

A difference of squares

This asks for the rate expression in b, using a fixed starting input 6. Find the average rate of f(x) = x2 + 5 from 6 to b.

Show the steps and the answer
start 6b = 924681030507090
y = x2 + 5, solid, beside y = 15x − 49, dashed.

b2+5−41=b2−36

=(b−6)(b+6)

rate = b + 6

start 6b = 924681030507090
The same picture again: y = x2 + 5, solid, beside y = 15x − 49, dashed.
  1. f(6)=62+5=36+5=41; f(b)=b2+5.Evaluate the fixed endpoint and substitute the letter at the other endpoint.
  2. f(b)−f(6)=(b2+5)−41=b2−36.Subtract the whole starting output; the constants combine: 5 − 41 = −36.
  3. b2−36=b2−62=(b−6)(b+6).Multiplying these two factors gives b2 + 6b − 6b − 36; the middle terms cancel.
  4. (b−6)(b+6)b−6=b+6, with b≠6.The input change is b − 6, and that nonzero factor multiplies the whole numerator.
Answer
b + 6, for b ≠ 6. Sort the interval endpoints according to whether b is greater or less than 6.
Double-check Choose b = 9: f(9) = 86, so 86−419−6 = 453 = 15 = 9 + 6.

Rung 4 of 7

A coefficient and a difference of squares

This asks for a rate expression rather than one rate number. Find the average rate of f(x) = 3x2 − 5 on [2, b], with b > 2.

Show the steps and the answer

3b2−12

=3(b−2)(b+2)

rate = 3b + 6

  1. f(2)=3(2)2−5=3(4)−5=7; f(b)=3b2−5.Evaluate the fixed input 2 and then substitute the endpoint b.
  2. f(b)−f(2)=(3b2−5)−7=3b2−12.Subtract the whole starting output; −5 − 7 = −12.
  3. 3b2−12=3(b2−4).Both terms contain a factor of 3.
  4. b2−4=(b−2)(b+2).Multiplying back gives b2 + 2b − 2b − 4 = b2 − 4.
  5. 3(b−2)(b+2)b−2=3(b+2)=3b+6, with b≠2.The input difference b − 2 is a nonzero factor of the entire top.
Answer
  • 3b + 6, for b ≠ 2. The stated interval [2, b] uses b > 2
  • for b < 2 use [b, 2].
Double-check The original check b = 3 gives f(3) = 22 and 22−73−2 = 15 = 3(3) + 6. With b = 5, f(5) = 70 and 70−75−2 = 633 = 21.

Rung 5 of 7

Average rate on [5, a] for a shifted parabola, then find a

Let f(x) = (x − 1)2 − 3, which expands to f(x) = x2 − 2x − 2. (a) Write the average rate of change of f between x = 5 and x = a as a simplified expression in a, with a ≠ 5. (b) The average rate of change on [5, a] is 10. Find a.

Show the steps and the answer
(5, 13)(7, 33)−224688162432
Graph of f(x) = (x − 1)2 − 3 with the endpoints (5, 13) and (7, 33).
  1. f(5)=52−2(5)−2=25−10−2=13.The fixed numerical endpoint is evaluated first, so one output is a plain number.
  2. f(a)=a2−2a−2.Substituting the letter a for x gives the other output as an expression.
  3. Rate =f(a)−f(5)a−5=(a2−2a−2)−13a−5=a2−2a−15a−5, with a≠5.The average rate is the change in output over the matching change in input, taken in the same order. a ≠ 5 keeps the denominator nonzero.
  4. Factor a2 − 2a − 15. Look for two numbers with product −15 and sum −2. These are −5 and 3, so a2 − 2a − 15 = (a − 5)(a + 3). Check: (a−5)(a+3)=a2+3a−5a−15=a2−2a−15.This is the sum-and-product puzzle for a monic quadratic. Multiplying back confirms the factorization is correct.
  5. (a−5)(a+3)a−5=a+3, for a≠5.(a − 5) multiplies the whole numerator and matches the denominator, so it cancels. The restriction a ≠ 5 from the original quotient still applies.
  6. Set a+3=10, so a=7. Since 7 > 5, the interval is [5, 7].The given rate equals the simplified expression. The value a = 7 is allowed because it is not 5.
Answer
(a) The rate is a + 3, for a ≠ 5. (b) a = 7.
Double-check In the original quotient: f(7) = 49 − 14 − 2 = 33, and f(5) = 13. Then 33−137−5 = 202 = 10, which matches the given rate.

Work to write

  1. f(5) = 13
  2. f(a)=a2−2a−2
  3. a2−2a−15a−5,a≠5
  4. a2−2a−15=(a−5)(a+3)
  5. rate = a + 3, a ≠ 5
  6. a + 3 = 10 ⇒ a = 7
  7. check: 33−137−5 = 10

Rung 6 of 7

Finding an unknown endpoint from a given average rate of change

Let f(x) = x2 − 6x + 4, which can also be written f(x) = (x − 3)2 − 5. (a) Write the average rate of change of f between the inputs 5 and a, where a ≠ 5, as a simplified expression in a. (b) The average rate of change of f between 5 and a equals 3. Find a and state the interval, written with the smaller endpoint first.

Show the steps and the answer
(4, −4)(5, −1)2468−6−4−2246
Graph of f(x) = (x − 3)2 − 5 with the points (4, −4) and (5, −1).
  1. f(5)=52−6(5)+4=25−30+4=−1The fixed endpoint is a number, so its output can be computed directly.
  2. f(a)=a2−6a+4Substituting the letter a for x gives the output at the other endpoint.
  3. rate =f(a)−f(5)a−5=(a2−6a+4)−(−1)a−5=a2−6a+5a−5Average rate of change is the change in output divided by the matching change in input, with the inputs in the same order as the outputs.
  4. a2 − 6a + 5 = (a − 1)(a − 5). Check: (a−1)(a−5)=a2−5a−a+5=a2−6a+5Factoring shows whether the numerator contains the factor (a − 5). Multiplying back confirms the factoring.
  5. rate =(a−1)(a−5)a−5=a−1, for a≠5(a − 5) multiplies the whole numerator, so it cancels. The restriction a ≠ 5 stays because the original quotient is undefined at a = 5.
  6. a−1=3, so a=4Setting the simplified rate equal to the given value gives a linear equation in a.
  7. a=4 satisfies a≠5. Since 4 < 5, the sorted interval is [4, 5].The solution must satisfy the original restriction. An interval is written with the smaller endpoint first.
Answer
(a) The average rate of change is a − 1, with a ≠ 5. (b) a = 4, and the interval is [4, 5].
Double-check f(4) = 16 − 24 + 4 = −4 and f(5) = −1. In the original quotient, f(4)−f(5)4−5 = −4−(−1)−1 = −3−1 = 3, which matches the given rate.

Work to write

  1. f(5)=−1
  2. f(a)=a2−6a+4
  3. a2−6a+5a−5
  4. a2−6a+5=(a−1)(a−5)
  5. rate = a − 1, a ≠ 5
  6. a − 1 = 3 so a = 4
  7. interval [4, 5]
  8. check: −4−(−1)4−5 = 3

Rung 7 of 7

Work backward with a fraction

This asks for an endpoint giving a specified average rate. Let g(x) = 12x. Starting at 3, find b > 3 so the average rate on [3, b] is −1.

Show the steps and the answer
=−4 ÷ b−1do the same thing to both sides
The same picture a third time: A balance: −4 ÷ b on the left, −1 on the right.
  1. Require b≠0 for g(b), and b≠3 for the input change. g(3)=123=4; g(b)=12b.Both the function output and the rate quotient need nonzero bottoms.
  2. g(b)−g(3)=12b−4=12b−4bb=12−4bb.Write the whole number 4 over bottom b by multiplying it by bb = 1.
  3. 12−4b=−4(b−3).Multiplying −4 through the parentheses gives −4b + 12, the same numerator.
  4. The rate is −4(b−3)b÷(b−3)=−4(b−3)b·1b−3=−4b.Dividing by b − 3 means multiplying by its reciprocal; the factor cancels only because b ≠ 3.
  5. Set −4b=−1, then multiply both sides by b: −4=−b.This finds the input that gives rate −1; multiplying by b is allowed because b ≠ 0.
  6. Multiply both sides by −1: b=4.The same multiplication preserves the equality and makes b positive.
  7. 4 > 3, 4≠0, and 4≠3.The candidate satisfies the requested interval and both original restrictions.
Answer
b = 4.
Double-check In the original rate, g(4) = 3 and g(3) = 4, so 3−44−3 = −11 = −1.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: Cancel a from a2 + 2a − 35 against the a in a − 5.
Those a pieces belong to sums. They are not one factor multiplying the whole top and bottom.
✓ Instead: Factor the top as (a + 7)(a − 5), then cancel the whole factor a − 5 for a ≠ 5.
✗ Not this: The simplified expression a + 7 lets a = 5 have rate 12.
The original quotient at a = 5 has input difference 0 and is undefined.
✓ Instead: Keep a ≠ 5 even after the denominator disappears.
✗ Not this: Any expression that is 0 at a = 5 automatically has a visible factor a − 5.
A zero value alone is not a factoring calculation for every kind of function. Here the product must be shown or justified for the particular expression.
✓ Instead: Multiply (a + 7)(a − 5) back out to prove this example's polynomial factorization.
✗ Not this: In every lesson the letter a is always the left endpoint.
Letters name different jobs in different questions. In [5, a] with a > 5, a is the right endpoint.
✓ Instead: Identify the fixed input and the changing endpoint before computing.

Tips and tricks

  • If the problem gives a rate and asks for an endpoint, build the rate expression first and solve second.
  • Multiply proposed factors back out; this checks the numerator before cancellation.
  • Keep the original excluded input beside every final expression.
  • Write interval endpoints in increasing order. Reversing both subtractions preserves the rate, but reversing only one changes its sign.

7 The trap

Canceling pieces of sums instead of whole multiplied factors, or forgetting the excluded endpoint after cancellation.

Lesson 5 of 10

The difference quotient: average rate of change from x to x + h

Subject: MathematicsLeans on: Distributive propertyLeans on: Subtracting a whole expressionLeans on: Like termsLeans on: Canceling a nonzero factorOpen this lesson
Leans on 0Comes back in 0

1 In plain words

The difference quotient has two ordinary actions hidden in its name. A difference is the result of subtracting: how much the output changed. A quotient is the result of dividing: share that change over the input change. Picture starting at a road marker and walking a chosen distance. Call your starting input x and your signed step h. A positive h goes right on the number line; a negative h goes left. Your ending input is x + h. The difference quotient measures average climb per input unit over that trip. It is the same average rate you already know, written so one answer can cover many starts and steps.

2 Picture it

x + h6 − 2 · inputinputoutput6 − 2(x + h)
The whole expression x + h replaces the input of g.

Reminder

Reminder: Distributive property, Subtracting a whole expression, Like terms, Canceling a nonzero factor
  • Distributive property. A number outside parentheses multiplies each term: −2(x + h) = −2x − 2h.
  • Subtracting a whole expression. The minus means multiply by −1: −(6 − 2x) = −6 + 2x.
  • Like terms. Like terms have the same letters with the same powers: −2x + 2x = 0. Terms x2 and x cannot be combined.
  • Canceling a nonzero factor. −2hh = −2 · hh = −2 only for h ≠ 0.

3 Why

Put starting input x and ending input x + h into the endpoint formula. The bottom is (x + h) − x = h, and the top is f(x + h) − f(x). That is why the formula has this form. For an expanded polynomial, f(x + h) contains a copy of f(x); subtracting removes that copy, leaving terms with h. For g(x) = 6 − 2x, that is 6 − 2x − 2h minus 6 − 2x. Other kinds of functions may require combining fractions first, rather than showing cancellation term by term.

4 The rule

Difference quotient =f(x+h)−f(x)h, with h≠0 and both outputs defined.

It is the average rate on [x,x+h] if h > 0, and on the sorted interval [x+h,x] if h < 0.

The same idea, five ways

Say it

Say 'difference in outputs divided by the input step'. Read f(x + h) as 'f of x plus h'.

Write it

Start at input x, change the input by a nonzero signed amount h, and compare the two outputs per input unit.

In math
  • f(x+h)−f(x)h,h≠0
  • h=(x+h)−x
  • [x, x + h] if h > 0; [x + h, x] if h < 0
Like

Average climb per mile between a chosen road marker and another marker a signed step away.

See it
0123456+3lands on 5
The starting input 2 plus step 3 gives ending input 5.

The same idea, other ways

With numbers first

For x2 starting at 3, a step 1 gives 16−91 = 7. A step 0.5 gives 12.25−90.5 = 6.5. Steps 0.1 and 0.01 give 6.1 and 6.01. The table shows every rate is 6 + h. The fully worked first rung computes every column.

input houtput average rate170.56.50.16.10.016.01↓ evaluate: input given, read the output below it
h: 1, 0.5, 0.1, 0.01; average rate: 7, 6.5, 6.1, 6.01.
A line before its algebra

For g(x) = 6 − 2x, start at 1 and step 1: 2−41 = −2. Start at 3 and step 2: −4−02 = −2. Start at −2 and step 0.5: 9−100.5 = −2. All give −2; the worked subtraction shows why.

(x, g(x))(x + h, g(x + h))check startcheck endrun 1rise −224−448
y=−2x+6.
What it is for

If d(t) is miles driven after t hours, d(t+h)−d(t)h gives the average speed across those times when h > 0. For positive s and h on a growing square tile, 2sh + h2 is the extra area and 2s + h is extra area per inch of growth. At s = 10 and h = 1, that rate is 21. A reusable answer also covers many intervals: 2x + h at x = 3, h = 0.5 gives 6.5.

s²shsshh²hsh
An area model with sides s + h and s + h.
The same function machine

In f(x + h), f names the machine and x + h is its one input. It is not f multiplied by x + h. For a square machine, put in 3 + 1 = 4 and get 16, rather than adding 1 to the old output 9.

x + hsquare inputinputoutput(x + h)²
The parentheses contain one input for the named function.
A smaller observation window

The square table's rates 7, 6.5, 6.1, and 6.01 move toward 6 as the positive step shrinks. For smoothly changing travel, shrinking the time window makes an average speed approach the speedometer reading at the starting time. Here the immediate task is building reusable interval rates; no zero step is ever divided by.

input houtput average rate170.56.50.16.10.016.01↓ evaluate: input given, read the output below it
The same picture again: h: 1, 0.5, 0.1, 0.01; average rate: 7, 6.5, 6.1, 6.01.

Part 1: When the start is a number

The letter x only names the starting input. Replacing it by 4 gives an interval between 4 and 4 + h; the method is unchanged.

  • For h > 0, the interval is [4, 4 + h]. For h < 0, it is [4 + h, 4].
  • The input difference in starting-to-ending order is always h.
(4, 19)(6, 29)run 1rise 5461620242832
Choosing the starting input 4 leaves the same average line rate 5.

Worked example

Difference quotient starting at x = 3

Let f(x) = x2 − 4x + 1. (a) Find and simplify the difference quotient f(3+h)−f(3)h, with h ≠ 0. (b) Use your result with h = −1 to find the average rate of change of f on the matching interval, and name that interval.

Show the steps and the answer
(3, −2)(2, −3)−2246−4−2246
Graph of f(x) = x2 − 4x + 1 = (x − 2)2 − 3, marking the start point (3, −2) and the point (2, −3) reached…
  1. Find the fixed output first: f(3)=32−4(3)+1=9−12+1=−2.The start is the number 3, so f(3) is a single number. Computing it once keeps the subtraction simple.
  2. Plug in: f(3+h)=(3+h)2−4(3+h)+1=(9+6h+h2)−12−4h+1=h2+2h−2.Every x is replaced by the whole input (3 + h). The square (3 + h)2 is multiplied out to 9 + 6h + h2, and −4 is distributed over both terms of (3 + h).
  3. Subtract: f(3+h)−f(3)=(h2+2h−2)−(−2)=h2+2h−2+2.Writing −(f(3)) in parentheses lets the minus sign reach the whole value. Subtracting −2 means adding 2.
  4. Simplify: h2+2h−2+2=h2+2h.The terms that do not depend on h, −2 and +2, cancel. This always happens for a fully expanded polynomial.
  5. Cancel h: h2+2hh=h(h+2)h=h+2, for h≠0.Factor h shows that h multiplies the entire numerator. Since h ≠ 0, we may divide it out. The restriction h ≠ 0 stays.
  6. Part (b): put h=−1 into h+2 to get −1+2=1. The interval runs from 3+(−1)=2 to 3. Since h < 0, write it in sorted order as [2, 3].When h is negative, x + h lies to the left of x. The average rate therefore belongs to the interval [x + h, x].
Answer
(a) f(3+h)−f(3)h = h + 2, h ≠ 0. (b) With h = −1, the average rate of change is 1 on the interval [2, 3].
Double-check Compute directly: f(2) = 4 − 8 + 1 = −3 and f(3) = −2. The average rate on [2, 3] is −2−(−3)3−2 = 11 = 1, which matches h + 2 at h = −1. A second check with h = 1: f(4) = 16 − 16 + 1 = 1, so 1−(−2)1 = 3, and h + 2 = 3. ✓

Work to write

  1. f(3)=9−12+1=−2
  2. f(3+h)=(3+h)2−4(3+h)+1=h2+2h−2
  3. f(3+h)−f(3)=(h2+2h−2)−(−2)=h2+2h
  4. h2+2hh=h(h+2)h=h+2,h≠0
  5. h = −1: h + 2 = 1; interval [2, 3] since h < 0

Part 2: When the start is 3 for a square

Use the same replacement in a square formula. Multiplying out creates the output change 6h + h2. Factoring h turns this into the reusable rate 6 + h.

  • The rate is 6 + h, with h ≠ 0.
  • This is 2x + h with the starting input set to x = 3.
  • Section 1.1 used a and a + h; here the general start is named x.
x²xhxxhh²hxh
For positive lengths x and h, the four areas explain the square expansion.

Worked example

Difference quotient of a quadratic starting at x = 2

Let f(x) = 2x2 + 4x − 2. (a) Find and simplify the difference quotient f(2+h)−f(2)h, with h ≠ 0. (b) Use your result with h = 12 to find the average rate of change of f on the matching interval, and name that interval.

Show the steps and the answer
(2, 14)(2.5, 20.5)−4−224−66121824
The parabola f(x) = 2x2 + 4x − 2 = 2(x + 1)2 − 4, with the points (2, 14) and (2.5, 20.5).
  1. Find f(2) first: f(2)=2(2)2+4(2)−2=8+8−2=14.The fixed output f(2) is needed for the subtraction, and the start is a plain number.
  2. Plug in 2+h: f(2+h)=2(2+h)2+4(2+h)−2. Expand (2+h)2=4+4h+h2, so f(2+h)=8+8h+2h2+8+4h−2=2h2+12h+14.Every x is replaced by the whole input (2 + h). The square must be multiplied out before it is doubled.
  3. Subtract: f(2+h)−f(2)=(2h2+12h+14)−(14)=2h2+12h.The parentheses let the minus reach the whole of f(2). The constant terms, which do not depend on h, cancel.
  4. Factor and cancel h: 2h2+12hh=h(2h+12)h=2h+12, for h≠0.h multiplies the entire numerator, so the nonzero factor h divides out. The restriction h ≠ 0 stays.
  5. Part (b): with h=12, the rate is 2(12)+12=13. Because h > 0, the interval is [2,2+12]=[2,52].The difference quotient is the average rate on [x, x + h] when h > 0.
Answer
(a) f(2+h)−f(2)h = 2h + 12, h ≠ 0. (b) The average rate of change is 13 on the interval [2, 52].
Double-check Direct computation gives f(52) = 2(254) + 4(52) − 2 = 252 + 10 − 2 = 412. Then f(52)−f(2)52−2 = 412−1412 = 13212 = 13, which matches 2h + 12 at h = 12.

Work to write

  1. f(2) = 14
  2. f(2+h)=2(4+4h+h2)+4(2+h)−2=2h2+12h+14
  3. f(2+h)−f(2)=(2h2+12h+14)−(14)=2h2+12h
  4. h(2h+12)h=2h+12,h≠0
  5. h = 12: average rate = 13 on [2, 52]

Part 3: A negative h goes left

h is a signed input change, not necessarily a positive length. The ending input is still start plus h. Keep the output subtraction in that same travel order; sorting the interval reverses both differences and preserves the ratio.

  • h < 0 means the ending input is smaller than the starting input.
  • For x = 4 and h = −3, the endpoints are 4 and 1, so the sorted interval is [1, 4].
012345−3lands on 1
A signed step h = −3 goes left from 4 to 1; the sorted interval is [1, 4].

Worked example

Difference quotient of a downward parabola, then a step backward with h = −3

Let f(x) = −x2 + 6x − 5. (a) Find and simplify the difference quotient f(x+h)−f(x)h, with h ≠ 0. (b) Use your result with x = 5 and h = −3 to find the average rate of change of f on the matching interval, and name that interval.

Show the steps and the answer
(2, 3)(5, 0)246−4−2246
Graph of f(x) = −x2 + 6x − 5 = −(x − 3)2 + 4, with vertex (3, 4).
  1. f(x+h)=−(x+h)2+6(x+h)−5=−(x2+2xh +h2) +6x+6h−5=−x2−2xh −h2+6x+6h−5Plug in: every x is replaced by the whole input (x + h). The square is expanded before the leading minus is applied to all three of its terms.
  2. f(x+h)−(f(x))= (−x2−2xh −h2+6x+6h−5) −(−x2+6x−5)=−x2−2xh −h2+6x+6h−5+x2−6x+5Subtract: f(x) goes in parentheses so the minus reaches every term. −x2 becomes +x2, +6x becomes −6x, and −5 becomes +5.
  3. Numerator =−2xh −h2+6hSimplify: −x2 + x2 = 0, 6x − 6x = 0 and −5 + 5 = 0. The terms independent of h cancel, as they must for a fully expanded polynomial.
  4. −2xh−h2+6hh=h(−2x−h+6)h=−2x−h+6, with h≠0Cancel h: h is a factor of the entire numerator and h ≠ 0, so it divides out. The restriction h ≠ 0 stays.
  5. With x = 5 and h = −3: −2(5)−(−3)+6=−10+3+6=−1Substitute the given values into the simplified quotient. Subtracting a negative h adds 3.
  6. x+h=5+(−3)=2, so the interval is [2, 5]Since h < 0, the step goes left from x = 5 to x + h = 2. The interval is written sorted, from the smaller endpoint to the larger.
Answer
(a) f(x+h)−f(x)h = −2x − h + 6, for h ≠ 0. (b) The average rate of change of f on [2, 5] is −1.
Double-check Compute directly: f(5) = −25 + 30 − 5 = 0 and f(2) = −4 + 12 − 5 = 3. Then f(5)−f(2)5−2 = 0−33 = −1, which matches. The original form gives the same value: f(2)−f(5)−3 = 3−0−3 = −1.

Work to write

  1. f(x + h) = −x2 − 2xh − h2 + 6x + 6h − 5
  2. f(x + h) − f(x) = −2xh − h2 + 6h
  3. f(x+h)−f(x)h=−2x−h+6,h≠0
  4. x=5,h=−3:−10+3+6=−1
  5. Interval [2, 5]; average rate of change = −1

Part 4: An area-growth use

A difference quotient compares the growth of a square's area with the growth of its side. The picture splits the added area into two strips and a corner when the lengths and growth are positive.

  • A square of side s has area s2.
  • Growing the side by h adds 2sh + h2 square units.
  • For s > 0 and h > 0, area gain per side-length gain is 2s + h.
s²shsshh²hsh
The same picture again: An area model with sides s + h and s + h.

Worked example

What the difference quotient measures on a tile

This asks how much area is gained per inch of side growth. A square tile's side grows from 10 inches to 11 inches. Find the average area gain per inch.

Show the steps and the answer
s²shsshh²hsh
The same picture a third time: An area model with sides s + h and s + h.
  1. Starting area: 102=100 square inches. Ending area: 112=121 square inches.A square's area is its side times itself.
  2. Area change: 121−100=21 square inches. Side change: 11−10=1 inch.Subtract matching endpoints in the same order.
  3. 211=21 square inches per inch.The difference quotient divides the area difference by the side-length difference.
Answer
21 square inches per inch of side growth.
Double-check The two strips add 10 · 1 + 1 · 10 = 20 square inches, and the corner adds 12 = 1, giving 21. The general result 2s + h gives 2(10) + 1 = 21.

5 How

  1. Plug in: replace every x in the formula by the whole input (x + h); multiply out the needed products.
  2. Subtract: write −(f(x)) with parentheses, then let that minus reach every term.
  3. Simplify: combine like terms; for a fully expanded polynomial, the terms independent of h cancel.
  4. Cancel h: when h multiplies the entire numerator, divide that nonzero factor out; keep every original restriction.

The difference quotient in four moves

1
Is the starting input a fixed number instead of x?
YesUse that number in f(start + h) and f(start); the bottom is still h.
NoKeep x as the unspecified starting input.
↓
2
Is there a squared input?
YesMultiply (x + h)(x + h) into x2 + 2xh + h2 before applying any outside coefficient.
NoCheck whether another power or a fraction needs work.
↓
3
Is there a cubed input?
YesUse the cubing-a-sum refresher to get x3 + 3x2h + 3xh2 + h3.
NoContinue with the given formula.
↓
4
Is the input in a fraction bottom?
YesKeep each whole bottom grouped, find a common bottom, and subtract the tops before dividing by h.
NoCombine polynomial like terms directly.
↓
5
After full polynomial expansion, does a term independent of h remain?
YesRecheck substitution and signs. This check applies to expanded polynomials; other functions can need different simplification.
NoFactor h from the whole numerator if possible.
↓
6
Is h negative?
YesKeep its sign in the quotient and describe the sorted interval [x + h, x].
NoFor h > 0 use [x, x + h]; h = 0 is excluded.
  1. Plug in the whole input x + h.
  2. Subtract the whole starting output f(x).
  3. Simplify using the right arithmetic for the formula.
  4. Cancel only a whole nonzero common factor, and state the restrictions.

6 Worked example

Worked example

Difference quotient of a line

This asks for the average rate over any nonzero input step h starting at any input x. Find g(x+h)−g(x)h for g(x) = 6 − 2x. The plan is plug in, subtract, simplify, then cancel h.

Show the steps and the answer
(x, g(x))(x + h, g(x + h))check startcheck endrun 1rise −224−448
The same picture again: y = −2x + 6.
  1. Require h≠0. Write g(x+h)=6−2(x+h).x + h is one whole input. The −2 must multiply all of it; writing 6 − 2x + h would multiply only x.
  2. g(x+h)=6−2x−2h.The distributive property means the multiplier reaches each term: (−2) · x = −2x and (−2) · h = −2h.
  3. Write the output change as (6−2x−2h)−(6−2x).Subtract the entire starting output, not one selected term.
  4. −(6−2x)=−6+2x, so the output change is 6−2x−2h−6+2x.The hidden sign on 6 is +. Multiplying the whole starting expression by −1 changes +6 to −6 and −2x to +2x.
  5. 6−2x−2h−6+2x=−2h.The +6 and −6 cancel, and −2x and +2x cancel. The expanded polynomial contains a copy of g(x), removed by subtraction.
  6. −2hh=−2·hh=−2·1=−2.The nonzero factor h divides by itself to give 1.
Answer
  • −2, with h ≠ 0.
  • For each 1 unit right, g drops 2 output units. This rate holds for both positive and negative signed steps.
Double-check Use x = 2 and h = 3: g(2) = 6 − 4 = 2 and g(5) = 6 − 10 = −4, so −4−23 = −63 = −2. Also g(0) = 6 and g(1) = 4 show a drop of 2 for a move of 1 right.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 6

Difference quotient with numbers: f(x) = x2 − 3x + 5 at x = 1, h = 3

Let f(x) = x2 − 3x + 5. Use x = 1 and h = 3, so x + h = 4. The table lists f(1) = 3 and f(4) = 9. Find the difference quotient f(x+h)−f(x)h and state the interval it is the average rate of change on.

input xoutput f(x) = x² − 3x + 51349
x:1,4;f(x)=x2−3x+5:3,9.
Show the steps and the answer
input xoutput f(x) = x² − 3x + 51349↓ evaluate: input given, read the output below it
Values of f(x) = x2 − 3x + 5 at x = 1 and x + h = 4.
  1. Plug in: f(1+3)=(1+3)2−3(1+3)+5. Multiply out: (1+3)2=1+2·1·3+32=1+6+9, and −3(1+3)=−3−9. So f(1+3)=1+6+9−3−9+5=9.Every x in the formula is replaced by the whole input (x + h) = (1 + 3). Expanding before adding shows which pieces come from h. The total matches the table value f(4) = 9.
  2. Subtract: f(1+3)−(f(1))=(1+6+9−3−9+5)−(1−3+5)=1+6+9−3−9+5−1+3−5.Writing −(f(x)) with parentheses lets the minus sign reach every term of f(1) = 1 − 3 + 5.
  3. Simplify: the pairs 1−1,−3+3 and 5−5 cancel, leaving 6+9−9=6. So the numerator is 6. As a check, 9−3=6.The terms that do not involve h (the pieces of f(1)) cancel. Only terms built from h = 3 remain.
  4. Cancel h: 63=2. Here h=3≠0, and both f(1) and f(4) are defined.Dividing by h needs h ≠ 0. Since h > 0, the interval is [x, x + h] = [1, 4].
Answer
f(4)−f(1)3 = 9−33 = 2. This is the average rate of change of f on [1, 4].
Double-check The general difference quotient is (x+h)2−3(x+h)+5−(x2−3x+5)h = 2xh+h2−3hh = 2x + h − 3 for h ≠ 0. At x = 1, h = 3 this gives 2 + 3 − 3 = 2, which matches. The slope of the secant line through (1, 3) and (4, 9) is 9−34−1 = 2.

Work to write

  1. x + h = 1 + 3 = 4
  2. f(4) = 16 − 12 + 5 = 9 and f(1) = 1 − 3 + 5 = 3
  3. f(4)−(f(1))=9−3=6
  4. 63 = 2, with h = 3 ≠ 0
  5. Average rate of change on [1, 4] is 2

Rung 2 of 6

Difference quotient of the constant function f(x) = 4

Let f(x) = 4 for every real number x. Its graph is the horizontal line y = 4, which passes through (−2, 4) and (3, 4). (a) Find and simplify the difference quotient f(x+h)−f(x)h, with h ≠ 0. (b) Use your result with x = 3 and h = −5, so x + h = −2. Find the average rate of change of f on the matching interval, and name that interval.

(−2, 4)(3, 4)−6−4−2246−2246
The horizontal line y = 4, the graph of the constant function f(x) = 4, through (−2, 4) and (3, 4).
Show the steps and the answer
  1. Plug in: f(x+h)=4.The rule for f has no x in it. Replacing every x by (x + h) changes nothing, so the output is 4 for any input.
  2. Subtract with parentheses: f(x+h)−(f(x))=4−(4).Writing −(f(x)) with parentheses makes the minus sign reach the whole output f(x) = 4.
  3. Simplify the numerator: 4−4=0, so the difference quotient is 0h.The two outputs are equal, so they cancel completely. Every term is independent of h, so nothing is left.
  4. Divide: 0h=0 for h≠0.Zero divided by any nonzero number is 0. The restriction h ≠ 0 stays, because the quotient is undefined at h = 0.
  5. Part (b): with x=3 and h=−5, the difference quotient is 0. Since h < 0, sort the endpoints x+h=−2 and x=3 to get the interval [−2,3].The rule says that when h < 0, the quotient is the average rate on the sorted interval [x + h, x].
Answer
(a) f(x+h)−f(x)h = 0 for all h ≠ 0. (b) The average rate of change of f on [−2, 3] is 0.
Double-check Compute directly from the two points: f(−2)−f(3)−2−3 = 4−4−5 = 0−5 = 0. This matches part (b). It also agrees with the graph, because a horizontal line has slope 0 and the output never changes.

Work to write

  1. f(x + h) = 4
  2. f(x+h)−(f(x))=4−(4)=0
  3. f(x+h)−f(x)h=0h=0,h≠0
  4. h = −5 < 0, so the interval is [−2, 3]
  5. Average rate of change on [−2, 3] = 0

Rung 3 of 6

Multiplication changes the whole input

This asks for a rate formula for every start and every nonzero step. Find the difference quotient of f(x) = 4x.

Show the steps and the answer
startendrun 1rise 42462610141822
This line rises 4 for each unit right.
  1. f(x+h)=4(x+h)=4x+4h.The whole new input replaces x, and 4 multiplies each piece.
  2. Output change: (4x+4h)−(4x)=4x+4h−4x=4h.Subtract the entire original output; +4x and −4x cancel.
  3. 4hh=4·hh=4, with h≠0.A common nonzero multiplier h divides out.
Answer
4, with h ≠ 0.
Double-check At x = 2 and h = 3, f(5) = 20 and f(2) = 8, so 20−83 = 123 = 4.

Rung 4 of 6

A negative slope with a constant term

This asks for the average rate over any nonzero input step h starting at any input x. Find g(x+h)−g(x)h for g(x) = 6 − 2x. The plan is plug in, subtract, simplify, then cancel h.

Show the steps and the answer
(x, g(x))(x + h, g(x + h))check startcheck endrun 1rise −224−448
The same picture a third time: y = −2x + 6.
  1. Require h≠0. Write g(x+h)=6−2(x+h).x + h is one whole input. The −2 must multiply all of it; writing 6 − 2x + h would multiply only x.
  2. g(x+h)=6−2x−2h.The distributive property means the multiplier reaches each term: (−2) · x = −2x and (−2) · h = −2h.
  3. Write the output change as (6−2x−2h)−(6−2x).Subtract the entire starting output, not one selected term.
  4. −(6−2x)=−6+2x, so the output change is 6−2x−2h−6+2x.The hidden sign on 6 is +. Multiplying the whole starting expression by −1 changes +6 to −6 and −2x to +2x.
  5. 6−2x−2h−6+2x=−2h.The +6 and −6 cancel, and −2x and +2x cancel. The expanded polynomial contains a copy of g(x), removed by subtraction.
  6. −2hh=−2·hh=−2·1=−2.The nonzero factor h divides by itself to give 1.
Answer
  • −2, with h ≠ 0.
  • For each 1 unit right, g drops 2 output units. This rate holds for both positive and negative signed steps.
Double-check Use x = 2 and h = 3: g(2) = 6 − 4 = 2 and g(5) = 6 − 10 = −4, so −4−23 = −63 = −2. Also g(0) = 6 and g(1) = 4 show a drop of 2 for a move of 1 right.

Rung 5 of 6

Difference quotient of a downward parabola starting at x = 1

Let f(x) = −x2 − 2x + 5. Its graph is the downward-opening parabola y = −(x + 1)2 + 6, with vertex (−1, 6). It passes through (1, 2) and (3, −10). (a) Find and simplify the difference quotient f(1+h)−f(1)h, with h ≠ 0. (b) Use your result with h = 2 to find the average rate of change of f on the matching interval, and name that interval.

(1, 2)(3, −10)−4−224−12−8−448
The parabola y = −(x + 1)2 + 6 with vertex (−1, 6).
Show the steps and the answer
  1. Find the starting output: f(1)=−(1)2−2(1)+5=−1−2+5=2.The difference quotient needs f(1), so the fixed start x = 1 is evaluated first. The result matches the point (1, 2) on the graph.
  2. Plug in 1 + h: f(1+h)=−(1+h)2−2(1+h)+5=−(1+2h+h2)−2−2h+5=−1−2h−h2−2−2h+5=2−4h−h2.Every x is replaced by the whole input (1 + h). The square is expanded before the leading minus sign is distributed.
  3. Subtract: f(1+h)−f(1)=(2−4h−h2)−(2)=2−4h−h2−2.Writing −(f(1)) with parentheses makes the minus sign reach the whole starting output.
  4. Simplify the numerator: 2−2−4h−h2=−4h−h2.The terms that do not depend on h cancel, as they must for a fully expanded polynomial.
  5. Factor and cancel h: −4h−h2h=h(−4−h)h=−4−h, for h≠0.h multiplies the entire numerator. It is a nonzero factor, so it can be divided out. The restriction h ≠ 0 stays attached to the result.
  6. For (b), substitute h=2: −4−2=−6. The inputs are 1 and 1+2=3, so the interval is [1, 3].With h > 0, the difference quotient is the average rate of change on [x, x + h] = [1, 3].
Answer
(a) f(1+h)−f(1)h = −4 − h, for h ≠ 0. (b) With h = 2, the average rate of change of f on [1, 3] is −6. On average, f falls 6 units for each 1-unit increase in x.
Double-check Compute directly from the given points: f(3) = −9 − 6 + 5 = −10 and f(1) = 2. Then −10−23−1 = −122 = −6, which matches −4 − 2 = −6.

Work to write

  1. f(1) = 2
  2. f(1+h)=−(1+h)2−2(1+h)+5=2−4h−h2
  3. f(1+h)−f(1)=(2−4h−h2)−(2)=−4h−h2
  4. −4h−h2h=h(−4−h)h=−4−h,h≠0
  5. h=2:−4−2=−6
  6. Interval: [1, 3]; average rate of change = −6

Rung 6 of 6

Difference quotient with a negative step on y = x2 + 2x − 3

Let f(x) = x2 + 2x − 3. Its graph is the upward-opening parabola y = (x + 1)2 − 4, with vertex (−1, −4). It passes through (−1, −4) and (2, 5). (a) Find and simplify the difference quotient f(x+h)−f(x)h, with h ≠ 0. (b) Use your result with x = 2 and h = −3 to find the average rate of change of f on the matching interval, and name that interval.

(−1, −4)(2, 5)−4−224−44
The parabola y = x2 + 2x − 3 = (x + 1)2 − 4, with vertex (−1, −4) and the point (2, 5).
Show the steps and the answer
  1. f(x+h)=(x+h)2+2(x+h)−3=x2+2xh +h2+2x+2h−3Plug in: every x becomes the whole input (x + h). Then (x + h)2 = x2 + 2xh + h2, and 2(x + h) = 2x + 2h.
  2. f(x+h)−f(x)=x2+2xh +h2+2x+2h−3−(x2+2x−3)=x2+2xh +h2+2x+2h−3−x2−2x+3Subtract: f(x) goes in parentheses, so the minus reaches every one of its terms. This turns −3 into +3.
  3. f(x+h)−f(x)=2xh +h2+2hSimplify: x2 − x2 = 0, 2x − 2x = 0 and −3 + 3 = 0. The terms independent of h cancel, as they must for a fully expanded polynomial.
  4. 2xh+h2+2hh=h(2x+h+2)h=2x+h+2, with h≠0Cancel h: h is a factor of the entire numerator and h ≠ 0, so it divides out. The restriction h ≠ 0 still holds.
  5. With x = 2 and h = −3: 2(2)+(−3)+2=4−3+2=3Substitute into the simplified quotient. Here −3 ≠ 0, so this value of h is allowed.
  6. x+h=2+(−3)=−1, so the interval is [−1,2]The step h is negative, so x + h lies to the left of x. Sorting the endpoints gives [x + h, x] = [−1, 2].
Answer
(a) f(x+h)−f(x)h = 2x + h + 2, for h ≠ 0. (b) The average rate of change of f on [−1, 2] is 3.
Double-check Use the graph's points directly. f(2) = 4 + 4 − 3 = 5 and f(−1) = 1 − 2 − 3 = −4. Then f(2+(−3))−f(2)−3 = −4−5−3 = −9−3 = 3. In left-to-right order, 5−(−4)2−(−1) = 93 = 3. Both match the formula's value.

Work to write

  1. f(x + h) = x2 + 2xh + h2 + 2x + 2h − 3
  2. f(x + h) − f(x) = x2 + 2xh + h2 + 2x + 2h − 3 − (x2 + 2x − 3) = 2xh + h2 + 2h
  3. 2xh+h2+2hh = 2x + h + 2, h ≠ 0
  4. x=2,h=−3:2(2)+(−3)+2=3
  5. x + h = −1, so the interval is [−1, 2]
  6. Average rate of change on [−1, 2] = 3

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: f(x + h) = f · x + f · h.
f is the function's name, not a number multiplying its input.
✓ Instead: For f(x) = x2, f(x + h) = (x + h)2.
✗ Not this: f(x + h) = f(x) + f(h).
With f(x) = x2, x = 3 and h = 1, f(4) = 16, but f(3) + f(1) = 9 + 1 = 10.
✓ Instead: Feed the whole new input x + h into the formula.
✗ Not this: f(x + h) = f(x) + h.
For the same square example, f(4) = 16, but f(3) + 1 = 10. Adding to an input happens before the rule runs.
✓ Instead: For g(x) = 6 − 2x, g(x + h) = 6 − 2x − 2h.
✗ Not this: A leftover 12 − 2h is the correct output change for g(x) = 6 − 2x.
The starting +6 was added rather than subtracted. The proper subtraction removes the copy 6 − 2x.
✓ Instead: (6−2x−2h)−(6−2x)=−2h.
✗ Not this: h must always be positive, and [x, x + h] is ordered for every h.
If x = 4 and h = −3, x + h = 1 lies to the left of 4.
✓ Instead: The quotient allows h < 0; use the sorted interval [1, 4]. Only h = 0 is always forbidden.

Tips and tricks

  • Remember the name: difference means subtraction, quotient means division.
  • Here x names the starting input, not an unknown you automatically solve for. h names the signed input change.
  • The four moves are plug in, subtract, simplify, cancel a valid whole factor.
  • For a fully expanded polynomial numerator, every term independent of h cancels. Do not apply that wording blindly to square roots or other unexpanded functions.
  • Check algebra with x = 2 and h = 3; values 0 and 1 often hide missing factors and wrong powers.
  • The assigned algebra skill is an interval ending at a letter or having step h. This is the same endpoint-rate idea from section 1.1.

7 The trap

Adding h to the old output, treating f as a multiplier, or allowing a zero input step. Substitute x + h into the whole function formula and keep h ≠ 0.

Lesson 6 of 10

Difference quotient of a quadratic

Subject: MathematicsLeans on: Dividing powersLeans on: Distributing a multiplierLeans on: Subtracting a whole expressionOpen this lesson
Leans on 0Comes back in 0

1 In plain words

A quadratic polynomial uses constants, multiples of the input, and a nonzero multiple of the squared input. Think of enlarging a square carpet: the old square stays in one corner, two strips appear beside it, and a corner fills the gap. Multiplying out (x + h)2 records all four pieces: x2, xh, xh, and h2. That is why x2 + h2 loses something. Subtracting the old output removes the old square, leaving the pieces due to the input change. If every remaining term contains h, write each as h times something and divide out that common multiplier. A number in front multiplies every piece. The final rate can depend on where you start and how far you step.

2 Picture it

x²xhxxhh²hxh
An area model with sides x + h and x + h.

Reminder

Reminder: Dividing powers, Distributing a multiplier, Subtracting a whole expression
  • Dividing powers. h2 ÷ h = (h · h) ÷ h = h for h ≠ 0. One of the two h factors is removed.
  • Distributing a multiplier. 3(x2 + 2xh + h2) = 3x2 + 6xh + 3h2. The 3 reaches every term.
  • Subtracting a whole expression. −(3x2 − 4x + 1) = −3x2 + 4x − 1. The outside −1 changes all three signs.

3 Why

Multiplying (x + h)(x + h) gives x2 + xh + hx + h2 = x2 + 2xh + h2. Subtracting x2 leaves h(2x + h), so dividing by h gives 2x + h. For a fully expanded polynomial, the terms independent of h reproduce the original f(x) and cancel when it is subtracted. This is a polynomial check, not a universal description of all functions. The rate varies on a curve: for 3x2 − 4x + 1, steps of 1 beginning at 1 and at 2 give different rates, 5 and 11.

4 The rule

(x+h)2=x2+2xh +h2.

For f(x)=x2, the difference quotient is 2x+h, with h≠0. Multiply any outside coefficient into every expanded term, subtract all of f(x), then factor the whole numerator before canceling h.

The same idea, five ways

Say it

Say 'multiply out the square, subtract the old output, factor the step, then divide'.

Write it

The extra square pieces produce the output change, and dividing by the input step gives the average rate.

In math
  • (x + h)2 = x2 + 2xh + h2
  • For f(x) = x2: f(x+h)−f(x)h = 2x + h, h ≠ 0
  • [x, x + h] if h > 0; [x + h, x] if h < 0
Like

A square carpet grows by two strips and a corner, then its added area is compared with the side-length change.

See it
x²xhxxhh²hxh
The same picture again: An area model with sides x + h and x + h.

The same idea, other ways

Four pieces in the square

The old x2 square, two xh strips, and h2 corner account for the entire new square. The picture describes positive lengths; the algebraic identity still holds when x or h is negative.

x²xhxxhh²hxh
The same picture a third time: An area model with sides x + h and x + h.
Follow the matching pairs

For 3x2 − 4x + 1, subtracting the original output creates +3x2 with −3x2, −4x with +4x, and +1 with −1. Each pair adds to 0, leaving 6xh + 3h2 − 4h.

+3x2 and −3x2 cancel

−4x and +4x cancel

+1 and −1 cancel

Give the letters numerical meaning

For the harder function, from 1 to 2 the rate is 5−01 = 5. From 2 to 3 it is 16−51 = 11. The same step h = 1 gives a different rate because the curve is steeper in a different place.

(1, 0)(2, 5)2261014
y = 3(x − 0.6667)2 − 0.3333, solid, beside y = 5x − 5, dashed.

Part 1: A coefficient outside the square

Multiplying a square by a number means multiplying its entire area by that number. It does not mean moving that number inside the square and squaring it too.

  • 5(x + h)2 = 5x2 + 10xh + 5h2.
  • The outside coefficient acts after the square is computed.
start x = 1end x + h = 3246183042
y = 5x2, solid, beside y = 20x − 15, dashed.

Worked example

A coefficient multiplies every square piece

This asks for the same square rate when every output is multiplied by 5. Find the difference quotient of f(x) = 5x2.

Show the steps and the answer
start x = 1end x + h = 3246183042
The same picture again: y = 5x2, solid, beside y = 20x − 15, dashed.

5(x+h)2

= 5x2 + 10xh + 5h2

rate = 10x + 5h

start x = 1end x + h = 3246183042
The same picture a third time: y = 5x2, solid, beside y = 20x − 15, dashed.
  1. f(x+h)=5(x+h)2=5(x2+2xh +h2).Multiply out the square before applying its outside coefficient 5.
  2. f(x+h)=5x2+10xh +5h2.The 5 multiplies all three terms, including the middle term.
  3. Output change: (5x2+10xh +5h2) −(5x2)=10xh +5h2.Subtract the whole original output; 5x2 cancels with −5x2.
  4. 10xh +5h2=h(10x+5h).Write 10xh = h · 10x and 5h2 = h · 5h.
  5. h(10x+5h)h=10x+5h, with h≠0.Cancel the nonzero multiplier of the whole numerator.
Answer
10x + 5h, with h ≠ 0.
Double-check With x = 1 and h = 2, f(3) = 45 and f(1) = 5. The direct rate is 45−52 = 402 = 20, and 10(1) + 5(2) = 20.

Part 2: A square plus a linear term

Each appearance of x needs the new input. The square gives 2xh + h2 after subtraction; the extra linear term gives another 2h.

  • For f(x) = x2 + 2x, rate = 2x + h + 2, with h ≠ 0.
start x = 2end x + h = 5246515253545
For x = 2 and h = 3, the dashed ruler joins outputs 8 and 35 with slope 9.

Worked example

Difference quotient of f(x) = 3x2 − 6x and a secant slope

Let f(x) = 3x2 − 6x. The graph of y = 3x2 − 6x passes through (0, 0) and (2, 0), and its lowest point is (1, −3). (a) Find and simplify the difference quotient f(x+h)−f(x)h for h ≠ 0. (b) Use your result with x = 1 and h = 1 to find the slope of the secant line through (1, −3) and (2, 0).

(1, −3)(2, 0)−224−4−22468
The parabola y = 3x2 − 6x = 3(x − 1)2 − 3 with lowest point (1, −3) and the point (2, 0).
Show the steps and the answer
  1. f(x+h)=3(x+h)2−6(x+h)=3(x2+2xh +h2) −6x−6h=3x2+6xh +3h2−6x−6hEvery x is replaced by (x + h). The square expands as x2 + 2xh + h2, and the outside multipliers 3 and −6 must reach every term inside their parentheses.
  2. f(x+h)−f(x)= (3x2+6xh +3h2−6x−6h) −(3x2−6x)=3x2+6xh +3h2−6x−6h−3x2+6xAll of f(x) is kept in parentheses after the minus, so both of its terms change sign.
  3. Combine terms: 3x2−3x2=0 and −6x+6x=0, leaving 6xh +3h2−6hThe terms that do not contain h cancel, as they always do in a difference quotient.
  4. 6xh+3h2−6hh=h(6x+3h−6)h=6x+3h−6, for h≠0Every remaining term contains h, so h factors out of the whole numerator. It cancels with the denominator only when h ≠ 0.
  5. For part (b), substitute x = 1 and h = 1: 6(1)+3(1)−6=3With x = 1 and h = 1 the difference quotient is the slope between the inputs 1 and 2, and these are the points (1, −3) and (2, 0).
Answer
(a) f(x+h)−f(x)h = 6x + 3h − 6, for h ≠ 0. (b) The secant slope through (1, −3) and (2, 0) is 3.
Double-check Compute directly: f(1) = 3 − 6 = −3 and f(2) = 12 − 12 = 0, so 0−(−3)2−1 = 3, which matches. A second check uses x = 2 and h = 1: f(3) = 27 − 18 = 9, so 9−01 = 9, and the formula gives 12 + 3 − 6 = 9.

Work to write

  1. f(x + h) = 3x2 + 6xh + 3h2 − 6x − 6h
  2. f(x + h) − f(x) = 6xh + 3h2 − 6h
  3. h(6x+3h−6)h=6x+3h−6,h≠0
  4. x=1,h=1:6+3−6=3
  5. Secant slope = 3

Part 3: A three-term quadratic

A three-term function needs every substitution and every subtraction sign. The constant term disappears in the difference, while the linear term leaves a multiple of h.

  • For f(x) = 3x2 − 4x + 1, rate = 6x + 3h − 4, with h ≠ 0.
  • 6xh = h · 6x; 3h2 = h · 3h; −4h = h · (−4).
(1, 0)(2, 5)2261014
The same picture again: y = 3(x − 0.6667)2 − 0.3333, solid, beside y = 5x − 5, dashed.

Worked example

Difference quotient with a number in front of x2

This asks for the average-rate formula for a quadratic with three terms. Find f(x+h)−f(x)h for f(x) = 3x2 − 4x + 1.

Show the steps and the answer
(1, 0)(2, 5)2261014
The same picture a third time: y = 3(x − 0.6667)2 − 0.3333, solid, beside y = 5x − 5, dashed.
  1. f(x+h)=3(x+h)2−4(x+h)+1.The new input replaces every x position, including the linear term.
  2. f(x+h)=3(x2+2xh +h2) −4x−4h+1.Multiply out the square first, then let −4 multiply both terms of x + h.
  3. f(x+h)=3x2+6xh +3h2−4x−4h+1.The coefficient 3 multiplies x2, 2xh, and h2 separately.
  4. Output change: (3x2+6xh +3h2−4x−4h+1) −(3x2−4x+1).The whole original output stays in parentheses until its minus sign is distributed.
  5. =3x2+6xh +3h2−4x−4h+1−3x2+4x−1.The outside −1 changes +3x2 to −3x2, −4x to +4x, and +1 to −1.
  6. =6xh +3h2−4h.The explicit matching pairs are +3x2 with −3x2, −4x with +4x, and +1 with −1.
  7. 6xh =h·6x,3h2=h·3h, and −4h=h·(−4). Therefore the output change is h(6x+3h−4).Each remaining term contains the same multiplier h.
  8. h(6x+3h−4)h=6x+3h−4, with h≠0.A nonzero h multiplying the whole top cancels with h on the bottom.
Answer
6x + 3h − 4, with h ≠ 0.
Double-check Use x = 2 and h = 3. f(5) = 75 − 20 + 1 = 56 and f(2) = 12 − 8 + 1 = 5, so 56−53 = 513 = 17. The result gives 6(2) + 3(3) − 4 = 17. The wrong expression 6x + 3h2 − 4 gives 35, so this check catches the power error.

Part 4: A negative squared term

The minus outside x2 negates the entire square after it has been multiplied out. Subtracting the starting negative square then adds a matching positive x2.

  • −(x + h)2 = −x2 − 2xh − h2.
  • For −x2 + 6x, rate = −2x − h + 6, with h ≠ 0.
start x = 2end x + h = 5246481216
y = −(x − 3)2 + 9, solid, beside y = −x + 10, dashed.

Worked example

A negative square coefficient

This asks for the rate formula when a minus sign is outside the square. Find the difference quotient of f(x) = −x2 + 6x.

Show the steps and the answer
start x = 2end x + h = 5246481216
The same picture again: y = −(x − 3)2 + 9, solid, beside y = −x + 10, dashed.

−(x+h)2

= −x2 − 2xh − h2

rate = −2x − h + 6

start x = 2end x + h = 5246481216
The same picture a third time: y = −(x − 3)2 + 9, solid, beside y = −x + 10, dashed.
  1. f(x+h)=−(x+h)2+6(x+h).The new input stays inside the square; the outside minus remains outside.
  2. =−(x2+2xh +h2) +6x+6h=−x2−2xh −h2+6x+6h.The outside minus reaches all three square terms, while 6 multiplies both input pieces.
  3. Output change: (−x2−2xh −h2+6x+6h) −(−x2+6x).Subtract the whole original output in parentheses.
  4. =−x2−2xh −h2+6x+6h+x2−6x.Multiplying the starting expression by −1 changes −x2 to +x2 and +6x to −6x.
  5. =−2xh −h2+6h=h(−2x−h+6).The −x2 and +x2 cancel; 6x and −6x cancel. The remaining factors after taking out h are −2x, −h, and 6.
  6. h(−2x−h+6)h=−2x−h+6, with h≠0.The entire numerator has a nonzero factor h that divides out.
Answer
−2x − h + 6, with h ≠ 0.
Double-check At x = 2 and h = 3, f(5) = −25 + 30 = 5 and f(2) = −4 + 12 = 8. The direct rate 5−83 = −1 matches −2(2) − 3 + 6 = −1.

5 How

  1. Plug in: replace every x by (x + h), multiply out the square, and apply each outside multiplier.
  2. Subtract: keep all of f(x) in parentheses after a minus, then flip every term's sign.
  3. Simplify: combine matching polynomial terms; the terms independent of h cancel.
  4. Cancel h: write every remaining term as h times something, factor h from the whole numerator, and cancel only for h ≠ 0.

The quadratic difference quotient in four moves

1
Is there a number in front of the square?
YesSquare the whole input first, then multiply every resulting term by that number.
NoUse x2 + 2xh + h2.
↓
2
Is there a minus outside the square?
YesAfter expanding, change every square term's sign.
NoKeep the signs from the square expansion.
↓
3
Have all original-output signs been reversed in the subtraction?
YesCombine the matching terms.
NoRewrite −(f(x)) before removing the parentheses.
↓
4
In this fully expanded polynomial, does a term without h remain?
YesRecheck the expansion and subtraction; the copy of f(x) should cancel.
NoWrite each term as h times something.
↓
5
Does the whole numerator now have a factor h?
YesCancel it for h ≠ 0 and interpret the average rate.
NoRecheck the algebra rather than canceling selected pieces.
  1. Plug in and multiply the square into four products.
  2. Subtract the complete original formula with parentheses.
  3. Simplify by grouping like terms.
  4. Factor a whole nonzero h and keep the original restriction.

6 Worked example

Worked example

Difference quotient for a garden plot with a compost corner

A rectangular garden plot is x meters wide and 2x meters long. A 2 square-meter corner is set aside for a compost bin, so the planted area is A(x) = 2x2 − 2 square meters. The graph of y = 2x2 − 2 passes through (2, 6) and (3, 16), and its lowest point is (0, −2). (a) Find and simplify the difference quotient A(x+h)−A(x)h for h ≠ 0. (b) Use your result with x = 2 and h = 1 to find the average rate of change of the planted area as the width grows from 2 m to 3 m.

Show the steps and the answer
(2, 6)(3, 16)(0, −2)24481216
Graph of the planted area A(x) = 2x2 − 2, with lowest point (0, −2) and the points (2, 6) and (3, 16) that…
  1. A(x+h)=2(x+h)2−2=2(x2+2xh +h2) −2=2x2+4xh +2h2−2Replace every x by (x + h), expand with (x + h)2 = x2 + 2xh + h2, and multiply the outside coefficient 2 into every expanded term.
  2. A(x+h)−A(x)= (2x2+4xh +2h2−2) −(2x2−2)=2x2+4xh +2h2−2−2x2+2Keep all of A(x) in parentheses after the minus, then flip the sign of each of its terms. The −2 becomes +2.
  3. 2x2−2x2=0 and −2+2=0, so the numerator is 4xh +2h2.Combine matching terms. The terms that do not contain h cancel.
  4. 4xh+2h2h=h(4x+2h)h=4x+2h, for h≠0Every remaining term contains h. Factor h from the whole numerator, then cancel it, which is allowed only because h ≠ 0.
  5. With x = 2 and h = 1: 4(2)+2(1)=8+2=10Substitute into the simplified difference quotient. The result is the average rate of change of A on [2, 3].
Answer
(a) A(x+h)−A(x)h = 4x + 2h, for h ≠ 0. (b) 10 square meters per meter. The planted area grows by 10 m2 for each meter of added width, on average, from 2 m to 3 m.
Double-check From the graph points, A(2) = 2(4) − 2 = 6 and A(3) = 2(9) − 2 = 16. So 16−63−2 = 10, which matches 4(2) + 2(1) = 10.

Work to write

  1. A(x + h) = 2x2 + 4xh + 2h2 − 2
  2. A(x + h) − A(x) = (2x2 + 4xh + 2h2 − 2) − (2x2 − 2) = 4xh + 2h2
  3. h(4x+2h)h=4x+2h,h≠0
  4. x = 2, h = 1: 4(2) + 2(1) = 10 m2 per m
  5. Check: 16−63−2 = 10

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 7

Difference quotient for the area of a strip of five square tiles

A decorative border is a straight strip of 5 identical square tiles placed side by side. If each tile has side length x inches, the total area of the strip is A(x) = 5x2 square inches. (a) Find and simplify the difference quotient A(x+h)−A(x)h for h ≠ 0. (b) Use your result to find the average rate of change of the area when the side length grows from x = 4 inches to x = 4.2 inches, so h = 0.2.

Show the steps and the answer
(4, 80)2461133557799
Graph of the strip's area A(x) = 5x2, marking the point (4, 80).
  1. Plug in: A(x+h)=5(x+h)2=5(x2+2xh +h2) =5x2+10xh +5h2.Every x is replaced by (x + h). The square expands as (x + h)2 = x2 + 2xh + h2, and the outside coefficient 5 must multiply every expanded term.
  2. Subtract: A(x+h)−A(x)= (5x2+10xh +5h2) −(5x2)=5x2+10xh +5h2−5x2.All of A(x) goes in parentheses after the minus sign, and the sign of each term inside is then flipped.
  3. Simplify: 5x2−5x2=0, so the numerator is 10xh +5h2.Matching terms are combined. The terms with no h cancel, and every remaining term contains h.
  4. Cancel h: 10xh+5h2h=h(10x+5h)h=10x+5h, for h≠0.Write each term as h times something and factor h from the whole numerator. The common factor h may be canceled only because h ≠ 0.
  5. Evaluate at x = 4 and h = 0.2: 10(4)+5(0.2)=40+1=41.The simplified difference quotient is the average rate of change from x to x + h, so the numbers are substituted directly.
Answer
(a) A(x+h)−A(x)h = 10x + 5h, for h ≠ 0. (b) 41 square inches of area per inch of side length.
Double-check Compute the change in area directly. A(4.2) = 5(4.2)2 = 5(17.64) = 88.2, and A(4) = 5(16) = 80. Then 88.2−800.2 = 8.20.2 = 41, which matches 10(4) + 5(0.2) = 41.

Work to write

  1. A(x + h) = 5(x + h)2 = 5x2 + 10xh + 5h2
  2. A(x + h) − A(x) = 5x2 + 10xh + 5h2 − 5x2 = 10xh + 5h2
  3. 10xh+5h2h = h(10x+5h)h = 10x + 5h, h ≠ 0
  4. At x = 4, h = 0.2: 10(4) + 5(0.2) = 41

Rung 2 of 7

A coefficient multiplies every square piece

This asks for the same square rate when every output is multiplied by 5. Find the difference quotient of f(x) = 5x2.

Show the steps and the answer
start x = 1end x + h = 3246183042
The same picture, shown 4 times now: y = 5x2, solid, beside y = 20x − 15, dashed.

5(x+h)2

= 5x2 + 10xh + 5h2

rate = 10x + 5h

start x = 1end x + h = 3246183042
The same picture, shown 5 times now: y = 5x2, solid, beside y = 20x − 15, dashed.
  1. f(x+h)=5(x+h)2=5(x2+2xh +h2).Multiply out the square before applying its outside coefficient 5.
  2. f(x+h)=5x2+10xh +5h2.The 5 multiplies all three terms, including the middle term.
  3. Output change: (5x2+10xh +5h2) −(5x2)=10xh +5h2.Subtract the whole original output; 5x2 cancels with −5x2.
  4. 10xh +5h2=h(10x+5h).Write 10xh = h · 10x and 5h2 = h · 5h.
  5. h(10x+5h)h=10x+5h, with h≠0.Cancel the nonzero multiplier of the whole numerator.
Answer
10x + 5h, with h ≠ 0.
Double-check With x = 1 and h = 2, f(3) = 45 and f(1) = 5. The direct rate is 45−52 = 402 = 20, and 10(1) + 5(2) = 20.

Rung 3 of 7

Difference quotient of f(x) = 2x2 − 4x

Let f(x) = 2x2 − 4x. The graph of y = 2x2 − 4x passes through (0, 0) and (3, 6), and its lowest point is (1, −2). (a) Find and simplify the difference quotient f(x+h)−f(x)h for h ≠ 0. (b) Use your result with x = 0 and h = 3 to find the slope of the secant line through (0, 0) and (3, 6).

(0, 0)(3, 6)(1, −2)−224−22468
Parabola y = 2x2 − 4x = 2(x − 1)2 − 2 with vertex (1, −2).
Show the steps and the answer
  1. Replace every x by (x + h): f(x+h)=2(x+h)2−4(x+h).f(x + h) means the input x + h goes into each place x appears, in both the square term and the linear term.
  2. Expand the square: (x+h)2=x2+2xh +h2, so 2(x+h)2=2x2+4xh +2h2.The outside coefficient 2 multiplies every term of the expanded square, not just the first one.
  3. Distribute the −4: −4(x+h)=−4x−4h. So f(x+h)=2x2+4xh +2h2−4x−4h.The multiplier −4 applies to both x and h.
  4. Subtract all of f(x) in parentheses: f(x+h)−f(x)= (2x2+4xh +2h2−4x−4h) −(2x2−4x)=2x2+4xh +2h2−4x−4h−2x2+4x.The minus sign applies to every term of f(x), so −4x becomes +4x.
  5. Combine like terms: 2x2−2x2=0 and −4x+4x=0, leaving 4xh +2h2−4h.The terms that do not contain h cancel. Every remaining term should contain h.
  6. Factor h from the whole numerator: 4xh +2h2−4h=h(4x+2h−4).Each term has a factor of h. Only a factor of the entire numerator can be canceled.
  7. Divide: h(4x+2h−4)h=4x+2h−4, for h≠0.h divided by h equals 1 only when h is not zero.
  8. For part (b), substitute x = 0 and h = 3: 4(0)+2(3)−4=0+6−4=2.The difference quotient with x = 0 and h = 3 is the slope from the point at x = 0 to the point at x = 0 + 3 = 3.
Answer
(a) f(x+h)−f(x)h = 4x + 2h − 4, for h ≠ 0. (b) The secant slope through (0, 0) and (3, 6) is 2.
Double-check Compute the slope directly. f(0) = 2(0)2 − 4(0) = 0 and f(3) = 2(9) − 12 = 6. Then 6−03−0 = 63 = 2, which matches 4(0) + 2(3) − 4 = 2. A second test with x = 1 and h = 1: f(2) = 8 − 8 = 0 and f(1) = 2 − 4 = −2, so 0−(−2)1 = 2. The formula gives 4(1) + 2(1) − 4 = 2. ✓

Work to write

  1. f(x + h) = 2(x + h)2 − 4(x + h) = 2x2 + 4xh + 2h2 − 4x − 4h
  2. f(x + h) − f(x) = 2x2 + 4xh + 2h2 − 4x − 4h − (2x2 − 4x) = 4xh + 2h2 − 4h
  3. =h(4x+2h−4)
  4. f(x+h)−f(x)h=4x+2h−4,h≠0
  5. x=0,h=3:4(0)+2(3)−4=2

Rung 4 of 7

A coefficient, a linear term, and a constant

This asks for the average-rate formula for a quadratic with three terms. Find f(x+h)−f(x)h for f(x) = 3x2 − 4x + 1.

Show the steps and the answer
(1, 0)(2, 5)2261014
The same picture, shown 4 times now: y = 3(x − 0.6667)2 − 0.3333, solid, beside y = 5x − 5, dashed.
  1. f(x+h)=3(x+h)2−4(x+h)+1.The new input replaces every x position, including the linear term.
  2. f(x+h)=3(x2+2xh +h2) −4x−4h+1.Multiply out the square first, then let −4 multiply both terms of x + h.
  3. f(x+h)=3x2+6xh +3h2−4x−4h+1.The coefficient 3 multiplies x2, 2xh, and h2 separately.
  4. Output change: (3x2+6xh +3h2−4x−4h+1) −(3x2−4x+1).The whole original output stays in parentheses until its minus sign is distributed.
  5. =3x2+6xh +3h2−4x−4h+1−3x2+4x−1.The outside −1 changes +3x2 to −3x2, −4x to +4x, and +1 to −1.
  6. =6xh +3h2−4h.The explicit matching pairs are +3x2 with −3x2, −4x with +4x, and +1 with −1.
  7. 6xh =h·6x,3h2=h·3h, and −4h=h·(−4). Therefore the output change is h(6x+3h−4).Each remaining term contains the same multiplier h.
  8. h(6x+3h−4)h=6x+3h−4, with h≠0.A nonzero h multiplying the whole top cancels with h on the bottom.
Answer
6x + 3h − 4, with h ≠ 0.
Double-check Use x = 2 and h = 3. f(5) = 75 − 20 + 1 = 56 and f(2) = 12 − 8 + 1 = 5, so 56−53 = 513 = 17. The result gives 6(2) + 3(3) − 4 = 17. The wrong expression 6x + 3h2 − 4 gives 35, so this check catches the power error.

Rung 5 of 7

A negative square coefficient

This asks for the rate formula when a minus sign is outside the square. Find the difference quotient of f(x) = −x2 + 6x.

Show the steps and the answer
start x = 2end x + h = 5246481216
The same picture, shown 4 times now: y = −(x − 3)2 + 9, solid, beside y = −x + 10, dashed.

−(x+h)2

= −x2 − 2xh − h2

rate = −2x − h + 6

start x = 2end x + h = 5246481216
The same picture, shown 5 times now: y = −(x − 3)2 + 9, solid, beside y = −x + 10, dashed.
  1. f(x+h)=−(x+h)2+6(x+h).The new input stays inside the square; the outside minus remains outside.
  2. =−(x2+2xh +h2) +6x+6h=−x2−2xh −h2+6x+6h.The outside minus reaches all three square terms, while 6 multiplies both input pieces.
  3. Output change: (−x2−2xh −h2+6x+6h) −(−x2+6x).Subtract the whole original output in parentheses.
  4. =−x2−2xh −h2+6x+6h+x2−6x.Multiplying the starting expression by −1 changes −x2 to +x2 and +6x to −6x.
  5. =−2xh −h2+6h=h(−2x−h+6).The −x2 and +x2 cancel; 6x and −6x cancel. The remaining factors after taking out h are −2x, −h, and 6.
  6. h(−2x−h+6)h=−2x−h+6, with h≠0.The entire numerator has a nonzero factor h that divides out.
Answer
−2x − h + 6, with h ≠ 0.
Double-check At x = 2 and h = 3, f(5) = −25 + 30 = 5 and f(2) = −4 + 12 = 8. The direct rate 5−83 = −1 matches −2(2) − 3 + 6 = −1.

Rung 6 of 7

A fresh three-term quadratic

This asks for the rate expression when a square coefficient, a negative linear coefficient, and a constant all occur. Find the difference quotient of f(x) = 6x2 − 7x + 2.

Show the steps and the answer
start x = 1end x + h = 324515253545
y = 6(x − 0.5833)2 − 0.0417, solid, beside y = 17x − 16, dashed.

change = 12xh + 6h2 − 7h

=h(12x+6h−7)

rate = 12x + 6h − 7

start x = 1end x + h = 324515253545
The same picture again: y = 6(x − 0.5833)2 − 0.0417, solid, beside y = 17x − 16, dashed.
  1. f(x+h)=6(x+h)2−7(x+h)+2.Replace every x by the whole new input.
  2. =6(x2+2xh +h2) −7x−7h+2=6x2+12xh +6h2−7x−7h+2.The square expansion is multiplied term by term by 6, and −7 reaches both input pieces.
  3. Output change: (6x2+12xh +6h2−7x−7h+2) −(6x2−7x+2).Keep all three original terms grouped for subtraction.
  4. =6x2+12xh +6h2−7x−7h+2−6x2+7x−2.The starting terms +6x2, −7x, and +2 all reverse signs.
  5. =12xh +6h2−7h=h(12x+6h−7).The matching h-independent polynomial terms cancel. Each remaining term is h times 12x, 6h, or −7.
  6. h(12x+6h−7)h=12x+6h−7, with h≠0.Cancel the nonzero multiplier of the whole numerator.
Answer
12x + 6h − 7, with h ≠ 0.
Double-check At x = 1 and h = 2, f(1) = 6 − 7 + 2 = 1 and f(3) = 54 − 21 + 2 = 35. The direct rate 35−12 = 342 = 17 matches 12(1) + 6(2) − 7 = 17.

Rung 7 of 7

Difference quotient of 4x2 with a negative step

Let f(x) = 4x2. The graph of y = 4x2 passes through (1, 4) and (2, 16). (a) Find and simplify the difference quotient f(x+h)−f(x)h for h ≠ 0. (b) Use your result with x = 2 and the negative step h = −1. This gives the slope of the secant line through (2, 16) and (1, 4).

(1, 4)(2, 16)−22481216
The parabola y = 4x2 with the points (2, 16) and (1, 4).
Show the steps and the answer
  1. Plug in: f(x+h)=4(x+h)2=4(x2+2xh +h2) =4x2+8xh +4h2.Every x is replaced by (x + h). The square expands as x2 + 2xh + h2, and the outside coefficient 4 multiplies every expanded term.
  2. Subtract: f(x+h)−f(x)= (4x2+8xh +4h2) −(4x2)=4x2+8xh +4h2−4x2.All of f(x) goes in parentheses after the minus sign, so its sign flips.
  3. Simplify: 4x2−4x2=0, so the numerator is 8xh +4h2.The terms that do not contain h cancel, which always happens in a difference quotient.
  4. Cancel h: 8xh+4h2h=h(8x+4h)h=8x+4h, for h≠0.Factor h from the whole numerator first, then cancel it. The cancellation is allowed only because h ≠ 0.
  5. Use the negative step: with x=2 and h=−1,8x+4h=8(2)+4(−1)=16−4=12.The formula holds for any nonzero h, negative values included. Here x + h = 1, so the result is the slope between x = 2 and x = 1.
Answer
(a) f(x+h)−f(x)h = 8x + 4h, h ≠ 0. (b) With x = 2 and h = −1, the value is 12.
Double-check Compute directly: f(1) = 4(1)2 = 4 and f(2) = 4(2)2 = 16. Then f(1)−f(2)−1 = 4−16−1 = −12−1 = 12. This matches 8(2) + 4(−1) = 12. As a second check, the pattern 2x + h for x2, multiplied by 4, gives 8x + 4h.

Work to write

  1. f(x + h) = 4(x + h)2 = 4x2 + 8xh + 4h2
  2. f(x + h) − f(x) = 4x2 + 8xh + 4h2 − 4x2 = 8xh + 4h2
  3. h(8x+4h)h=8x+4h,h≠0
  4. x=2,h=−1:8(2)+4(−1)=12

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: (x+h)2=x2+h2.
Multiplying the two parentheses also creates xh and hx. For x = 3 and h = 1, the wrong shortcut gives 10 instead of 16.
✓ Instead: (x + h)2 = x2 + 2xh + h2.
✗ Not this: 3(x+h)2=(3x+3h)2.
The left squares first then multiplies by 3; the right puts 3 inside the square and squares that factor too. At x = 1, h = 1, they give 12 and 36.
✓ Instead: 3(x + h)2 = 3x2 + 6xh + 3h2.
✗ Not this: 3h2÷h=3h2.
h2 = h · h, so dividing out one nonzero h leaves h. With h = 2, 12 ÷ 2 = 6, not 12.
✓ Instead: 3h2 ÷ h = 3h, for h ≠ 0.
✗ Not this: A check at x = 1 and h = 1 proves the h power is correct.
At 1, h and h2 have the same value. The wrong 6x + 3h2 − 4 can match the correct result there.
✓ Instead: Use x = 2 and h = 3: the correct result is 17, while that wrong expression gives 35.

Tips and tricks

  • Square first, multiply the outside number second. Keep those two actions on separate lines.
  • Write the matching pairs in the subtraction so each canceled term has a visible partner.
  • Meaning matters: x is where you start, and h is the signed amount the input changes.
  • A square-area picture uses positive lengths; an algebraic square identity also accepts negative values.
  • Keep h ≠ 0 beside the answer, even after h has canceled.

7 The trap

Losing the middle term, squaring an outside coefficient, or forgetting that h2 ÷ h = h. Show the full square expansion before subtracting.

Lesson 7 of 10

Difference quotient of a fraction

Subject: MathematicsLeans on: Multiplication by 1Leans on: A common denominatorLeans on: Dividing by a nonzero letterLeans on: Subtracting a grouped topOpen this lesson
Leans on 0Comes back in 0

1 In plain words

When the input sits in the bottom of a fraction, the output change becomes one fraction minus another with different bottoms. Think of converting two currencies into the same currency before subtracting: you cannot compare the counts while they use different-sized units. Give the fractions one common bottom by multiplying their original bottoms. Multiply each top by the missing bottom too, so its value stays unchanged. Then subtract only the tops, keeping the second top in parentheses. In the reciprocal examples, the new top has an h factor. Dividing by the input step means multiplying by the fraction 1 over h, and that whole factor can cancel. Keep every forbidden zero bottom.

2 Picture it

(1, 2)(2, 1)2412345(1, 2)(2, 1)1234512345
y = 2/x, solid, beside y = −x + 3, dashed.

Reminder

Reminder: Multiplication by 1, A common denominator, Dividing by a nonzero letter, Subtracting a grouped top
  • Multiplication by 1. xx = 1 only for x ≠ 0. This is why multiplying both top and bottom by x preserves a fraction.
  • A common denominator. 13 − 15 = 515 − 315 = 215. Rename fractions before subtracting.
  • Dividing by a nonzero letter. A ÷ h = A · 1h for h ≠ 0. Dividing reverses multiplication.
  • Subtracting a grouped top. 2x−2(x+h)=2x−(2x+2h)=2x−2x−2h.

3 Why

A common bottom names the same-sized pieces, so their counts can be subtracted. Multiplying top and bottom by the same nonzero expression multiplies by 1 and preserves the value. For 2x, the common-bottom top is 2x − 2(x + h) = 2x − 2x − 2h = −2h. This shows the cancellation directly, rather than assuming every kind of function must look like a polynomial. Dividing by h multiplies by 1h, because that undoes multiplication by h. The restrictions remain necessary even after a factor disappears.

4 The rule

For f(x)=2x, combine the output difference over x(x+h), then divide by h to obtain −2x(x+h).

Require h≠0,x≠0, and x+h≠0. Other fraction bottoms need their own product as the common bottom.

The same idea, five ways

Say it

Say 'make the output fractions use one bottom, subtract their tops, then divide by the step'.

Write it

A fractional difference quotient uses the usual endpoint-rate formula with an extra common-denominator step.

In math
  • For f(x) = 2x: rate = −2x(x+h)
  • h ≠ 0; x ≠ 0; x + h ≠ 0
  • For h > 0: [x, x + h]; for h < 0: [x + h, x]
Like

Convert two amounts to matching-sized units before subtracting, then share the difference over the input change.

See it
(1, 2)(2, 1)2412345
The same picture again: y = 2/x, solid, beside y = −x + 3, dashed.

The same idea, other ways

With numbers first

13 − 15 = 515 − 315 = 215. The common bottom changes the piece size but not either fraction's value. The worked numerical part shows how each missing factor is chosen.

1/31/5
Bars for 1/3 and 1/5.
Keep a whole bottom together

The fraction 2x+h means 2 divided by the whole x + h. The bar groups that bottom like parentheses. You do not split it into 2x plus 2h.

2x+h means 2 ÷ (x + h)

whole bottom stays grouped

Name all three zero risks

For 2x, the starting output needs x ≠ 0, the ending output needs x + h ≠ 0, and the rate needs h ≠ 0. These are three separate restrictions, even though the final formula has no isolated h in its bottom.

start: x ≠ 0

end: x + h ≠ 0

step: h ≠ 0

Part 1: Common-denominator arithmetic before letters

The top of a fraction is its numerator, counting pieces. The bottom is its denominator, naming their size. Rename both fractions with the same bottom before subtracting.

  • Never subtract fraction bottoms.
  • Multiplying a top and bottom by the same nonzero value preserves the fraction.
1/31/5
The same picture again: Bars for 1/3 and 1/5.

Worked example

Common bottoms with numbers before letters

This asks how to subtract two fractions with different bottoms. Find 13 − 15.

Show the steps and the answer
1/31/5
The same picture a third time: Bars for 1/3 and 1/5.
  1. Use common bottom 3·5=15.Neither bottom divides the other; their product gives a bottom both fractions can use.
  2. 13=515, and 15=315.Multiply each top and bottom by its missing factor: 5 for the first fraction and 3 for the second.
  3. 515−315=215.With equal bottoms, subtract only the tops and keep the common bottom.
Answer
215.
Double-check 13 = 515, and 215 + 315 = 515, so adding the subtracted amount recovers the starting fraction.

Part 2: A sum in the bottom

A bottom such as x + 2 remains one grouped expression. Substitute into the entire bottom and use both resulting bottoms in the common product.

  • For 3x+2, the bottoms are x + 2 and x + h + 2.
  • Require x ≠ −2 and x + h ≠ −2, as well as h ≠ 0.
(2, 3/4)(5, 3/7)246123
For the curve 3x+2, the dashed ruler between x = 2 and x = 5 has slope −328 and is not part of f.

Worked example

Difference quotient of f(x) = 52x+3

Let f(x) = 52x+3. Find and simplify the difference quotient f(x+h)−f(x)h. State every restriction on x and h.

Show the steps and the answer
  1. Plug in: f(x+h)=52(x+h)+3=52x+2h+3.Every x, including the one below the fraction bar, is replaced by the whole new input x + h. The 2 multiplies both x and h.
  2. Set up the subtraction: f(x+h)−f(x)=52x+2h+3−52x+3.The difference quotient starts with the new output minus the original output.
  3. Use the common bottom (2x+2h+3)(2x+3): 5(2x+3)−5(2x+2h+3)(2x+2h+3)(2x+3).The common bottom is the product of the two original bottoms. Each top is multiplied by the factor its own bottom is missing. The second top stays grouped so the minus sign applies to all of it.
  4. Distribute: 10x+15−10x−10h−15(2x+2h+3)(2x+3)=−10h(2x+2h+3)(2x+3).The outside minus changes the sign of every term in the second product. Then 10x − 10x = 0 and 15 − 15 = 0, which leaves only −10h.
  5. Divide by h: −10h(2x+2h+3)(2x+3)·1h=−10(2x+2h+3)(2x+3).Dividing by h is the same as multiplying by 1h. Here h is a whole factor of the numerator and h ≠ 0, so it cancels.
  6. State the restrictions: h≠0,2x+3≠0 (so x≠−32), and 2x+2h+3≠0 (so x+h≠−32).The original quotient divides by h. Both f(x) and f(x + h) must be defined, so neither bottom can be zero. These restrictions still hold after cancelling.
Answer
f(x+h)−f(x)h = −10(2x+2h+3)(2x+3), with h ≠ 0, x ≠ −32, and x + h ≠ −32.
Double-check Try x = 0 and h = 1. Directly: f(1) = 55 = 1 and f(0) = 53, so 1−531 = −23. With the formula: −10(0+2+3)(0+3) = −1015 = −23. The two results match.

Work to write

  1. f(x+h)=52x+2h+3
  2. 5(2x+3)−5(2x+2h+3)(2x+2h+3)(2x+3)
  3. −10h(2x+2h+3)(2x+3)
  4. −10(2x+2h+3)(2x+3)
  5. h≠0,x≠−32,x+h≠−32

Part 3: A fixed starting input

The same method works when the start is the number 2. The bottom at the start is then 3, and the bottom at the end is 3 + h.

  • For b(x) = 1x+1 starting at 2, rate = −13(3+h).
  • Require h ≠ 0 and h ≠ −3.
(2, 1/3)(5, 1/6)2461
Starting at 2 with h = 3, the dashed ruler has slope −118; it is not part of b.

Worked example

Difference quotient of 6x−2 starting at x = 5

Let f(x) = 6x−2. Find and simplify the difference quotient f(5+h)−f(5)h. State every restriction on h.

Show the steps and the answer
  1. Plug in: f(5+h)=6(5+h)−2=6h+3, and f(5)=65−2=63.Every x, including the one below the fraction bar, is replaced by the whole new input. f(5) is left as 63 so that both outputs are fractions with visible bottoms.
  2. Subtract: f(5+h)−f(5)=6h+3−63=6·3−6(h+3)3(h+3).The product of the two bottoms, 3(h + 3), is a common bottom. Each top is multiplied by the factor its bottom lacks. The second top, 6(h + 3), stays grouped so the minus sign applies to all of it.
  3. Simplify the numerator: 18−6(h+3)=18−6h−18=−6h. The difference is −6h3(h+3).Distributing the minus over the grouped term changes the sign of each piece. The constants 18 and −18 then cancel.
  4. Divide by h: −6h3(h+3)·1h=−63(h+3)=−2h+3.Dividing by h is the same as multiplying by 1h. Since h ≠ 0, the whole factor h cancels. Finally −63 reduces to −2.
  5. State the restrictions: h≠0 and h≠−3.The quotient has h in its bottom, so h ≠ 0. The input 5 + h must not equal 2, which would make the bottom of f zero, so h ≠ −3. Both restrictions remain after simplifying.
Answer
f(5+h)−f(5)h = −2h+3, for h ≠ 0 and h ≠ −3.
Double-check Take h = 1. Then f(6) = 64 = 1.5 and f(5) = 2, so the quotient is 1.5−21 = −0.5. The formula gives −21+3 = −0.5. The two values match.

Work to write

  1. f(5 + h) = 6h+3 and f(5) = 63
  2. 6h+3−63=18−6(h+3)3(h+3)
  3. Numerator: 18 − 6h − 18 = −6h
  4. −6h3(h+3)·1h=−2h+3
  5. Restrictions: h ≠ 0, h ≠ −3

Part 4: The input in both top and bottom

Every occurrence of the input must change. A common bottom can create two products on top, so multiply out both before combining their difference.

  • For xx+1, substitute x + h above and below the bar.
  • The common top simplifies to h, giving rate 1(x+1)(x+h+1).
  • Require h ≠ 0, x ≠ −1, and x + h ≠ −1.
(2, 2/3)(5, 5/6)2461
y = −1/(x + 1) + 1, solid, beside y = 0.0556x + 0.5556, dashed.

Worked example

The input also appears in the top

This asks for an average-rate formula when the input occurs above and below the fraction bar. Find the difference quotient of f(x) = xx+1.

Show the steps and the answer
(2, 2/3)(5, 5/6)2461
The same picture again: y = −1/(x + 1) + 1, solid, beside y = 0.0556x + 0.5556, dashed.

f(x)=xx+1

common bottom:

(x + 1)(x + h + 1)

combined top simplifies to h

(2, 2/3)(5, 5/6)2461
The same picture a third time: y = −1/(x + 1) + 1, solid, beside y = 0.0556x + 0.5556, dashed.
  1. Require h≠0,x≠−1, and x+h≠−1.The step and both original function bottoms must be nonzero.
  2. f(x+h)=x+hx+h+1, and f(x)=xx+1.The new input replaces both occurrences of x in the first output.
  3. Use common bottom (x+1)(x+h+1). The first top becomes (x+h)(x+1), and the second becomes x(x+h+1).Each top gets multiplied by the whole bottom missing from its own fraction.
  4. The combined top is (x+h)(x+1)−x(x+h+1).After the bottoms match, subtract the entire second top.
  5. (x+h)(x+1)=x·x+x·1+h·x+h·1=x2+x+ xh +h.Each term of the first parentheses multiplies each term of the second.
  6. x(x+h+1)=x2+ xh +x.The outside x multiplies all three terms inside.
  7. The combined top is (x2+x+ xh +h) − (x2+ xh +x) =x2+x+ xh +h−x2− xh −x=h.Distribute the outside minus, then cancel the matched x2, xh, and x terms.
  8. The output change is h(x+1)(x+h+1). Divide by h: h(x+1)(x+h+1)·1h=1(x+1)(x+h+1).The common nonzero multiplier h cancels from the multiplied fraction.
Answer
1(x+1)(x+h+1), with h ≠ 0, x ≠ −1, and x + h ≠ −1.
Double-check At x = 2 and h = 3, f(5) = 56 and f(2) = 23. Direct rate: (56 − 46) ÷ 3 = 16 · 13 = 118, matching 13·6.

5 How

  1. Plug in: replace every input occurrence, including those below a fraction bar, by the whole new input.
  2. Subtract: use the product of the original bottoms as a common bottom, multiply each top by its missing factor, and keep the second top grouped.
  3. Simplify: distribute the outside minus and combine the numerator terms.
  4. Cancel a valid h factor: multiply by 1h, cancel only a whole nonzero factor, and keep all original restrictions.

The fractional difference quotient in four moves

1
Is the bottom exactly x for the original function?
YesIts two output bottoms are x and x + h; use x(x + h).
NoWrite the actual two whole bottoms and use their product.
↓
2
Does the input appear in the numerator too?
YesReplace it there as well; multiply out the new top products carefully.
NoKeep the constant numerator as given.
↓
3
Has the minus reached the entire second numerator?
YesCombine matching numerator terms.
NoKeep that numerator in parentheses and distribute the outside −1.
↓
4
Is the starting input a fixed number?
YesUse its value for the starting bottom and start-plus-h for the ending bottom.
NoKeep the general starting letter.
↓
5
Does any original bottom or h equal 0?
YesExclude that input or step before cancellation.
NoCancel a whole common h factor if one is present.
  1. Plug in and record every nonzero-bottom condition.
  2. Subtract using a common bottom for the two output fractions.
  3. Simplify the combined top while leaving the bottom as a product.
  4. Divide by the step through multiplication by 1h, and cancel only a whole factor.

6 Worked example

Worked example

Difference quotient with x in the denominator

This asks for a reusable average-rate formula when the input is in a fraction bottom. Find f(x+h)−f(x)h for f(x) = 2x.

Show the steps and the answer
(1, 2)(2, 1)2412345
The same picture a third time: y = 2/x, solid, beside y = −x + 3, dashed.
  1. Require h≠0,x≠0, and x+h≠0. Write f(x+h)=2x+h and f(x)=2x.The outer division by h and both original function outputs need nonzero denominators.
  2. The output change is 2x+h−2x. Use common bottom x(x+h).The two bottoms are different, so both fractions must be renamed using the same-sized pieces before subtracting.
  3. 2x+h=2xx(x+h).Multiply the first top and bottom by x; since x ≠ 0, this multiplies the value by xx = 1.
  4. 2x=2(x+h)x(x+h).Multiply the second top and bottom by x + h; its nonzero condition makes this multiplication by 1.
  5. Subtract the tops: 2x−2(x+h)x(x+h).The common bottom stays fixed. Keep the whole second top grouped so the minus reaches both pieces.
  6. 2x−(2x+2h)x(x+h)=2x−2x−2hx(x+h)=−2hx(x+h).Distribute 2 inside the second top, then distribute the outside minus to +2x and +2h. The +2x and −2x cancel.
  7. Divide by the input change h: −2hx(x+h)÷h=−2hx(x+h)·1h.Dividing by nonzero h is multiplying by its reciprocal, so this multiplication compares the output change with the input step.
  8. −2hx(x+h)·1h=−2hh·x(x+h)=−2x(x+h).Multiply tops and bottoms. The whole numerator −2h and denominator both contain the nonzero factor h, which cancels.
Answer
−2x(x+h), with h ≠ 0, x ≠ 0, and x + h ≠ 0.
Double-check Use x = 2 and h = 3. f(5) = 25 and f(2) = 1, so the direct rate is (25 − 1) ÷ 3 = −35 · 13 = −15. The result gives −22·5 = −15.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

The unit reciprocal

This asks for the average-rate expression for the reciprocal function. Find the difference quotient of f(x) = 1x.

Show the steps and the answer
(2, 1/2)(5, 1/5)2461
The dashed ruler through the two endpoint curve points has slope −110 and is not part of f.
  1. Require h≠0,x≠0, and x+h≠0. The output change is 1x+h−1x.All three original divisions need nonzero bottoms.
  2. Use common bottom x(x+h): 1x+h=xx(x+h) and 1x=x+hx(x+h).Multiply the first top and bottom by x and the second top and bottom by x + h; both multipliers are nonzero.
  3. Subtract: x−(x+h)x(x+h)=x−x−hx(x+h)=−hx(x+h).The second top stays in parentheses until its minus reaches both x and h.
  4. Divide by h: −hx(x+h)·1h=−hh·x(x+h).Multiplying by the reciprocal divides the output change by the input step.
  5. −hh·x(x+h)=−1x(x+h).The nonzero factor h cancels; the minus sign remains.
Answer
  • −1x(x+h), also written −1x(x+h).
  • Require h ≠ 0, x ≠ 0, and x + h ≠ 0.
Double-check At x = 2 and h = 3, (15 − 12) ÷ 3 = (210 − 510) · 13 = −310 · 13 = −110. The result gives −12·5 = −110.

Rung 2 of 5

A constant numerator of 2

This asks for a reusable average-rate formula when the input is in a fraction bottom. Find f(x+h)−f(x)h for f(x) = 2x.

Show the steps and the answer
(1, 2)(2, 1)2412345
The same picture, shown 4 times now: y = 2/x, solid, beside y = −x + 3, dashed.
  1. Require h≠0,x≠0, and x+h≠0. Write f(x+h)=2x+h and f(x)=2x.The outer division by h and both original function outputs need nonzero denominators.
  2. The output change is 2x+h−2x. Use common bottom x(x+h).The two bottoms are different, so both fractions must be renamed using the same-sized pieces before subtracting.
  3. 2x+h=2xx(x+h).Multiply the first top and bottom by x; since x ≠ 0, this multiplies the value by xx = 1.
  4. 2x=2(x+h)x(x+h).Multiply the second top and bottom by x + h; its nonzero condition makes this multiplication by 1.
  5. Subtract the tops: 2x−2(x+h)x(x+h).The common bottom stays fixed. Keep the whole second top grouped so the minus reaches both pieces.
  6. 2x−(2x+2h)x(x+h)=2x−2x−2hx(x+h)=−2hx(x+h).Distribute 2 inside the second top, then distribute the outside minus to +2x and +2h. The +2x and −2x cancel.
  7. Divide by the input change h: −2hx(x+h)÷h=−2hx(x+h)·1h.Dividing by nonzero h is multiplying by its reciprocal, so this multiplication compares the output change with the input step.
  8. −2hx(x+h)·1h=−2hh·x(x+h)=−2x(x+h).Multiply tops and bottoms. The whole numerator −2h and denominator both contain the nonzero factor h, which cancels.
Answer
−2x(x+h), with h ≠ 0, x ≠ 0, and x + h ≠ 0.
Double-check Use x = 2 and h = 3. f(5) = 25 and f(2) = 1, so the direct rate is (25 − 1) ÷ 3 = −35 · 13 = −15. The result gives −22·5 = −15.

Rung 3 of 5

Difference quotient when the bottom is x + 2

Let f(x) = 3x+2. Find and simplify the difference quotient f(x+h)−f(x)h. State every restriction on x and h.

Show the steps and the answer
  1. Plug in: f(x+h)=3(x+h)+2=3x+h+2, so the difference quotient is 3x+h+2−3x+2h.Every x, including the one inside the bottom x + 2, is replaced by the whole new input x + h.
  2. Subtract using the common bottom (x + h + 2)(x + 2): 3x+h+2−3x+2=3(x+2)−3(x+h+2)(x+h+2)(x+2).The common bottom is the product of the two original bottoms. Each top is multiplied by the factor its bottom is missing. The second top, 3(x + h + 2), stays grouped so the minus applies to all of it.
  3. Simplify the top: 3(x+2)−3(x+h+2)=3x+6−3x−3h−6=−3h. The difference is −3h(x+h+2)(x+2).Distributing the outside minus changes the sign of every term in the second group. The 3x terms cancel and the 6 terms cancel, leaving only −3h.
  4. Divide by h: −3h(x+h+2)(x+2)·1h=−3(x+2)(x+h+2).Dividing by h is multiplying by 1h. h is a whole factor of the top, and h ≠ 0, so it cancels. The original restrictions still hold after the cancellation.
  5. Record the restrictions: h≠0,x≠−2, and x+h≠−2.h ≠ 0 because we divided by h. The bottom x + 2 cannot be 0, so x ≠ −2. The bottom x + h + 2 cannot be 0, so x + h ≠ −2.
Answer
f(x+h)−f(x)h = −3(x+2)(x+h+2), with h ≠ 0, x ≠ −2, and x + h ≠ −2.
Double-check Take x = 1 and h = 1. Directly: f(2) = 34 and f(1) = 33 = 1, so f(2)−f(1)1 = 34 − 1 = −14. From the formula: −3(1+2)(1+1+2) = −33·4 = −14. The two values match.

Work to write

  1. f(x+h)=3x+h+2
  2. 3(x+2)−3(x+h+2)(x+h+2)(x+2)
  3. 3x+6−3x−3h−6=−3h
  4. −3h(x+h+2)(x+2)·1h
  5. −3(x+2)(x+h+2)
  6. h≠0,x≠−2,x+h≠−2

Rung 4 of 5

Difference quotient of f(x) = 4x−1 starting at x = 3

Let f(x) = 4x−1. Find and simplify the difference quotient f(3+h)−f(3)h. State every restriction on h.

Show the steps and the answer
  1. Plug in: f(3+h)=4(3+h)−1=4h+2 and f(3)=43−1=42.Every x, including the one below the fraction bar, is replaced by the whole new input. The fraction 42 stays unreduced so that both terms are still fractions with their own bottoms.
  2. Subtract over the common bottom 2(h + 2): 4h+2−42=4·2−4(h+2)2(h+2)=8−(4h+8)2(h+2).The product of the original bottoms is a common bottom. Each top is multiplied by the factor its bottom is missing. The second top stays grouped so that the minus sign applies to all of it.
  3. Simplify: 8−4h−82(h+2)=−4h2(h+2)=−2hh+2.Distributing the minus gives 8 − 4h − 8. The constants cancel and leave −4h. Then the common factor 2 is reduced.
  4. Divide by h: −2hh+2·1h=−2h+2, with h≠0 and h≠−2.Dividing by h is the same as multiplying by 1h. The whole factor h cancels only because h ≠ 0. The restriction h ≠ −2 comes from the original bottom (3 + h) − 1 ≠ 0, and it stays even though h itself has cancelled.
Answer
f(3+h)−f(3)h = −2h+2, with h ≠ 0 and h ≠ −2.
Double-check Take h = 1. Then f(4) = 43 and f(3) = 2, so f(4)−f(3)1 = 43 − 2 = −23. The formula gives −21+2 = −23, which matches. The general result −4(x−1)(x+h−1) at x = 3 also gives −42(h+2) = −2h+2.

Work to write

  1. f(3+h)=4h+2,f(3)=42
  2. 4h+2−42=8−4(h+2)2(h+2)
  3. =−4h2(h+2)=−2hh+2
  4. −2hh+2·1h=−2h+2
  5. h≠0,h≠−2

Rung 5 of 5

The input also appears in the top

This asks for an average-rate formula when the input occurs above and below the fraction bar. Find the difference quotient of f(x) = xx+1.

Show the steps and the answer
(2, 2/3)(5, 5/6)2461
The same picture, shown 4 times now: y = −1/(x + 1) + 1, solid, beside y = 0.0556x + 0.5556, dashed.

f(x)=xx+1

common bottom:

(x + 1)(x + h + 1)

combined top simplifies to h

(2, 2/3)(5, 5/6)2461
The same picture, shown 5 times now: y = −1/(x + 1) + 1, solid, beside y = 0.0556x + 0.5556, dashed.
  1. Require h≠0,x≠−1, and x+h≠−1.The step and both original function bottoms must be nonzero.
  2. f(x+h)=x+hx+h+1, and f(x)=xx+1.The new input replaces both occurrences of x in the first output.
  3. Use common bottom (x+1)(x+h+1). The first top becomes (x+h)(x+1), and the second becomes x(x+h+1).Each top gets multiplied by the whole bottom missing from its own fraction.
  4. The combined top is (x+h)(x+1)−x(x+h+1).After the bottoms match, subtract the entire second top.
  5. (x+h)(x+1)=x·x+x·1+h·x+h·1=x2+x+ xh +h.Each term of the first parentheses multiplies each term of the second.
  6. x(x+h+1)=x2+ xh +x.The outside x multiplies all three terms inside.
  7. The combined top is (x2+x+ xh +h) − (x2+ xh +x) =x2+x+ xh +h−x2− xh −x=h.Distribute the outside minus, then cancel the matched x2, xh, and x terms.
  8. The output change is h(x+1)(x+h+1). Divide by h: h(x+1)(x+h+1)·1h=1(x+1)(x+h+1).The common nonzero multiplier h cancels from the multiplied fraction.
Answer
1(x+1)(x+h+1), with h ≠ 0, x ≠ −1, and x + h ≠ −1.
Double-check At x = 2 and h = 3, f(5) = 56 and f(2) = 23. Direct rate: (56 − 46) ÷ 3 = 16 · 13 = 118, matching 13·6.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: 2x+h=2x+2h.
At x = 1 and h = 1 the left side is 22 = 1, but the right side is 2 + 2 = 4.
✓ Instead: Keep the entire denominator grouped: 2x+h = 2 ÷ (x + h).
✗ Not this: Subtract fraction tops and bottoms directly: 23 − 25 = 0−2 = 0.
The original fractions have different piece sizes. Their actual difference is 1015 − 615 = 415.
✓ Instead: Make the bottoms match, then subtract only the tops.
✗ Not this: The bottom x(x + h) is required for every function with a fraction.
For 3x+2, the actual output bottoms are x + 2 and x + h + 2.
✓ Instead: Use the product (x + 2)(x + h + 2) for that function.
✗ Not this: After canceling h, allow h = 0, x = 0, or x + h = 0.
Those choices make an original division undefined. A shortened expression cannot restore missing original outputs or a zero input span.
✓ Instead: Preserve h ≠ 0, x ≠ 0, and x + h ≠ 0 for the reciprocal function.

Tips and tricks

  • The fraction bar groups the entire bottom: 2x+h means 2 ÷ (x + h).
  • Leave the common bottom as a product so its nonzero factors and any cancellations stay visible.
  • Show the combined top before opening the parentheses. In 2x − 2(x + h), the minus must reach both 2x and 2h.
  • Check with x = 2 and h = 3, provided every original bottom is nonzero.
  • A formula built from endpoint values does not describe values at an input where the original function is undefined; keep domain gaps in mind when interpreting the interval.

7 The trap

Subtracting bottoms, splitting a denominator sum, or losing the minus on the second numerator. Use a common bottom, subtract the whole second top, then divide by h.

Lesson 8 of 10

Increasing, decreasing, and constant

Subject: MathematicsLeans on: Interval notationLeans on: Subtracting negativesLeans on: Point coordinatesOpen this lesson
Leans on 1Comes back in 0

1 In plain words

Picture a hiking trail seen from the side. You walk toward the right and watch its height. A stretch is increasing when every move right takes you higher, decreasing when every move right takes you lower, and constant when the height stays the same. The height can be below zero and still rise. You name each stretch by its inputs, the positions along the horizontal axis, rather than its heights. A turning point is where the graph switches between rising and falling. This course reports behavior using open intervals, which leave the boundary inputs out. That reporting convention describes stretches. It does not make an endpoint disappear from the function's domain.

2 Picture it

(0, 0)turns at x = 0−222468(0, 0)turns at x = 0−3−2−11232468
Left to right, y = x2 drops until x = 0 and then climbs: decreasing on (−∞, 0), increasing on (0, ∞).

Reminder

Reminder: Interval notation, Subtracting negatives, Point coordinates
  • Interval notation. (1, 3) means 1 < x < 3. [1, 3] includes the endpoints. Infinity always has a parenthesis.
  • Subtracting negatives. −1 − (−2) = −1 + 2 = 1: output −1 is an increase from −2.
  • Point coordinates. In (2, 5), 2 is the horizontal input and 5 is the vertical output.

3 Why

Choose any two inputs a < b within one stretch. The change in input, b − a, is positive. If f(b) > f(a), the change in output is positive, so the average rate of change is positive. If f(b) < f(a), that rate is negative. Equal outputs give rate 0. This explains the three rows of the table. The letters a and b name any two inputs, rather than fixed endpoints to solve for. We split a graph at turns, flat boundaries, and breaks because a single label must describe the entire stretch between them.

4 The rule

On an interval, whenever a < b: increasing means f(a) < f(b); decreasing means f(a) > f(b); constant means f(a)=f(b).

This course reports intervals of behavior with open endpoints. These are comparisons between inputs, not a direction assigned to one point.

The same idea, five ways

Say it

'the output rises, falls, or stays the same as the input moves right.'

Write it

'f is increasing on (1, 3), decreasing on (3, 4), and constant on (4, 5).'

In math
  • 1 < x < 3
  • (1, 3)
  • {x : 1 < x < 3}
  • If 1 < a < b < 3, then f(a) < f(b).
  • Decreasing: a < b gives f(a) > f(b).
  • Constant: a < b gives f(a) = f(b).
Like

Walk east along an uphill, downhill, or level trail.

See it
13(1, 3)
After 'increasing on,' (1, 3) names all inputs strictly between the boundaries.

The same idea, other ways

As a picture

On y = x2, you move downhill toward 0 from the left, then uphill after 0. The vertex is the turning point of this U shaped graph, called a parabola. The heights are positive on both sides, but one side decreases and the other increases.

(0, 0)−222468
Read left to right to name direction, rather than checking whether the height is positive.
With two numbers

For f(x) = x2 on (0, ∞), let a = 1 and b = 3. Then f(1) = 1 < 9 = f(3). The letters are names for those inputs. On (−∞, 0), inputs −3 and −1 move right while outputs 9 and 1 move down.

input xoutput x²−39−111139
Compare inputs in increasing order within one side of the turn.
With recorded prices

The gas-price columns rise at each recorded year from 2005 through 2008, fall from 2008 to 2009, and rise at each recorded year afterward. These are sampled yearly data. They describe recorded changes and do not tell you the price at every time between the recorded years.

input yearoutput price ($ per gallon)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68
Compare successive columns; the data do not establish unmeasured behavior between years.
Directions and rate signs

The behavior table links each direction to the sign of every average rate taken within that stretch. The drawn columns repeat all three comparisons.

input behavioroutput average rate signincreasing: climbspositivedecreasing: dropsnegativeconstant: levelzero
The three columns preserve the direction, name, and average-rate facts of the behavior table.
Graph moving left to rightNameSign of every average rate of change there
climbsincreasingpositive
dropsdecreasingnegative
stays levelconstantzero

Part 1: Increasing: larger inputs give larger outputs

Think of walking up a ramp. You can begin below street level and still climb. Increasing describes the direction of the outputs as the inputs move right.

  • Any a < b within the interval gives f(a) < f(b).
  • Every average rate of change within that interval is positive.
  • Increasing does not mean positive outputs.
−4−2−4−3−2−11
The negative heights still rise as the input moves right.

Worked example

Increasing below zero

For f(x) = x on (−3, 0), name the output's direction.

Show the steps and the answer
−2−1−2−3−2−1
Direction and output sign are different facts.
  1. f(−2)=−2 and f(−1)=−1, so these outputs rise.−2 < −1, and −1 is greater than −2.
  2. The line rises throughout the interval: increasing on (−3,0).For this formula any larger input is itself a larger output.
Answer
Increasing on (−3, 0), with negative outputs there.
Double-check −1−(−2)−1−(−2) = 1, a positive average rate.

Part 2: Decreasing: larger inputs give smaller outputs

Walking down a ramp lowers your height as you move forward. On a decreasing stretch, the output becomes smaller as the input becomes larger, even when every output stays above zero.

  • Any a < b within the interval gives f(a) > f(b).
  • Every average rate of change within that interval is negative.
2246
The line falls to the right while the visible heights are nonnegative.

Worked example

Positive heights can decrease

Read the direction of f(x) = 6 − 2x on (0, 3).

Show the steps and the answer
422246
Positive outputs 4 and 2 go down from left to right.
  1. At x=1 the height is 4; at x=2 it is 2.6 − 2(1) = 4 and 6 − 2(2) = 2.
  2. Report decreasing on (0, 3).Every larger input subtracts a larger amount from 6.
Answer
Decreasing on (0, 3).
Double-check 2−42−1 = −2, a negative average rate.

Part 3: Constant: the output stays level

Imagine a shelf at one fixed height. Walking along it changes your position but never its height. Constant means every input in the stretch produces the same output.

  • Any a < b within the interval gives f(a) = f(b).
  • Every average rate of change there is 0.
  • A constant stretch is neither strictly increasing nor strictly decreasing.
212345
Every input shown gives height 4.

Worked example

Finding the level stretches on a water-tank graph

A graph shows the water depth h(t), in feet, in a tank, where the horizontal axis is labeled t (time in hours). The graph is made of straight segments with no breaks. It starts at a closed dot at (0, 2) and rises straight to (3, 5). It stays level at depth 5 from (3, 5) to (7, 5). It falls straight from (7, 5) to (9, 1). It stays level at depth 1 from (9, 1) to a closed dot at (12, 1), where the graph ends. Using open intervals of t, give where h is increasing, where h is decreasing, and where h is constant.

Show the steps and the answer
  1. Read the horizontal-axis label. The input is t (hours), so every interval is an interval of t-values, not depths.Intervals of behavior are always reported as intervals of the input variable.
  2. Mark the turning inputs and flat boundaries: t=3 (rising becomes level), t=7 (level becomes falling), and t=9 (falling becomes level). There are no breaks.The behavior can only change at these inputs, so they split the graph into separate stretches.
  3. Mark the domain ends: closed dots at t=0 and t=12. Both boundaries are finite, and no interval may use infinity.A finite end dot gives a finite boundary. Only arrows or stated continuation justify infinity.
  4. On 0 to 3, moving right, the depth goes from 2 up to 5. Whenever a < b there, h(a) < h(b), so h is increasing.Increasing means larger inputs always give larger outputs.
  5. On 3 to 7, the depth stays at 5. For any a < b there, h(a)=h(b)=5, so h is constant.Constant means every pair of inputs in the stretch gives equal outputs.
  6. On 7 to 9, the depth goes from 5 down to 1. Whenever a < b there, h(a) > h(b), so h is decreasing.Decreasing means larger inputs always give smaller outputs.
  7. On 9 to 12, the depth stays at 1. For any a < b there, h(a)=h(b)=1, so h is constant again.This is a second level stretch. It is separate from the first because a decreasing stretch lies between them.
  8. Write each behavior with open endpoints: increasing (0, 3); decreasing (7, 9); constant (3, 7) and (9, 12), written (3,7)∪(9,12).This course reports intervals of behavior with open endpoints. The two constant stretches are joined piece by piece with 'and' or ∪, never merged into (3, 12).
Answer
Increasing on (0, 3). Decreasing on (7, 9). Constant on (3, 7) and (9, 12), that is, (3, 7) ∪ (9, 12).
Double-check Test inputs from each interval. Increasing: h(1) is about 3 and h(2) is about 4, so 3 < 4. Decreasing: h(7.5) = 4 and h(8.5) = 2, so 4 > 2. Constant: h(4) = h(6) = 5 and h(10) = h(11) = 1. The intervals (0, 3), (3, 7), (7, 9) and (9, 12) cover the domain from 0 to 12, leaving out only the boundary inputs, and they do not overlap. The combined interval (3, 12) would be wrong, because h(8) = 3 differs from h(4) = 5.

Work to write

  1. Input variable: t
  2. Increasing: (0, 3)
  3. Decreasing: (7, 9)
  4. Constant: (3, 7) ∪ (9, 12)

5 How

  1. Begin at the left of the whole given graph. Read the horizontal-axis label: the input might be x or t.
  2. Mark turning inputs, flat boundaries, and breaks.
  3. Mark actual domain ends. A finite end dot gives a finite boundary; arrows or stated continuation can justify infinity.
  4. Move right within each separate stretch. Name it increasing, decreasing, or constant.
  5. Write open intervals of inputs. Keep pieces on opposite sides of a break separate.
  6. Give each requested behavior on its own line. Say 'and' between stretches; ∪ joins the listed stretches piece by piece.

Read, split, name each stretch

1
Does the graph break at an excluded input or a vertical asymptote?
YesSplit the interval at that input. A vertical asymptote is a vertical line the curve approaches while outputs grow in size.
NoContinue within the connected piece.
↓
2
Does the graph stop at a finite end dot, with no stated continuation?
YesUse that input as the finite boundary. A viewing-window edge alone does not prove the graph stops.
NoUse infinity only when arrows or the stated domain justify continuation.
↓
3
Does the graph turn or become flat?
YesSplit before labeling the stretches.
NoKeep the current stretch together.
↓
4
Does the height rise as you move right?
YesIncreasing.
NoFalling is decreasing; unchanged height is constant.
  1. Find the input axis and actual domain ends.
  2. Split at turns, flat boundaries, and gaps.
  3. Read left to right on each separate piece.
  4. Write open intervals of inputs.

6 Worked example

Worked example

Reading increasing and decreasing intervals from a downward parabola

The graph shows y = f(x), where f(x) = -(x - 2)2 + 3. The horizontal axis is labeled x. The graph is a parabola that opens downward. Its highest point (turning point) is at (2, 3). It also passes through (0, -1), (1, 2), (3, 2) and (4, -1). Arrows on both ends show that the graph continues forever to the left and to the right, so the domain is all real numbers. State the open interval(s) on which f is increasing, the open interval(s) on which it is decreasing, and any interval on which it is constant.

(2, 3)(0, -1)(4, -1)−2246−6−4−224
Graph of f(x) = -(x - 2)2 + 3, a downward parabola with its highest point at (2, 3), passing through (0, -1)…
Show the steps and the answer
  1. Start at the far left of the graph and read the horizontal-axis label. It is x, so every interval is written in terms of x-values.Intervals of behavior describe inputs, so we must know which variable is the input.
  2. Mark the turning input x=2, where the graph changes from rising to falling at (2, 3). There are no flat stretches and no breaks.Behavior can only change at turning inputs, flat boundaries or breaks, so these split the domain into stretches.
  3. Mark the domain ends. Both ends carry arrows, so the graph continues without end in both directions. The left boundary is -∞ and the right boundary is ∞.Arrows show the graph continues, which justifies using infinity instead of a finite endpoint.
  4. Move right through the stretch from -∞ to 2. Here the graph rises. For example, at 0 < 1 we have f(0)=-1<f(1)=2. This stretch is increasing.Increasing means that whenever a < b on the interval, f(a) < f(b).
  5. Move right through the stretch from 2 to ∞. Here the graph falls. For example, at 3 < 4 we have f(3)=2>f(4)=-1. This stretch is decreasing.Decreasing means that whenever a < b on the interval, f(a) > f(b).
  6. Look for a stretch where the output stays the same. There is none, because the curve never runs level over an interval.Constant requires f(a) = f(b) for all a < b in an interval. A single highest point is not an interval.
  7. Write each behavior on its own line, using open intervals of x: increasing on (-∞,2), decreasing on (2,∞), constant on no interval.This course reports intervals with open endpoints. The turning input x = 2 belongs to neither stretch.
Answer
Increasing on (-∞, 2). Decreasing on (2, ∞). Constant on no interval.
Double-check Compare inputs on each side of x = 2 using the formula f(x) = -(x - 2)2 + 3. On the left, f(-1) = -(-3)2 + 3 = -6 and f(0) = -1. Since -1 < 0 and -6 < -1, the function is increasing. On the right, f(4) = -1 and f(5) = -(3)2 + 3 = -6. Since 4 < 5 and -1 > -6, the function is decreasing. Both results agree with the answer.

Work to write

  1. Turning input: x = 2 (highest point at (2, 3)); arrows on both ends, so the domain is (-∞, ∞)
  2. Increasing on (-∞, 2)
  3. Decreasing on (2, ∞)
  4. Constant: on no interval

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 6

Reading one rising line: where is it increasing, decreasing, or constant?

The graph shows y = g(x), where g(x) = 12x + 1. The horizontal axis is labeled x. The graph is a straight line with arrows at both ends, so it continues forever to the left and to the right. It passes through (-4, -1), (0, 1) and (2, 2). Give the open intervals on which g is increasing, decreasing, and constant.

(-4, -1)(0, 1)(2, 2)−6−4−2246−4−224
The line y = g(x) = 12x + 1 rises from left to right through (-4, -1), (0, 1) and (2, 2), with arrows at both…
Show the steps and the answer
  1. Start at the far left of the graph and read the horizontal-axis label: the input is x.Intervals of behavior are written in terms of the input, so we must know which variable the intervals describe.
  2. Look for turning inputs, flat stretches, and breaks. There are none. The graph is one unbroken straight line with slope 12.A non-vertical line with nonzero slope never changes direction and never levels off, so there are no boundaries inside it.
  3. Check the domain ends. Both ends have arrows, so the domain runs from -∞ to ∞.Arrows show that the graph continues without end, which justifies using infinity. A finite end dot would have given a finite boundary instead.
  4. Move right along the single stretch. Compare the inputs -4<2: g(-4)=-1 and g(2)=2, and -1<2. The outputs rise as x moves right, so the stretch is increasing.Increasing means that whenever a < b, f(a) < f(b). A positive slope makes this true for every pair of inputs on the line.
  5. Write the behavior as an open interval of inputs: increasing on (-∞,∞).This course reports intervals with open endpoints, and ∞ is never included in any case.
  6. Report each requested behavior on its own line. Decreasing: none. Constant: none.No pair of inputs a < b gives g(a) > g(b) or g(a) = g(b), so neither behavior occurs anywhere.
Answer
Increasing on (-∞, ∞). Decreasing: none. Constant: none.
Double-check Test another pair of inputs, -4 < 0: g(-4) = 12(-4) + 1 = -1 and g(0) = 1. Since -1 < 1, the output rises again. The slope 12 is positive, which confirms that the line rises everywhere and has no flat or falling part.

Work to write

  1. Input variable: x; no turning points, flat stretches, or breaks
  2. Domain: arrows at both ends, so (-∞, ∞)
  3. Increasing on (-∞, ∞)
  4. Decreasing: none
  5. Constant: none

Rung 2 of 6

A parabola that opens upward with its turn at x = -3

The graph shows y = h(x), where h(x) = 12(x + 3)2 - 2. The horizontal axis is labeled x. The graph is a parabola that opens upward, with arrows on both ends, so it continues forever to the left and to the right. Its lowest point (turning point) is at (-3, -2). It also passes through (-7, 6), (-5, 0), (-4, -1.5), (-2, -1.5), (-1, 0) and (1, 6). Give the open intervals where h is increasing, decreasing, and constant.

(-3, -2)(-5, 0)(-1, 0)(-7, 6)(1, 6)−8−6−4−22−4−2246
The graph of y = h(x) = 12(x + 3)2 - 2 opens upward, with its lowest point at (-3, -2).
Show the steps and the answer
  1. Read the horizontal-axis label. It is x, so every interval will be a set of x-values.Intervals of behavior describe inputs. They do not describe outputs.
  2. Mark the turning input: the lowest point is (-3,-2), so mark x=-3. There are no flat pieces and no breaks.The graph can change from falling to rising only at a turning input. Flat boundaries and breaks would also split stretches, but this graph has none.
  3. Check the domain ends. Both ends have arrows, so the inputs run from -∞ to ∞.Arrows show that the graph continues without end. That justifies using -∞ and ∞ as boundaries.
  4. Move right on the stretch left of x=-3. Take -5<-4: h(-5)=0 and h(-4)=-1.5, so h(-5)>h(-4). Also h(-7)=6>h(-5)=0. Name this stretch decreasing.Decreasing means that whenever a < b, f(a) > f(b). The outputs fall as the inputs grow.
  5. Move right on the stretch right of x=-3. Take -2<-1: h(-2)=-1.5 and h(-1)=0, so h(-2)<h(-1). Also h(-1)=0<h(1)=6. Name this stretch increasing.Increasing means that whenever a < b, f(a) < f(b). The outputs rise as the inputs grow.
  6. Write the open intervals: decreasing on (-∞,-3) and increasing on (-3,∞). No stretch is constant.This course reports intervals with open endpoints. The turning input x = -3 is not placed in either interval, because a single point has no direction by itself. No two different inputs give equal outputs along a stretch, so there is no constant interval.
Answer
  • Increasing: (-3, ∞)
  • Decreasing: (-∞, -3)
  • Constant: none
Double-check Compare two inputs on each side of the turn. On the left, -7 < -5 and h(-7) = 12(16) - 2 = 6 > h(-5) = 12(4) - 2 = 0, so h is decreasing. On the right, -1 < 1 and h(-1) = 0 < h(1) = 12(16) - 2 = 6, so h is increasing. The symmetric points (-4, -1.5) and (-2, -1.5) lie on opposite sides of x = -3, which agrees with the turn being at x = -3.

Work to write

  1. Horizontal axis is x; turning input x = -3; arrows on both ends
  2. Decreasing on (-∞, -3)
  3. Increasing on (-3, ∞)
  4. Constant: none

Rung 3 of 6

A parabola that stops at two solid end dots

The graph shows y = p(t), where p(t) = (t - 1)2 - 4, drawn only for -2 ≤ t ≤ 3. The horizontal axis is labeled t. The graph is a parabola that opens upward. It starts at a solid (filled) dot at (-2, 5) and ends at a solid (filled) dot at (3, 0). There are no arrows. Its lowest point (turning point) is at (1, -4). It also passes through (-1, 0), (0, -3) and (2, -3). State the open intervals on which p is increasing, decreasing, and constant.

(-2, 5)(1, -4)(3, 0)(-1, 0)(0, -3)(2, -3)−4−224−44
The graph of p(t) = (t - 1)2 - 4 for -2 ≤ t ≤ 3.
Show the steps and the answer
  1. Start at the left end of the graph and read the horizontal axis. It is labeled t, so every interval is an interval of t-values.Intervals of behavior describe inputs, and the input variable here is t, not x.
  2. Mark the turning input t=1, where the lowest point (1,-4) is. There are no flat pieces and no breaks.Behavior can change only at turning inputs, at the ends of flat pieces, or at breaks.
  3. Mark the domain ends t=-2 and t=3. Both are solid end dots and there are no arrows.A finite end dot gives a finite boundary. Without arrows the graph does not continue, so -∞ and ∞ are not used.
  4. Move right from t=-2 to t=1. The heights go 5, 0, -3,-4, so this stretch is decreasing.For a < b in this stretch, p(a) > p(b). For example, p(-1) = 0 > p(0) = -3.
  5. Move right from t=1 to t=3. The heights go -4,-3, 0, so this stretch is increasing.For a < b in this stretch, p(a) < p(b). For example, p(2) = -3 < p(3) = 0.
  6. Write each stretch as an open interval of t: decreasing on (-2,1) and increasing on (1, 3). No stretch is constant.This course reports intervals of behavior with open endpoints, even when the end dots are solid. No two inputs in any stretch give equal outputs, so p is never constant.
Answer
  • Increasing: (1, 3)
  • Decreasing: (-2, 1)
  • Constant: none
Double-check Test two inputs inside each interval. On (-2, 1): p(-1.5) = 6.25 - 4 = 2.25 and p(0.5) = 0.25 - 4 = -3.75. Since 2.25 > -3.75, p is decreasing there. On (1, 3): p(1.5) = 0.25 - 4 = -3.75 and p(2.5) = 2.25 - 4 = -1.75. Since -3.75 < -1.75, p is increasing there. Both intervals stop at the finite ends -2 and 3, with no ∞.

Work to write

  1. Input variable: t
  2. Turning input: t = 1; domain ends: t = -2 and t = 3 (solid dots, no arrows)
  3. Increasing on (1, 3)
  4. Decreasing on (-2, 1)
  5. Constant: none

Rung 4 of 6

A break splits the branches

Read the behavior of f(x) = 1x. Why are there two decreasing intervals?

Show the steps and the answer
−4−224−4−224
Each branch falls separately; 0 is an excluded input.
  1. Exclude input 0.10 has no value.
  2. On the left, outputs at −2 and −1 fall from −12 to −1; on the right, outputs at 1 and 2 fall from 1 to 12.Each branch decreases as its inputs move right.
  3. Keep (−∞,0) and (0,∞) separate.No connected behavior interval can cross the missing input 0.
Answer
  • Decreasing on (−∞, 0) and on (0, ∞), each branch separately.
  • Increasing on no interval.
  • Constant on no interval.
Double-check Across the break, −1 < 1 but f(−1) = −1 < 1 = f(1), so the function is not decreasing across the whole union.

Rung 5 of 6

A level stretch between sloping pieces

The connected straight pieces continue at both ends. They fall to (−3, −1), rise to (0, 4), stay at height 4 until x = 2, then rise again. Report all behaviors.

Show the steps and the answer
(−3, −1)flat beginsflat ends−4−224−22468
Dots sample connected straight pieces; the sample height at −1 is rounded and both ends continue.
  1. Mark −3, 0, and 2.Those inputs bound changes in direction or flatness.
  2. Read down, up, level, up as you move right.These are the directions of the given pieces.
  3. List the flat interval separately.Equal outputs there fail the strictly increasing comparison.
Answer
  • Decreasing on (−∞, −3).
  • Increasing on (−3, 0) and (2, ∞).
  • Constant on (0, 2).
Double-check Height 4 repeats across the middle stretch, so its average rate is 0.

Rung 6 of 6

Ends, a turn, and a level piece

This complete graph consists of straight segments joining corners (−4, −1), (−2, 3), (1, −2), (3, −2), and (5, 4). It stops at included endpoints. Report all behaviors.

filled endturnflat startsflat endsfilled end−4−224−224
Dots sample the connected straight segments.
Show the steps and the answer
  1. Mark endpoints −4 and 5 and inner boundaries −2, 1, and 3.Those inputs split the given segments.
  2. Read up, down, level, up across the successive segments.The segments connect the stated corners with no additional turns.
  3. Keep the two rising intervals separate and finite.The falling and level intervals lie between them and the graph stops.
Answer
  • Increasing on (−4, −2) and (3, 5).
  • Decreasing on (−2, 1).
  • Constant on (1, 3).
  • Domain: [−4, 5].
Double-check Each answer describes one segment direction. The endpoint inputs remain in the domain.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: Increasing means positive outputs.
For f(x) = x on (−3, 0), every output is negative but larger inputs give larger outputs.
✓ Instead: Read whether height rises as the input moves right.
✗ Not this: (1, 3) always names a point.
After 'increasing on' it names all inputs between 1 and 3; after 'passes through' it names input 1 and output 3.
✓ Instead: Read the surrounding words.
✗ Not this: Because 1x decreases on both branches, it decreases across their whole union.
−1 < 1 but f(−1) = −1 < 1 = f(1), which fails the decreasing comparison.
✓ Instead: Report the two decreasing branches separately.
✗ Not this: Using (0, ∞) for x means its domain excludes 0.
0 = 0 is defined. Domain [0, ∞) and course behavior interval (0, ∞) answer different questions.
✓ Instead: Keep the domain answer separate from the open behavior-interval answer.

Tips and tricks

  • Read right, then name the height's direction.
  • Put inputs on interval boundaries and keep outputs for heights.
  • Use 'and' between behavior pieces. With ∪, mean the listed intervals piece by piece, especially across a gap.
  • Use the stated domain, endpoint marks, and arrows to distinguish an actual end from a viewing-window edge.

7 The trap

Writing the intervals in y-values, for example 'increasing from −2 to 5' because the heights run from −2 to 5. Intervals of increase and decrease are always inputs (x-values or t-values), read along the horizontal axis.

Lesson 9 of 10

Local maxima and minima

Subject: MathematicsLeans on: Ordered pairsLeans on: Comparing negativesLeans on: Squaring and substitutionOpen this lesson
Leans on 0Comes back in 0

1 In plain words

Think of one hill on a roller coaster. Its top can beat nearby heights even when a taller hill stands farther away. That is what local means: compare close neighbors on both sides. At a hilltop point (−2, 16), the local maximum is the height 16, and it occurs at input x = −2. A valley gives a local minimum in the same way. Relative maximum and relative minimum are other names for these ideas. Together they are local extrema. Rising then falling gives a hilltop; falling then rising gives a valley. The formal rule below also allows tied nearby heights, so a flat stretch needs separate attention.

2 Picture it

outer samplelocal max (−2, 16)local min (2, −16)outer sample−4−224−32−161632outer samplelocal max (−2, 16)local min (2, −16)outer sample−4−3−2−11234−32−161632
The dots sample the given smooth curve f(x) = x3 − 12x.

Reminder

Reminder: Ordered pairs, Comparing negatives, Squaring and substitution
  • Ordered pairs. (−2, 16) means input −2, output 16. f(−2) = 16 says the same thing.
  • Comparing negatives. −1.633 is greater than −1.667 because it is closer to zero on the number line.
  • Squaring and substitution. (−2)2 = 4, but −(−2)2 = −4. Parentheses keep the substituted input together.

3 Why

Near a hilltop, points on the left are lower because the curve climbs toward the top; points on the right are lower because it falls away. This makes the top higher than nearby points on both sides. A valley reverses those comparisons. We compare within an open interval around the input, so an endpoint without domain points on both sides does not qualify under this text's definition. The value is an output because maximum and minimum describe height. The rule uses 'at least' and 'at most,' so equal nearby heights are allowed; strict hilltops and valleys are the turning cases.

4 The rule

A local maximum at input b has f(b)≥f(x) for every x in some open interval around b contained in the domain.

A local minimum has f(b)≤f(x) there. Its value is f(b); its location is b; its point is (b, f(b)). Equal nearby heights are allowed by these nonstrict inequalities. A strict local extremum uses > or < for every nearby x≠b.

The same idea, five ways

Say it

'local maximum of 16 at x = −2' and 'local minimum of −16 at x = 2.'

Write it

'The local maximum value is 16; it occurs at input −2, at point (−2, 16).'

In math
  • Location: b.
  • Value: f(b).
  • Point: (b, f(b)).
  • Local maximum: f(b) ≥ f(x) for all x in a small open domain interval around b.
  • Local minimum: f(b) ≤ f(x) in that neighborhood.
Like

The tallest or shortest house within your neighborhood, with nearby houses on both sides.

See it
left neighborvalue 3 at input 2right neighbor24−11234
The input 2 locates the vertex; the output 3 is its local maximum value.

The same idea, other ways

As a picture

For the given curve f(x) = x3 − 12x, the hilltop point is (−2, 16) and the valley point is (2, −16). The ends continue past those heights. These turns can be local extrema without being the whole graph's highest or lowest outputs.

local max 16local min −16−4−224−32−161632
These dots are exact samples of the given smooth curve, whose stated turns are at −2 and 2.
As a neighborhood story

The tallest house on your block need not be the tallest in the whole city. A local maximum compares within one neighborhood of inputs. The neighborhood must include domain points to the left and right, as visiting houses on both sides of your address would.

13turn input b
An open neighborhood around b = 2 includes inputs on both sides of 2.
Why the comparison holds

For f(x) = 3 − (x − 2)2, the square is never negative. Thus f(x) ≤ 3 for every input, and f(2) = 3. At any other input the square is positive, so the height is lower than 3. This proves the hilltop comparison, rather than relying on two sample neighbors.

neighborb = 2neighbor24−11234
The parabola's vertex is (2, 3), and the square argument proves that other heights are lower.
With recorded costs

Among adjacent recorded years, the 2008 gas price $3.30 exceeds the 2007 and 2009 prices, while the 2009 price $2.41 is below the 2008 and 2010 prices. These are a sampled peak and dip useful for discussing observed costs. They are not a proof of continuous local extrema between recorded years.

input yearoutput price ($ per gallon)20072.8420083.320092.4120102.84↓ evaluate: input given, read the output below it
Adjacent sampled columns support an observed peak in 2008 and dip in 2009; unrecorded prices are unknown.

Part 1: Local maximum: hilltop value, input, and point

A hilltop has an address and a height. The input is its address along the axis; the output is its height. A parabola is the U shaped graph of a quadratic, possibly opening down. Its vertex is its turning point.

  • Increasing then decreasing gives a strict local maximum.
  • A maximum value is an output; its input says where it occurs.
  • The formal nonstrict comparison also allows nearby equal outputs.
leftvertexright24−11234
Over the open interval (1, 3) around input 2, no point is higher than the vertex.

Worked example

Read a hilltop

For f(x) = 3 − (x − 2)2, find the local maximum value, where it occurs, and its point.

Show the steps and the answer
23224−11234
y=−(x−2)2+3.
  1. f(2)=3−02=3.Input 2 makes the square zero.
  2. At every different input, (x−2)2>0, so f(x) < 3.A nonzero real number has a positive square.
  3. The value is 3, the location is x=2, and the point is (2, 3).The ordered pair puts the input first and the output second.
Answer
  • Local maximum value: 3.
  • Location: x = 2.
  • Point: (2, 3).
Double-check f(1) = f(3) = 2 agrees with the pictured neighboring heights. The square argument supplies the proof for all nearby inputs.

Part 2: Local minimum: valley value, input, and point

At a valley floor, both nearby sides lead upward. You still give its height and its address separately. Minimum describes the lower output, rather than the smallest input.

  • Decreasing then increasing gives a strict local minimum.
  • The value is the output at the valley.
  • The formal nonstrict comparison allows ties.
leftvertexright−2−22
The valley height is −2 at input −1.

Worked example

Read a negative valley

Find the local minimum of f(x) = (x + 1)2 − 2.

Show the steps and the answer
−1−2−1−2−22
y=(x+1)2−2.
  1. f(−1)=02−2=−2.x + 1 equals zero at input −1.
  2. Every other input gives (x+1)2>0, so its output exceeds −2.Adding a positive number to −2 raises it.
  3. Report local minimum of −2 at x=−1.The output supplies the value and the input supplies the location.
Answer
  • Local minimum of −2 at x = −1.
  • Point: (−1, −2).
Double-check f(−2) = −1 and f(0) = −1 are both above −2. The square comparison proves the result beyond these samples.

Part 3: Ties on flat stretches and shoulders

Imagine a level shelf instead of a pointed hill. Its interior has nearby heights equal on both sides. 'At least as high' and 'at most as high' both permit equality, so the formal nonstrict definition can call those points both local types. They are not isolated turning points.

  • For a constant function, every interior input is a nonstrict local maximum and a nonstrict local minimum.
  • A flat stretch has no strict local extrema, because nearby equal heights fail the greater-than and less-than comparisons.
  • A shoulder at the edge of a flat stretch may satisfy one nonstrict comparison without an up-then-down turn. Compare both sides.
samesamesame−221234
Equal nearby heights satisfy the nonstrict rule but create no strict turn.

Worked example

Proving a strict local maximum on a downward parabola

The graph shows g(x) = −2(x − 1)2 + 6, defined for all real numbers. Points read from the graph are (−1, −2), (0, 4), (1, 6), (2, 4) and (3, −2). The curve keeps falling to the left of (−1, −2) and to the right of (3, −2). Show that g has a strict local maximum at x = 1. Give its value, its location and its point. Then explain why g has no other local maximum or minimum.

(−1, −2)(0, 4)(1, 6)(2, 4)(3, −2)−224−4−22468
Downward parabola g(x) = −2(x − 1)2 + 6 with the points (−1, −2), (0, 4), (1, 6), (2, 4) and (3, −2) marked.
Show the steps and the answer
  1. Scan the listed points from left to right. The heights go −2, 4, 6, then fall to 4 and −2.A local maximum shows up where the graph changes from rising to falling. Here that change happens at x = 1.
  2. Check for flat stretches. The heights on either side of x=1 are 4 and 4. Both are less than 6. The curve is a parabola, so it has no flat pieces.The nonstrict rule allows equal nearby heights. A strict maximum needs every nearby height to be lower, so we must rule out ties.
  3. Choose the open interval (0, 2) around b=1. It lies inside the domain, which is all real numbers, so there are inputs on both sides of 1.A local extremum needs a neighbourhood inside the domain on both sides of b. An endpoint would not qualify.
  4. Prove the strict inequality algebraically. For any x≠1,(x−1)2>0. Multiplying by −2 gives −2(x−1)2<0. Adding 6 gives g(x)<6=g(1).A strict local maximum needs g(1) > g(x) for every nearby x ≠ 1. The algebra shows this holds for every x ≠ 1 in (0, 2), and in fact for every real x ≠ 1.
  5. Read the location as the input x=1 and the value as the output g(1)=−2(0)2+6=6. The point is (1, 6).The location is the horizontal input b, the value is f(b), and the point is the pair (b, f(b)).
  6. Look at every other input c≠1. If c < 1, the graph is still rising at c, so slightly to the right of c it is higher. If c > 1, the graph is falling at c, so slightly to the left of c it is higher. No turn from falling to rising occurs anywhere.At every c ≠ 1 there is a higher point arbitrarily close, so c cannot be a local maximum. With no change from falling to rising, there is no local minimum.
Answer
g has a strict local maximum at x = 1 with value 6. The point is (1, 6). For every x ≠ 1, g(x) = 6 − 2(x − 1)2 < 6. g has no other local maximum and no local minimum.
Double-check Test inputs close to 1 on each side. g(0.9) = −2(−0.1)2 + 6 = 5.98 < 6 and g(1.1) = −2(0.1)2 + 6 = 5.98 < 6, which confirms the strict inequality. The listed points also agree with the formula: g(0) = −2(1) + 6 = 4 and g(3) = −2(4) + 6 = −2.

Work to write

  1. Rising then falling at x = 1: heights 4, 6, 4
  2. Neighbourhood (0, 2) lies in the domain on both sides of 1
  3. For x ≠ 1: (x − 1)2 > 0, so −2(x − 1)2 < 0, so g(x) < 6 = g(1)
  4. Strict local maximum at x = 1, value 6, point (1, 6)
  5. No other local extrema: no other change of direction

Part 4: Read supplied estimates without a graphing calculator

A supplied graph may place a turn between labeled grid lines. Read its input and height as estimates. Checking nearby formula values can catch a misread sign or height, but a few samples cannot prove where every turn occurs.

  • The supplied graph of f(x) = 2x + x3 has a local maximum near (−2.449, −1.633) and a local minimum near (2.449, 1.633), with coordinates rounded.
  • The local maximum may be lower than a different local minimum, because the comparisons are in separate neighborhoods.
  • Input 0 is excluded, so behavior intervals must split there.
estimated local maxestimated local min−6−4−2246−4−224
Rounded sample dots represent the given two smooth branches.

Worked example

Checking both branches of estimated heights for local extrema

A function f is defined only on the interval [-3, 3]. Between consecutive integer inputs its graph rises or falls steadily, with no extra bumps. Estimated heights read from the graph are: f(-3) = 2, f(-2) = 5, f(-1) = 7, f(0) = 5, f(1) = 1, f(2) = 3, f(3) = 6. Find every local maximum and local minimum of f. For each one, give its value and its location in a full sentence and state its point (b, f(b)). Explain why the endpoints x = -3 and x = 3 are not counted.

input xoutput f(x) (estimated height)−32−25−1705112336
A table of 7 values of f(x) (estimated height), for x from −3 to 3.
Show the steps and the answer
input xoutput f(x) (estimated height)−32−25−1705112336↓ evaluate: input given, read the output below it
Estimated heights of f on [-3, 3].
  1. Scan the heights from left to right: 2→5→7 is rising, 7→5→1 is falling, and 1→3→6 is rising again.Local extrema can only occur where the graph changes from rising to falling or from falling to rising.
  2. Mark the turns. At x=-1 the graph changes from rising to falling. At x=1 it changes from falling to rising.Rising then falling identifies an isolated local maximum. Falling then rising identifies an isolated local minimum.
  3. Check both branches at x=-1. On the left, f(-2)=5<7. On the right, f(0)=5<7. Since the graph climbs steadily up to x=-1 and descends steadily after it, every x near -1 with x≠-1 has f(x) < 7.A local maximum needs f(b) ≥ f(x) on an open interval around b, on both sides. Here the inequality is strict, so the maximum is strict.
  4. Check both branches at x=1. On the left, f(0)=5>1. On the right, f(2)=3>1. So every x near 1 with x≠1 has f(x) > 1.A local minimum needs f(b) ≤ f(x) on both sides of b. Again there are no equal heights nearby, so the minimum is strict.
  5. Look for flat stretches or shoulders. The table shows no repeated heights next to each other, and the graph rises or falls steadily between the table inputs.The nonstrict rule would also allow equal nearby heights, so level stretches must be inspected. There are none here.
  6. Examine the endpoints. At x=-3 (height 2) the graph only extends to the right. At x=3 (height 6) it only extends to the left.The domain is [-3, 3], so no open interval around -3 or around 3 lies inside the domain. Endpoints are considered separately, for absolute extrema.
  7. Read the horizontal input as the location and the vertical output as the value. The maximum has location -1 and value 7. The minimum has location 1 and value 1.The location is b, the value is f(b), and the point is (b, f(b)).
Answer
f has a local maximum value of 7 at x = -1, at the point (-1, 7). It has a local minimum value of 1 at x = 1, at the point (1, 1). The endpoints x = -3 and x = 3 are not local extrema because there is no domain interval on both sides of them.
Double-check Compare each candidate with its two neighbors. 5 < 7 > 5 confirms a peak at x = -1, and 5 > 1 < 3 confirms a valley at x = 1. The endpoint height 6 at x = 3 is larger than 1 and smaller than 7, but no input exists to its right, so it cannot be tested on both branches.

Work to write

  1. Rising then falling at x = -1: f(-2) = 5 < f(-1) = 7 > f(0) = 5
  2. Falling then rising at x = 1: f(0) = 5 > f(1) = 1 < f(2) = 3
  3. Local maximum value 7 at x = -1; point (-1, 7)
  4. Local minimum value 1 at x = 1; point (1, 1)
  5. x = -3 and x = 3 are endpoints with no interval on both sides, so they are not local extrema

Part 5: Retained sample data for the cubic picture

Sample dots are like a few photographed positions along a trail. They record exact or rounded heights at specific inputs. The given smooth curve supplies its behavior between those photographed positions.

  • These tables retain every plotted sample from the original cubic picture and add outer samples past the local heights.
  • Noninteger table heights are rounded to two decimal places.
  • Finite samples support the stated curve; they do not prove every nearby comparison or every turn.
input xoutput x³ − 12x, rounded−4.25−25.77−4−16−3.75−7.73−3.5−0.88−3.254.67−39−2.7512.20−2.514.38−2.2515.61
These retained sample values trace the given cubic to its left local maximum.

Worked example

Locating the turns of a cubic from a table of sample heights

The cubic f(x) = x3 − 12x + 4 is defined for all real numbers. Its heights at the integer inputs from −4 to 4 are: f(−4) = −12, f(−3) = 13, f(−2) = 20, f(−1) = 15, f(0) = 4, f(1) = −7, f(2) = −12, f(3) = −5, f(4) = 20. Between consecutive integer inputs the graph rises or falls steadily, with no extra bumps. Find every local maximum and local minimum of f in this window. Give the value and the location of each, and the point (b, f(b)).

input xoutput f(x) = x³ − 12x + 4−4−12−313−220−115041−72−123−5420
A table of 9 values of f(x) = x3 − 12x + 4, for x from −4 to 4.
Show the steps and the answer
input xoutput f(x) = x³ − 12x + 4−4−12−313−220−115041−72−123−5420↓ evaluate: input given, read the output below it
Sample heights of f(x) = x3 − 12x + 4 at x = −4 to 4.
  1. List the change in height between neighbouring samples: −12→13 (rise), 13→20 (rise), 20→15 (fall), 15→4 (fall), 4→−7 (fall), −7→−12 (fall), −12→−5 (rise), −5→20 (rise).A turn shows up as a switch from rising to falling or from falling to rising, so we scan the sign of each change.
  2. At x=−2 the graph rises into it (13→20) and falls out of it (20→15). Mark x=−2 as a local maximum.Rising then falling identifies an isolated local maximum. Because the graph changes steadily between integers, every x in the open interval (−3, −1) other than −2 has f(x) < 20.
  3. At x=2 the graph falls into it (−7→−12) and rises out of it (−12→−5). Mark x=2 as a local minimum.Falling then rising identifies an isolated local minimum. On the open interval (1, 3), every x other than 2 has f(x) > −12.
  4. Check for flat stretches or shoulders. No two neighbouring samples are equal, and every other interior sample lies on a steady rise or a steady fall. So no other input qualifies under the nonstrict rule.The formal rule also allows equal nearby heights. Equal neighbours would be the only way to get an extra local extremum, and there are none here.
  5. Do not count x=−4 or x=4. f(4)=20 equals the local maximum value, but it is not a turn.The domain is all real numbers, so −4 and 4 are only the edges of the table, not ends of the graph. The cubic keeps falling to the left of −4 and keeps rising to the right of 4, so neither is a turn. Equal heights at different places do not create a new extremum.
  6. Read each location from the input column and each value from the output column.The location is the horizontal input b, the value is the vertical output f(b), and the point is (b, f(b)).
Answer
f has a local maximum value of 20 at x = −2, at the point (−2, 20). It has a local minimum value of −12 at x = 2, at the point (2, −12). Both are strict local extrema.
Double-check Use the derivative of the cubic: f′(x) = 3x2 − 12 = 3(x − 2)(x + 2), which is zero only at x = −2 and x = 2. f′ is positive for x < −2, negative for −2 < x < 2, and positive for x > 2. That confirms a maximum at −2 and a minimum at 2. Recompute the heights: f(−2) = −8 + 24 + 4 = 20 and f(2) = 8 − 24 + 4 = −12.

Work to write

  1. Heights rise from 13 to 20, then fall to 15, around x = −2, so x = −2 is a local maximum.
  2. Heights fall from −7 to −12, then rise to −5, around x = 2, so x = 2 is a local minimum.
  3. x = ±4 are only table edges; the domain is all reals and the graph is still falling or rising there, so they are not extrema.
  4. The local maximum value is 20 at x = −2, point (−2, 20).
  5. The local minimum value is −12 at x = 2, point (2, −12).

5 How

  1. Scan the given graph for changes from rising to falling or falling to rising.
  2. Rising then falling identifies an isolated local maximum; falling then rising identifies an isolated local minimum.
  3. Also inspect flat stretches and shoulders when applying the formal nonstrict rule. Compare both sides, including any equal heights.
  4. Require a domain neighborhood on both sides. An end of the graph is considered separately for absolute extrema.
  5. Read the horizontal input as the location and the vertical output as the value.
  6. Give value and location in a full sentence; give the ordered pair if the question asks for the point.

Find local values and locations

1
Does the graph rise and then fall?
YesThe turn gives a strict local maximum.
NoCheck for a valley or flat comparison.
↓
2
Does it fall and then rise?
YesThe turn gives a strict local minimum.
NoA flat stretch or shoulder can still satisfy the nonstrict inequalities; compare both sides.
↓
3
Are nearby domain points available on both sides of this input?
YesApply the local comparison within an open neighborhood.
NoThis text does not call the domain endpoint local; check it for an absolute extremum.
↓
4
Are you using a finite table of samples rather than a given full curve?
YesCall the result a sampled peak or dip unless additional information establishes the full nearby behavior.
NoRead the full given curve's neighborhood.
  1. Check changes of direction and any level stretches.
  2. Compare close domain points on both sides.
  3. Name maximum or minimum using the correct output comparison.
  4. Read the value and input separately and report both.

6 Worked example

Worked example

Reading a local maximum from a downward-opening parabola

The graph shows f(x) = -(x - 2)2 + 5, defined for all real numbers. Points read from the graph are (0, 1), (1, 4), (2, 5), (3, 4) and (4, 1). The curve keeps falling to the left of (0, 1) and to the right of (4, 1). Find every local maximum and local minimum of f. Give each value and location in a sentence, and give the point.

(0, 1)(1, 4)(2, 5)(3, 4)(4, 1)−2246−4−2246
Graph of f(x) = -(x - 2)2 + 5 through (0, 1), (1, 4), (2, 5), (3, 4) and (4, 1).
Show the steps and the answer
  1. Scan the graph from left to right. Through (0, 1) and (1, 4) up to (2, 5) the heights increase: 1, 4, 5. After x=2 they decrease: 5, 4, 1.A local extremum shows up where the graph changes from rising to falling or from falling to rising.
  2. The graph rises and then falls at x=2, so x=2 is an isolated local maximum. Nowhere does the graph fall and then rise, so there is no isolated local minimum.Rising then falling marks a local maximum. Falling then rising would mark a local minimum.
  3. Check for flat stretches or shoulders. There are none. The curve is never level over an interval, and the points on each side of x=2 are lower: f(1)=4<5 and f(3)=4<5. In general f(x)=5-(x-2)2<5 for every x≠2.The nonstrict rule asks us to compare both sides, including equal heights. Here every nearby height is strictly smaller, so the maximum is strict.
  4. Confirm the domain. f is defined for all real numbers, so an open interval such as (1, 3) lies around x=2 inside the domain. The graph has no endpoints.A local extremum needs domain values on both sides. Ends of a graph are considered separately, for absolute extrema.
  5. Read the coordinates of the turning point. The horizontal input is 2, which is the location. The vertical output is f(2)=5, which is the value.The location of an extremum is its input b, its value is f(b), and its point is (b, f(b)).
  6. State the result: f has a local maximum value of 5 at x=2, at the point (2, 5). f has no local minimum.The answer must give the value and the location in a full sentence, plus the ordered pair the question asks for.
Answer
f has a local maximum value of 5 at x = 2. The point is (2, 5). f has no local minimum.
Double-check f(2) = -(0)2 + 5 = 5. On the open interval (1, 3), f(x) = 5 - (x - 2)2 ≤ 5, with equality only at x = 2. For example, f(1.5) = 4.75 and f(2.5) = 4.75, and both are less than 5. Since f rises for x < 2 and falls for x > 2, it never falls and then rises, so there is no local minimum.

Work to write

  1. The graph rises up to x = 2 and falls after it.
  2. f(x) ≤ f(2) = 5 for every x in the open interval (1, 3).
  3. Local maximum value 5 at x = 2, point (2, 5).
  4. There is no local minimum.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 5

Read a hilltop

For f(x) = 3 − (x − 2)2, find the local maximum value, where it occurs, and its point.

Show the steps and the answer
23224−11234
The same picture again: y = −(x − 2)2 + 3.
  1. f(2)=3−02=3.Input 2 makes the square zero.
  2. At every different input, (x−2)2>0, so f(x) < 3.A nonzero real number has a positive square.
  3. The value is 3, the location is x=2, and the point is (2, 3).The ordered pair puts the input first and the output second.
Answer
  • Local maximum value: 3.
  • Location: x = 2.
  • Point: (2, 3).
Double-check f(1) = f(3) = 2 agrees with the pictured neighboring heights. The square argument supplies the proof for all nearby inputs.

Rung 2 of 5

Read a negative valley

Find the local minimum of f(x) = (x + 1)2 − 2.

Show the steps and the answer
−1−2−1−2−22
The same picture again: y = (x + 1)2 − 2.
  1. f(−1)=02−2=−2.x + 1 equals zero at input −1.
  2. Every other input gives (x+1)2>0, so its output exceeds −2.Adding a positive number to −2 raises it.
  3. Report local minimum of −2 at x=−1.The output supplies the value and the input supplies the location.
Answer
  • Local minimum of −2 at x = −1.
  • Point: (−1, −2).
Double-check f(−2) = −1 and f(0) = −1 are both above −2. The square comparison proves the result beyond these samples.

Rung 3 of 5

Reading a local maximum from a downward parabola with its vertex left of the y-axis

The graph shows h(x) = −(x + 1)2 + 3, defined for all real numbers. Points read from the graph are (−3, −1), (−2, 2), (−1, 3), (0, 2) and (1, −1). The curve keeps falling to the left of (−3, −1) and to the right of (1, −1). Find every local maximum and every local minimum of h. For each local maximum, state its value and its location, and give the point.

(−3, −1)(−2, 2)(−1, 3)(0, 2)(1, −1)−4−22−6−4−224
Downward parabola h(x) = −(x + 1)2 + 3 with the marked points (−3, −1), (−2, 2), (−1, 3), (0, 2) and (1, −1).
Show the steps and the answer
  1. Scan the listed points from left to right: −1, 2, 3, 2, −1. The heights rise from x=−3 to x=−1 and fall from x=−1 to x=1.A local extremum shows up where the graph changes direction, so the first job is to find where rising turns into falling or falling turns into rising.
  2. The only change of direction is at x=−1, where the graph goes from rising to falling. This marks a local maximum at x=−1.Rising then falling identifies an isolated local maximum.
  3. Look for flat stretches or shoulders. There are none: the heights on the left of x=−1 (for example h(−2)=2) and on the right (for example h(0)=2) are both below 3, and the curve keeps falling beyond the listed points.The nonstrict rule h(b) ≥ h(x) must be checked on both sides of b, and any equal nearby heights would also need attention.
  4. Confirm with the formula: h(x)=−(x+1)2+3≤3=h(−1) for every x, because (x+1)2≥0. Equality holds only at x=−1, so the open interval (−2,0) works, and the maximum is strict.The formal definition needs h(b) ≥ h(x) for all x in some open interval around b. It is strict when the inequality is > for every nearby x ≠ b.
  5. Check for a local minimum. The graph never changes from falling to rising, and the domain is all real numbers, so there are no endpoints to consider. h has no local minimum.A local minimum needs falling then rising, with domain on both sides of the point. There is no such place.
  6. Read the location from the horizontal axis, x=−1, and the value from the vertical axis, h(−1)=3. The point is (−1,3).The location is the input b, the value is the output h(b), and the point is (b, h(b)).
Answer
h has a local maximum value of 3, located at x = −1, so the point is (−1, 3). This local maximum is strict. h has no local minimum.
Double-check Substitute x = −1: h(−1) = −(0)2 + 3 = 3. Neighbors: h(−1.5) = −0.25 + 3 = 2.75 and h(−0.5) = −0.25 + 3 = 2.75. Both are less than 3, which agrees with a strict local maximum at x = −1. A downward parabola (leading coefficient −1) has exactly one turning point, its vertex (−1, 3), so no local minimum is missing.

Work to write

  1. The graph rises up to x = −1 and falls after x = −1.
  2. h(x) = −(x + 1)2 + 3 ≤ 3 = h(−1) for all x, so h(−1) ≥ h(x) on an open interval around −1.
  3. h has a local maximum value of 3 at x = −1.
  4. The point is (−1, 3).
  5. h has no local minimum.

Rung 4 of 5

Proving a local minimum with the inequality on a restricted domain

A function is given by f(x) = (x − 1)2 − 4, defined only on the interval [−2, 3]. Points read from its graph are (−2, 5), (−1, 0), (0, −3), (1, −4), (2, −3) and (3, 0). The graph falls from x = −2 to x = 1 and rises from x = 1 to x = 3. Find every local maximum and local minimum of f. Justify each one with the defining inequality and say whether it is strict. For each, give the value, the location and the point.

(-2, 5)(-1, 0)(0, -3)(1, -4)(2, -3)(3, 0)−224−4−2246
The parabola y = (x − 1)2 − 4 with the points (−2, 5), (−1, 0), (0, −3), (1, −4), (2, −3) and (3, 0) marked.
Show the steps and the answer
  1. Scan the listed heights 5, 0, −3,−4,−3, 0. The graph falls from (−2,5) down to (1,−4) and then rises to (3, 0).Local extrema can only occur where the graph changes direction, or along flat stretches.
  2. The only change of direction is at x=1, from falling to rising. Mark x=1 as a candidate local minimum. The graph has no change from rising to falling, so there is no interior candidate for a local maximum.Falling then rising signals a local minimum. Rising then falling would signal a local maximum.
  3. Apply the inequality at b=1. f(x)−f(1)=[(x−1)2−4]−(−4)=(x−1)2. This is ≥0, so f(1)≤f(x) for every x in the open interval (0, 2), which lies inside [−2,3].A local minimum at b requires f(b) ≤ f(x) for all x in some open interval around b that is contained in the domain.
  4. Check strictness. For x≠1,(x−1)2>0, so f(1) < f(x) for every nearby x≠1. There is no flat stretch or equal nearby height.A strict local minimum needs < for every nearby x other than b. A square of a nonzero number is positive.
  5. Examine the ends x=−2 and x=3. No open interval around either end fits inside [−2,3], because the domain stops on one side. So neither end is a local maximum or local minimum. The height f(−2)=5 is handled separately, as the absolute maximum.The local definition requires domain points on both sides of b. Ends are considered only for absolute extrema.
  6. Read the horizontal input as the location: 1. Read the vertical output as the value: f(1)=(1−1)2−4=−4.The location is b, the value is f(b), and the point is (b, f(b)).
Answer
f has a strict local minimum of value −4 at x = 1, at the point (1, −4). f has no local maximum. The endpoints x = −2 and x = 3 are not local extrema because no open interval around them lies in the domain.
Double-check Test inputs near 1 on both sides. f(0.9) = (−0.1)2 − 4 = −3.99 and f(1.1) = (0.1)2 − 4 = −3.99. Both are greater than −4, as the strict inequality predicts. The listed points (0, −3) and (2, −3) are also higher than (1, −4).

Work to write

  1. Graph falls on [−2, 1] and rises on [1, 3]; the only turn is at x = 1.
  2. f(x) − f(1) = (x − 1)2 ≥ 0, so f(1) ≤ f(x) on (0, 2) ⊂ [−2, 3].
  3. (x − 1)2 > 0 for x ≠ 1, so the local minimum is strict.
  4. x = −2 and x = 3 are endpoints: no open interval in the domain on both sides, so they are not local extrema.
  5. Local minimum value −4 at x = 1; point (1, −4). No local maximum.

Rung 5 of 5

Checking a claimed peak against both neighbors

A graph of y = f(x) on the domain [-3, 3] is made of straight segments joining these points, read from the graph: (-3, 1), (-2, 4), (-1, 6), (0, 3), (1, -2), (2, 0), (3, 5). A student says f has a local maximum at x = -2 because f(-2) = 4 is higher than f(-3) = 1. Check this claim by comparing heights on both sides. Then find every local maximum and local minimum of f, giving the value and location of each in a sentence and the point as an ordered pair.

input xoutput f(x)−31−24−16031−22035
x:−3,−2,−1,0,1,2,3;f(x):1,4,6,3,−2,0,5.
Show the steps and the answer
input xoutput f(x)−31−24−16031−22035↓ evaluate: input given, read the output below it
Heights read from the graph of f at x = -3 through 3. The peak f(-1) = 6 and the valley f(1) = -2 are marked.
  1. Check the student's claim at x=-2 on the left side. f(-3)=1<4=f(-2), and the segment from x=-3 to x=-2 rises.A local maximum needs f(-2) ≥ f(x) for all nearby x on both sides. The left side passes.
  2. Check the right side of x=-2. f(-1)=6>4. Every x just to the right of -2 lies on the rising segment toward (-1,6), so f(x) > 4 there.One branch higher than f(-2) in every open interval around -2 breaks the condition f(-2) ≥ f(x). The claim fails, and x = -2 is not a local maximum.
  3. Scan the heights in order: 1, 4, 6, 3, -2, 0, 5. The graph rises to x=-1, falls to x=1, then rises to x=3.Local extrema in the interior sit where rising changes to falling or falling changes to rising.
  4. At x=-1, the left neighbor is f(-2)=4<6 and the right neighbor is f(0)=3<6. Both segments slope down away from (-1,6), so f(x) < 6 for nearby x≠-1.Rising then falling, checked on both branches, gives a strict local maximum.
  5. At x=1, the left neighbor is f(0)=3>-2 and the right neighbor is f(2)=0>-2. Both segments slope up away from (1,-2), so f(x)>-2 for nearby x≠1.Falling then rising, checked on both branches, gives a strict local minimum.
  6. Exclude x=-3 and x=3.They are ends of the domain, so no open interval around them lies inside the domain. Ends are considered separately, for absolute extrema.
  7. Read each location from the horizontal input and each value from the vertical output.The location is b, the value is f(b), and the point is (b, f(b)).
Answer
The student's claim is wrong: f(-1) = 6 > f(-2) = 4, so x = -2 is not a local maximum. f has a local maximum value of 6 at x = -1, at the point (-1, 6). It has a local minimum value of -2 at x = 1, at the point (1, -2). There are no other local extrema.
Double-check Each candidate is compared with its neighbor on both sides. For x = -1: 4 < 6 and 3 < 6. For x = 1: 3 > -2 and 0 > -2. At x = -2, the right neighbor 6 is greater than 4, so that point fails. The domain ends at x = -3 and x = 3 are left out because they have no neighbors on both sides.

Work to write

  1. Right of x = -2: f(-1) = 6 > 4, so x = -2 is not a local max
  2. f(-2) = 4 < 6 and f(0) = 3 < 6, so local max at x = -1
  3. Local maximum value 6 at x = -1, point (-1, 6)
  4. f(0) = 3 > -2 and f(2) = 0 > -2, so local min at x = 1
  5. Local minimum value -2 at x = 1, point (1, -2)
  6. x = -3 and x = 3 are domain ends, not local extrema

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: The local maximum is x = 1.
1 is the input location. In the worked graph the local maximum output is 2.
✓ Instead: Write 'local maximum of 2 at x = 1'; its point is (1, 2).
✗ Not this: Every local maximum must exceed every local minimum.
The supplied rational graph has local maximum about −1.633 on one branch and local minimum about 1.633 on another.
✓ Instead: Local comparisons use one neighborhood, so extrema in different neighborhoods need not be ordered that way.
✗ Not this: Every formal local extremum must be an isolated turn.
The nonstrict rule allows equality. Every interior input of a constant function satisfies both local comparisons.
✓ Instead: Distinguish nonstrict local extrema from strict hilltops, valleys, and isolated turning points.
✗ Not this: Two lower sampled neighbors prove a local maximum.
Unmeasured inputs between those samples could have higher outputs.
✓ Instead: Use a given full curve or an argument applying to every nearby input; use finite samples as supporting checks.

Tips and tricks

  • For a value, give the output and where it occurs. For a point, give (input, output).
  • In this text, a domain endpoint without nearby domain points on both sides is not local. It can still be an absolute extremum.
  • Graph exercises commonly emphasize isolated turns. If a flat stretch appears, state how the nonstrict rule treats it.
  • Only the TI-30XIIS is allowed on this course's exam. Read the supplied graph; nearby substitutions support a suspected shape but cannot certify all turns.

7 The trap

Saying 'the maximum is x = 1'. The maximum is the output, 2; x = 1 is where it happens. On the exam write both: 'local maximum of 2 at x = 1'.

Lesson 10 of 10

Toolkit behavior, then absolute maximum and minimum

Subject: MathematicsLeans on: Toolkit functionsLeans on: Domain and endpoint notationLeans on: Squaring negativesOpen this lesson
Leans on 0Comes back in 0

1 In plain words

Think of a toolkit as the basic shapes you keep in a drawer. First, revisit the nine toolkit functions from section 1.1 and see where each graph rises, falls, or stays level. Then widen your view from one nearby hill to an entire mountain range. An absolute maximum is the greatest output the function actually reaches anywhere in its domain, the full set of accepted inputs. An absolute minimum is the least reached output. One value can occur at several inputs. Some graphs have no highest or lowest output. An included endpoint can supply the record height even when it is not a local turning point.

2 Picture it

(0, 4)absolute max 4−22−8−44(0, 4)absolute max 4−3−2−1123−8−44
y = 4 − x2 reaches its greatest output 4 at the vertex; its formula continues downward without a least output.

Reminder

Reminder: Toolkit functions, Domain and endpoint notation, Squaring negatives
  • Toolkit functions. The nine toolkit functions are the basic graphs introduced in section 1.1. Each part above gives its formula, domain, behavior, and example.
  • Domain and endpoint notation. [−3, 2] includes both ends; [0, 2) includes 0 and excludes 2. A filled dot is included and a hollow dot is excluded.
  • Squaring negatives. (−3)2 = 9, so 4 − (−3)2 = 4 − 9 = −5.

3 Why

Absolute compares with every accepted input, so you must inspect the whole domain and check whether the proposed height is reached. A continuing graph may have outputs without any upper or lower bound. It may also approach a bound without reaching it, as 1x2 approaches 0 but stays positive. Neither case supplies a missing extremum. Finite domain endpoints deserve attention because their heights can exceed all interior heights. A highest or lowest included output remains the record even if it occurs at more than one input.

4 The rule

The absolute maximum is f(c) when c is in the domain and f(c)≥f(x) for every domain input x.

The absolute minimum is f(d) when d is in the domain and f(d)≤f(x) for every domain input x. The value must be attained, meaning reached at an accepted input.

The same idea, five ways

Say it

'absolute maximum of 4 at x = 0' or 'there is no absolute minimum.'

Write it

'The greatest output over the entire domain is 4, attained at input 0.'

In math
  • Absolute maximum: f(c) ≥ f(x) for every x in the domain, with c in the domain.
  • Absolute minimum: f(d) ≤ f(x) for every x in the domain, with d in the domain.
  • Maximum value: f(c).
  • Maximum point: (c, f(c)).
Like

The highest or lowest reached elevation in the entire mapped region.

See it
height 4, input 0−22−6−4−224
The output 4 is the record height; input 0 locates where it is reached.

The same idea, other ways

Neighborhood versus whole region

A local record compares nearby heights on both sides; an absolute record compares every height in the whole domain. An included endpoint can be absolute without being local under this text's two-sided definition. Flat ties can satisfy the nonstrict local definition without isolated turns.

Local: nearby on both sides

Absolute: every domain input

Value: output height

Location: input

A bound plus a reached value

For f(x) = 4 − x2, no output exceeds 4 because x2 ≥ 0. The input 0 actually gives 4. These two facts together prove the absolute maximum. Large squares make outputs arbitrarily negative, so there is no absolute minimum.

maximum reached−22−8−44
Height 4 is an upper bound reached at input 0; the stated formula has no lower bound.
With recorded costs

If the domain is exactly the eight recorded years, the gas table has greatest output $3.68 at 2012 and least output $2.31 at 2005. Those are absolute extrema of this finite data function. The table alone cannot give absolute extrema of all unmeasured prices between those years.

input yearoutput price ($ per gallon)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68↓ evaluate: input given, read the output below it
Compare every column in the specified finite domain before naming its absolute records.
An approached low height can be unattained

For 1x2, x = 10 gives 0.01 and x = 100 gives 0.0001. Both are positive. Every allowed x gives a positive output, and doubling its size divides the output by 4. There is always a smaller output, but 0 is never reached, so there is no absolute minimum.

−4−22424
Both branches approach height 0 without touching it; approaching a height is different from attaining it.
Toolkit functionIncreasing onDecreasing onAbsolute max and min
Constant, f(x) = cnonenone (constant on (−∞, ∞))c is both
Identity, f(x) = x(−∞,∞)noneneither
Absolute value, f(x) = |x|(0, ∞)(−∞,0)min 0 at x = 0; no max
Quadratic, f(x) = x2(0, ∞)(−∞,0)min 0 at x = 0; no max
Cubic, f(x) = x3(−∞,∞)noneneither
Reciprocal, f(x) = 1xnone(−∞, 0) and (0, ∞), each piece separatelyneither
Reciprocal squared, f(x) = 1x2(−∞,0)(0, ∞)neither
Square root, f(x) = x(0, ∞)nonemin 0 at x = 0; no max
Cube root, f(x) = x3(−∞,∞)noneneither

Part 1: Toolkit 1: constant function, f(x) = c

A constant function acts like a shelf at one fixed height. The letter c stands for a fixed number, such as 3, rather than an input you change. All inputs get that same output.

  • Formula: f(x) = c.
  • Domain: (−∞, ∞). Range: {c}.
  • Increasing: none. Decreasing: none. Constant on (−∞, ∞).
  • Under the formal nonstrict definition, c is both a local maximum and a local minimum at every input.
  • There are no strict local extrema or isolated turning points.
  • Absolute maximum and minimum: c, both attained at every input.
−2241234
For c = 3, every accepted input gives height 3.

Worked example

A decreasing line with one excluded endpoint

Let f(x) = -12x + 3 with domain -4 ≤ x < 2. On the graph, the left end at (-4, 5) is a filled dot, so x = -4 is included. The right end at (2, 2) is a hollow dot, so x = 2 is excluded. The graph is a straight line segment with slope -12 and no other pieces. Find the absolute maximum and the absolute minimum of f, if they exist. For each one, give the value and every input where it occurs.

(-4, 5) filled(2, 2) hollow−6−4−2246−2246
Graph of f(x) = -12x + 3, shown only on the domain -4 ≤ x < 2.
Show the steps and the answer
  1. Identify the toolkit shape. f(x)=-12x+3 is a linear function with slope -12 and y-intercept 3.A line has no turning point, so its highest and lowest heights on a restricted domain can only occur at the ends of that domain.
  2. Read the whole stated domain, -4≤x<2. The input x=-4 is included because it has a filled dot. The input x=2 is excluded because it has a hollow dot. There are no arrows, so the graph stops at both ends.The domain decides which inputs are accepted. The edges of the viewing window do not set the domain. The stated inequality and the dots do.
  3. Check for unbounded behavior. The domain is a bounded interval and the line is continuous on it. So the outputs neither rise nor fall without bound.Unbounded outputs would rule out an extremum. Here that does not happen, so we compare the heights at the ends.
  4. Compute the end heights. f(-4)=-12(-4)+3=2+3=5. f(2)=-12(2)+3=-1+3=2.These are the only candidates for extremes on a line segment.
  5. The slope is negative, so f decreases from left to right. For every accepted x with -4≤x<2, we have 2<f(x)≤5.A decreasing function gives its greatest output at the leftmost input and smaller outputs further right.
  6. Maximum: the height 5 is reached at the included input x=-4, and f(x)≤5 for every domain input. No other input gives 5, because the line is strictly decreasing.An absolute maximum must be attained at an accepted input and be at least as large as every output.
  7. Minimum: the height 2 occurs only at x=2, which is excluded. Every accepted input gives an output strictly greater than 2. For any reached height f(x0)>2, an input slightly to the right of x0 gives a smaller output.A height that is approached but never reached is not an absolute extremum. No reached output is the least, so there is no absolute minimum.
Answer
  • Absolute maximum: 5
  • Location: x = -4
  • Absolute minimum: none (outputs approach 2 but never reach it, since x = 2 is excluded)
Double-check Test some accepted inputs. f(0) = 3 and f(1.9) = 2.05. Both are at most 5 and both are greater than 2. So 5 is reached and is never exceeded. Inputs closer to 2 give outputs closer to 2 without ever equaling it, which confirms there is no minimum.

Work to write

  1. f(-4) = -12(-4) + 3 = 5, and x = -4 is included (filled dot)
  2. f(2) = 2, but x = 2 is excluded (hollow dot), so 2 is not attained
  3. Slope -12 < 0, so f decreases on -4 ≤ x < 2
  4. Absolute maximum: 5
  5. Location: x = -4
  6. Absolute minimum: none

Part 2: Toolkit 2: identity function, f(x) = x

The identity function returns whatever number you give it, like a machine that sends your card back unchanged. A larger input is therefore always a larger output.

  • Formula: f(x) = x.
  • Domain and range: (−∞, ∞).
  • Increasing on (−∞, ∞); decreasing and constant on no interval.
  • No local maximum or minimum, and no absolute maximum or minimum.
−4−224−4−224
The identity line rises across its whole real-number domain.

Worked example

Read a graph with no records

For f(x) = x, give behavior and extrema.

Show the steps and the answer
−4−224−4−224
The line continues through higher and lower outputs than any candidate record.
  1. The output increases whenever the input increases.Output and input are the same number.
  2. There are no local turns or flat comparisons and no absolute record heights.Every input has a larger-output neighbor on one side and smaller-output neighbor on the other; the domain extends without bound.
Answer
  • Increasing on (−∞, ∞).
  • Decreasing: none.
  • Constant: none.
  • Local maximum: none.
  • Local minimum: none.
  • Absolute maximum: none.
  • Absolute minimum: none.
Double-check Between inputs −2 and 3 the average rate is 3−(−2)3−(−2) = 1.

Part 3: Toolkit 3: absolute value function, f(x) = |x|

Absolute value gives a number's distance from zero, so neither a leftward nor a rightward distance is negative. As you approach zero from the left, that distance shrinks; after passing zero toward the right, it grows.

  • Formula: f(x) = |x|.
  • Domain: (−∞, ∞). Range: [0, ∞).
  • Decreasing on (−∞, 0); increasing on (0, ∞); constant on no interval.
  • Local and absolute minimum: 0 at x = 0.
  • No local maximum and no absolute maximum.
minimum−4−22424
The distance reaches 0 at the vertex and grows without bound outward.

Worked example

Read the V shaped toolkit graph

For f(x) = |x|, give its behavior and absolute extrema.

Show the steps and the answer
302−4−22424
The absolute-value minimum also satisfies the local valley comparison.
  1. The outputs at −3, 0, and 2 are 3, 0, and 2.Absolute value is distance from zero.
  2. Report decreasing before 0 and increasing after 0.The picture falls toward its vertex, then rises.
  3. The absolute minimum is 0 at x=0; there is no maximum.Distances cannot be negative, zero is reached, and arbitrarily large inputs have arbitrarily large distances.
Answer
  • Decreasing on (−∞, 0).
  • Increasing on (0, ∞).
  • Absolute minimum: 0 at x = 0.
  • Absolute maximum: none.
Double-check |−3| = 3 and |3| = 3 are both above |0| = 0.

Part 4: Toolkit 4: quadratic function, f(x) = x2

Squaring measures the area of a square made from a length's size. Both positive and negative inputs give nonnegative outputs. The graph is a parabola with vertex at (0, 0).

  • Formula: f(x) = x2.
  • Domain: (−∞, ∞). Range: [0, ∞).
  • Decreasing on (−∞, 0); increasing on (0, ∞); constant on no interval.
  • Local and absolute minimum: 0 at x = 0.
  • No local maximum and no absolute maximum.
(0, 0)−22246810
The parabola's vertex is its lowest output.

Worked example

Absolute extrema of a downward parabola on a half-open domain

Let f(x) = -(x - 2)2 + 5 with domain -1 < x ≤ 4. On the graph, the left end at (-1, -4) is a hollow dot, so x = -1 is excluded. The right end at (4, 1) is a filled dot, so x = 4 is included. The vertex is at (2, 5). Find the absolute maximum and the absolute minimum of f, if they exist, and state every input where each occurs.

(2, 5) vertex(-1, -4) hollow(4, 1) filled−2246−44
Graph of f(x) = -(x - 2)2 + 5, a downward parabola with vertex (2, 5).
Show the steps and the answer
  1. Identify the toolkit shape. f(x)=-(x-2)2+5 is the squaring toolkit function reflected over the x-axis, shifted right 2 and up 5. It is a downward-opening parabola with vertex (2, 5).The squaring toolkit accepts every real input. Only the stated domain restricts this problem, so we know where the vertex is and which way the parabola opens.
  2. Read the whole stated domain: -1<x≤4. The domain has no arrows and is bounded on both sides. The left endpoint x=-1 is excluded (hollow dot). The right endpoint x=4 is included (filled dot).Absolute extrema are judged only over the domain. The edges of the viewing window are not the domain endpoints; the stated inequality is.
  3. Check for unbounded behavior. The domain is a bounded interval, and every output satisfies f(x)≤5. Outputs do not rise without bound, and they do not fall without bound.Unbounded rising rules out an absolute maximum, and unbounded falling rules out an absolute minimum. Neither happens here, so we must check reached heights.
  4. Check the excluded endpoint. Compute f(-1)=-(-3)2+5=-9+5=-4. Since x=-1 is not in the domain, the height -4 is approached as x moves toward -1 from the right but is never reached.A height that is approached but never attained cannot be an absolute extremum.
  5. Find the greatest reached output. For every x, (x-2)2≥0, so -(x-2)2+5≤5. Equality holds only when x=2, and x=2 is in the domain. So f(2)=5≥f(x) for every domain input, and no other input gives 5.The absolute maximum must be attained at an accepted input and must be at least every other output. The vertex meets both conditions.
  6. Look for the least reached output. The other candidate is the included endpoint: f(4)=-(2)2+5=1. But inputs near -1 give outputs below 1. For example, f(-0.9)=-8.41+5=-3.41 and f(-0.99)=-8.9401+5=-3.9401.Every candidate for the absolute minimum must be compared with all domain inputs, not just the endpoints and the vertex.
  7. Conclude about the minimum. Every reached output is greater than -4. For any reached value y>-4, an input closer to -1 gives a smaller output. So no reached output is less than or equal to all others.The only possible floor, -4, is unattained. Without an attained least height, there is no absolute minimum.
Answer
  • Absolute maximum: 5, occurring only at x = 2.
  • Absolute minimum: none, because outputs approach -4 as x approaches the excluded endpoint x = -1, but -4 is never reached.
Double-check f(2) = -(0)2 + 5 = 5. Any other domain input makes (x - 2)2 > 0, so f(x) < 5, which confirms the maximum is unique. f(4) = 1 is reached but is not the least output, since f(-0.99) = -3.9401 < 1. The outputs keep decreasing toward -4 as x approaches -1, and -4 is not attained because x = -1 is excluded.

Work to write

  1. f is a downward parabola with vertex (2, 5); domain -1 < x ≤ 4
  2. f(-1) = -4, but x = -1 is excluded, so -4 is not attained
  3. (x - 2)2 ≥ 0 ⇒ f(x) ≤ 5, with equality only at x = 2
  4. Absolute maximum value: 5
  5. Location: x = 2
  6. f(4) = 1 is not least, since outputs near x = -1 are smaller
  7. No absolute minimum (the height -4 is approached, never reached)

Part 5: Toolkit 5: cubic function, f(x) = x3

A cubic returns an input multiplied by itself three times. Negative inputs give negative outputs, zero gives zero, and positive inputs give positive outputs. Moving right keeps raising the output even where the middle of the graph looks nearly level.

  • Formula: f(x) = x3.
  • Domain and range: (−∞, ∞).
  • Increasing on (−∞, ∞); decreasing and constant on no interval.
  • No local maximum or minimum and no absolute maximum or minimum.
  • The flattened appearance at 0 is not a turn or a constant interval.
−22−8−448
The cubic rises through its flattened middle without a local turn.

Worked example

A wide, flat-looking parabola reaches its lowest height at exactly one input

Let f(x) = 14(x + 1)2 + 2 with domain -4 < x ≤ 5. The graph is a wide upward-opening parabola with vertex (-1, 2). Near the vertex the curve looks almost flat: f(-2) = 2.25 and f(0) = 2.25. The left end at (-4, 4.25) is a hollow dot, so x = -4 is excluded. The right end at (5, 11) is a filled dot, so x = 5 is included. Find the absolute maximum and the absolute minimum of f, if they exist, and give every input where each one occurs.

(-1, 2)(-4, 4.25) hollow(5, 11) filled(-2, 2.25)(0, 2.25)−4−22462610
Graph of f(x) = ¼(x + 1)2 + 2 on -4 < x ≤ 5.
Show the steps and the answer
  1. Identify the toolkit shape: f is the squaring function x2, shifted 1 left and 2 up, and vertically compressed by a factor of 14. It opens upward with vertex (-1,2).The form a(x - h)2 + k with a = 14 > 0, h = -1, k = 2 gives an upward parabola whose vertex is (h, k).
  2. Read the whole domain -4<x≤5. It is a bounded interval. The left end x=-4 is excluded (hollow dot). The right end x=5 is included (filled dot). Nothing continues past either end.Only accepted inputs can produce an extreme value, so the endpoint dots must be checked first.
  3. Check for unbounded behavior. The domain is bounded, so the outputs cannot rise or fall without bound. Both extrema may exist.An extremum fails to exist by unboundedness only when the outputs grow or drop without limit on the domain.
  4. Find the lowest height. The vertex input x=-1 lies in the domain, and f(-1)=14(0)2+2=2. Solve f(x)=2: 14(x+1)2=0, so x=-1 is the only solution. The nearby values f(-2)=2.25 and f(0)=2.25 are close to 2 but larger.The flat look near the vertex is misleading. The squared term is positive for every x ≠ -1, so the height 2 is reached at one input only.
  5. Find the highest height. f increases as x moves away from -1. To the left, the farthest input is near -4, where f(-4)=14(9)+2=4.25, but x=-4 is excluded. To the right, f(5)=14(36)+2=9+2=11, and x=5 is included.The hollow dot at height 4.25 is not attained, but it is lower than 11 anyway. The included right end gives the greatest output.
  6. Compare the reached heights across the whole domain. Every output satisfies 2≤f(x)≤11. The value 2 occurs only at x=-1. The value 11 occurs only at x=5, because the only other solution of (x+1)2=36 is x=-7, which lies outside the domain.An absolute extremum is a reached output that is at least as large (or at least as small) as every other output, together with all inputs that produce it.
Answer
  • Absolute maximum: 11
  • Occurs at x = 5
  • Absolute minimum: 2
  • Occurs at x = -1 (only)
Double-check Test several domain inputs. f(-3.9) ≈ 4.10, f(-2) = 2.25, f(-1.1) = 2.0025, f(0) = 2.25, f(3) = 6, f(5) = 11. All of them lie between 2 and 11. Only x = -1 gives exactly 2. Solving f(x) = 11 gives x = 5 or x = -7, and -7 is outside -4 < x ≤ 5. The excluded end height 4.25 is never reached, and it is below 11.

Work to write

  1. Vertex (-1, 2) is in the domain; f(-1) = 2
  2. 14(x + 1)2 = 0 only when x = -1, so the minimum is reached once
  3. f(-4) = 4.25 is not attained (hollow dot)
  4. f(5) = 11 is attained (filled dot)
  5. Absolute maximum: 11 at x = 5
  6. Absolute minimum: 2 at x = -1

Part 6: Toolkit 6: reciprocal function, f(x) = 1x

The reciprocal is one divided by the input. You cannot feed in zero, so the graph has separate branches on the two sides of zero. Each branch falls when you read it toward the right.

  • Formula: f(x) = 1x.
  • Domain and range: (−∞, 0) ∪ (0, ∞).
  • Decreasing on (−∞, 0) and on (0, ∞), separately; increasing and constant on no interval.
  • No local maximum or minimum and no absolute maximum or minimum.
  • It is not decreasing across the entire disconnected domain: f(−1) = −1 < 1 = f(1).
−4−224−4−224
Both branches fall; the gap at 0 must remain a split in behavior answers.

Worked example

Decrease separately on both sides

Read behavior and extrema of f(x) = 1x.

Show the steps and the answer
−4−224−4−224
Read each continuous branch independently.
  1. The negative branch moves from f(−2)=−12 to f(−1)=−1.These input moves stay on one side of the excluded zero.
  2. The positive branch moves from f(1)=1 to f(2)=12.The larger positive denominator gives a smaller positive fraction.
  3. Report two decreasing intervals and no extrema.Each branch strictly falls, and outputs are unbounded above and below near the missing input.
Answer
  • Decreasing on (−∞, 0) and on (0, ∞), each separately.
  • Increasing: none.
  • Constant: none.
  • Local maximum: none.
  • Local minimum: none.
  • Absolute maximum: none.
  • Absolute minimum: none.
Double-check The across-branch pair −1 and 1 fails a decreasing comparison, so 'separately' matters.

Part 7: Toolkit 7: reciprocal squared, f(x) = 1x2

First square the input, then divide one by that square. Every allowed output is positive. Farther from zero gives a larger denominator and a smaller output. Closer to zero gives a smaller denominator and a larger output.

  • Formula: f(x) = 1x2.
  • Domain: (−∞, 0) ∪ (0, ∞). Range: (0, ∞).
  • Increasing on (−∞, 0); decreasing on (0, ∞); constant on no interval.
  • No local maximum or minimum.
  • No absolute minimum: outputs approach 0 without reaching it. No absolute maximum: outputs grow without bound near excluded input 0.
−4−22424
The two positive branches never touch height 0 and rise without bound near input 0.

Worked example

A reciprocal curve that sinks toward 0 but never gets there

Let f(x) = 6x with domain x ≥ 1. The left end at (1, 6) is a filled dot, so x = 1 is included. The domain continues to the right without end. A graphing window happens to stop at x = 12, but that is only where the picture is cut off. A table of values gives: f(1) = 6, f(2) = 3, f(3) = 2, f(6) = 1, f(12) = 12. Find the absolute maximum and the absolute minimum of f, if they exist, and give every input where each occurs.

input xoutput f(x) = 6/x16233261120.5
x: 1, 2, 3, 6, 12; f(x) = 6/x: 6, 3, 2, 1, 0.5.
Show the steps and the answer
input xoutput f(x) = 6/x16233261120.5↓ evaluate: input given, read the output below it
Values of f(x) = 6/x on its domain x ≥ 1.
  1. Identify the toolkit shape: f(x)=6x is the reciprocal function 1x stretched vertically by 6. On x > 0 it is decreasing, and as x grows its outputs approach 0 from above.Knowing the toolkit behavior tells us how the graph acts beyond the plotted points. The reciprocal has the horizontal asymptote y = 0.
  2. Read the whole stated domain: x≥1. It has an included left endpoint at x=1 and no right endpoint. The window edge at x=12 is not a domain endpoint.Extrema are judged over every accepted input, not only the visible part of the graph.
  3. Check for unbounded behavior. On x≥1 the outputs never rise above 6 and never fall below 0, so they are bounded in both directions.An absolute maximum is ruled out automatically only when the outputs rise without bound. An absolute minimum is ruled out automatically only when they fall without bound.
  4. Test the low level 0. Solving 6x=0 has no solution, because a fraction with numerator 6 is never 0. For any input a, the larger input 2a gives f(2a)=3a, which is less than f(a). So no reached output is the least.A height that is approached but never attained is not an absolute minimum. Since every reached height has a lower one beyond it, there is no absolute minimum.
  5. Find the greatest reached output. f is decreasing on the domain, so the largest output occurs at the smallest accepted input, x=1. That output is f(1)=6. Every other input x > 1 gives 6x<6, so x=1 is the only location.The value 6 is attained at the included endpoint, and f(1) ≥ f(x) for every x ≥ 1.
Answer
  • Absolute maximum: 6, at x = 1 only.
  • Absolute minimum: none. The outputs approach 0 but never reach it.
Double-check f(1) = 61 = 6, and the filled dot means x = 1 is allowed. Large inputs give f(600) = 1100 and f(6000) = 11000, which are positive and shrinking but never 0. So 0 is never attained, and no positive output is the smallest.

Work to write

  1. Domain x ≥ 1 continues right; the window edge at x = 12 is not an endpoint
  2. f is decreasing on x ≥ 1 and approaches 0
  3. 6x = 0 has no solution, so 0 is not attained
  4. Absolute maximum: 6
  5. Location of maximum: x = 1
  6. Absolute minimum: none

Part 8: Toolkit 8: square root function, f(x) = x

A square root asks which nonnegative number squares to the input. For example, 4 = 2 because 22 = 4. Real inputs must be nonnegative. The graph starts at the included input zero and rises toward the right.

  • Formula: f(x) = x.
  • Domain and range: [0, ∞).
  • This function increases throughout its domain; the course reports the open increasing interval (0, ∞).
  • Decreasing and constant on no interval.
  • Absolute minimum: 0 at x = 0. No absolute maximum.
  • The included endpoint 0 is not a local minimum under this text's two-sided-neighborhood definition.
included end246810−11234
The domain starts at included input 0, while the course reports increasing on (0, ∞).

Worked example

An endpoint can be an absolute minimum

For f(x) = x, give the domain, the course's increasing interval, and extrema.

Show the steps and the answer
included min23246810−11234
The lowest attained height is at the domain's included endpoint.
  1. Accept inputs x≥0, including 0.A real square cannot be negative and 0 = 0.
  2. Read the rising graph and report increasing on (0,∞).The course uses open behavior intervals, while the actual function also increases when comparing 0 with a larger input.
  3. The absolute minimum is 0 at x=0; there is no absolute maximum or local extremum.All outputs are nonnegative, 0 is reached, the curve continues upward without bound, and its endpoint has no domain neighborhood on both sides.
Answer
  • Domain: [0, ∞).
  • Increasing interval reported in this course: (0, ∞).
  • Absolute minimum: 0 at x = 0.
  • Absolute maximum: none.
  • Local maximum: none.
  • Local minimum: none.
Double-check 0 = 0, 4 = 2, and 9 = 3 are increasing, including the accepted endpoint.

Part 9: Toolkit 9: cube root function, f(x) = x3

A cube root reverses cubing. It asks what real number multiplied by itself three times makes the input. Negative inputs work: ∛(−8) = −2 because (−2)3 = −8. Larger inputs give larger cube roots.

  • Formula: f(x) = x3.
  • Domain and range: (−∞, ∞).
  • Increasing on (−∞, ∞); decreasing and constant on no interval.
  • No local maximum or minimum and no absolute maximum or minimum.
−8−448−22
The cube-root curve rises through every real input.

Worked example

Negative cube roots are accepted

Read behavior and extrema of f(x) = x3.

Show the steps and the answer
input xoutput ∛x−8−200273
Each output cubes back to its input, including the negative input.
  1. f(−8)=−2,f(0)=0, and f(27)=3.(−2)3 = −8, 03 = 0, and 33 = 27.
  2. Report increasing on (−∞,∞), with no extrema.The curve keeps rising with no turn, level interval, or upper or lower output bound.
Answer
  • Increasing on (−∞, ∞).
  • Decreasing: none.
  • Constant: none.
  • Local maximum: none.
  • Local minimum: none.
  • Absolute maximum: none.
  • Absolute minimum: none.
Double-check Cubing each listed output returns its input, so the root values are correct.

Part 10: Read the retained toolkit summary

A summary table is like a directory: it helps you retrieve a result after you have understood the separate graphs. These drawn columns repeat the original nine-function summary. Return to each function’s part if you need its reason.

  • The drawn tables retain the original increasing, decreasing, and absolute-extremum summary for all nine functions.
  • Read reciprocal decreasing intervals separately, since input 0 is excluded.
  • Constant c is both absolute values at every input, with no strict rising or falling interval.
input toolkit functionoutput increasing onconstantnoneidentity(−∞, ∞)absolute value(0, ∞)quadratic(0, ∞)cubic(−∞, ∞)reciprocalnonereciprocal squared(−∞, 0)square root(0, ∞)cube root(−∞, ∞)
The columns draw the retained summary’s increasing intervals; square-root domain inclusion remains [0, ∞).

Worked example

Reflected square root on a half-open domain: an endpoint minimum and an unreached top

Let f(x) = -x-2 + 6 with domain 2 < x ≤ 11. On the graph, the left end at (2, 6) is a hollow dot, so x = 2 is excluded. The right end at (11, 3) is a filled dot, so x = 11 is included. A table of values from the domain gives f(3) = 5, f(6) = 4 and f(11) = 3. Find the absolute maximum and the absolute minimum of f, if they exist. For each one that exists, give the value and every input where it occurs. If one does not exist, explain why.

input xoutput f(x) = -√(x - 2) + 63564113
x:3,6,11;f(x)=-(x-2)+6:5,4,3.
Show the steps and the answer
input xoutput f(x) = -√(x - 2) + 63564113↓ evaluate: input given, read the output below it
Values of f(x) = -x-2 + 6 at the domain inputs 3, 6 and 11.
  1. Identify the toolkit shape. The function f(x)=-x-2+6 is the square-root toolkit function x shifted right 2, reflected across the horizontal axis, then shifted up 6.x rises on x ≥ 0. The minus sign flips the graph vertically, so f falls steadily as x increases. A function that always falls has no turning point in the middle of its domain.
  2. Read the whole stated domain: 2<x≤11. The left end x=2 is excluded (hollow dot at (2, 6)). The right end x=11 is included (filled dot at (11, 3)). There are no arrows, so the graph does not continue past either end.An extremum must be attained at an input that is in the domain. Included and excluded ends must be noted before any heights are compared.
  3. Check for unbounded behavior. On 2<x≤11, the value x-2 lies between 0 and 3. So f(x) lies between 6-3=3 and 6-0=6.The outputs are trapped between 3 and 6. They neither rise nor fall without bound, so unboundedness does not rule out either extremum.
  4. Test the top level, 6. As x moves toward 2 from the right, f(x) approaches 6. Reaching 6 would need x-2=0, which means x=2, and x=2 is excluded. For every x in the domain, x-2>0, so f(x) < 6. Any accepted input x has a slightly smaller accepted input that gives a larger output.A height that is approached but never reached is not an absolute maximum. The hollow dot at (2, 6) marks an unattained level, and no reached output is the largest. So there is no absolute maximum.
  5. Test the bottom level, 3. At x=11,f(11)=-9+6=-3+6=3, and x=11 is in the domain. For every x in the domain, x-2≤9=3, so f(x)≥3. Equality needs x-2=3, which happens only at x=11.The value 3 is reached at an accepted input, and no domain input gives a lower output. That makes 3 the absolute minimum, and x = 11 is the only input that produces it.
  6. Compare with the table: f(3)=5,f(6)=4,f(11)=3. These outputs fall as x grows, all are at least 3, and all are below 6.Reached heights at sample inputs agree with the analysis: a falling shape with its lowest point at the included right end.
Answer
  • Absolute maximum: none. The outputs approach 6 as x approaches 2, but x = 2 is excluded, so 6 is never reached.
  • Absolute minimum: 3, at x = 11.
Double-check Substitute x = 11: -11-2 + 6 = -3 + 6 = 3. Take an input close to the excluded end, x = 2.01: 0.01 = 0.1, so f(2.01) = 5.9. This is below 6. Take x = 2.0001: f = 6 - 0.01 = 5.99, still below 6. The outputs climb toward 6 without reaching it, and every value stays at least 3.

Work to write

  1. f(x) = -x-2 + 6 is decreasing on 2 < x ≤ 11
  2. x = 2 is excluded (hollow dot), x = 11 is included (filled dot)
  3. f(x) → 6 as x → 2⁺, but f(x) < 6 for every domain input
  4. Absolute maximum: none (6 is approached but not attained)
  5. f(11)=-9+6=3
  6. Absolute minimum: 3
  7. Occurs at x = 11

Part 11: Included endpoints can hold absolute records

The endpoint is the end of the accepted input stretch, like the end of a road. A filled dot means the point belongs to the graph. It can be the highest or lowest point even when there is no nearby road on one side.

  • Check included endpoints when comparing all domain outputs.
  • This text's local definition needs an open domain neighborhood on both sides.
  • The domain restriction can change absolute extrema even when the formula stays the same.
included endvalleyincluded end−22246810
The full restricted domain has an absolute maximum at its left endpoint.

Worked example

Absolute extrema of f(x) = -(x + 2)2 + 7 on a closed interval

Let f(x) = -(x + 2)2 + 7 with domain -5 ≤ x ≤ 0. The graph is a downward-opening parabola with vertex (-2, 7). The left end at (-5, -2) is a filled dot, so x = -5 is included. The right end at (0, 3) is a filled dot, so x = 0 is included. Find the absolute maximum and the absolute minimum of f, and state where each occurs.

(-5, -2)(-2, 7)(0, 3)−6−4−22−448
Downward parabola f(x) = -(x + 2)2 + 7 on -5 ≤ x ≤ 0.
Show the steps and the answer
  1. Identify the toolkit shape. f(x)=-(x+2)2+7 is the squaring function x2 reflected over the x-axis, shifted 2 left and 7 up. It is a downward-opening parabola with vertex (-2,7).The squaring toolkit function accepts every real input. Here the stated domain -5 ≤ x ≤ 0 is the only restriction.
  2. Read the whole domain: -5≤x≤0, with filled dots at both ends and no arrows.Filled dots mean x = -5 and x = 0 are accepted inputs. The graph stops there, so the outputs cannot rise or fall without bound.
  3. Check for heights that are approached but never reached. Both endpoints are included and the curve is unbroken between them, so there are no hollow dots or unattained levels.Only heights reached at accepted inputs can be absolute extrema.
  4. Find the greatest output. Since (x+2)2≥0,f(x)=7-(x+2)2≤7 for every x. The value 7 is reached when x+2=0, which gives x=-2, and -5≤-2≤0.The vertex is the highest point of a downward parabola, and here it lies inside the domain.
  5. Find the least output. Evaluate both endpoints: f(-5)=-(-3)2+7=-9+7=-2 and f(0)=-(2)2+7=-4+7=3. Compare them: -2<3.A downward parabola decreases as you move away from the vertex. The lowest point on a closed interval is therefore at the endpoint farther from the vertex. That endpoint is x = -5, which is 3 units from x = -2, while x = 0 is only 2 units away.
  6. Confirm that -2 is the least reached output and find every input that gives it. On the domain, -3≤x+2≤2, so (x+2)2≤9 and f(x)≥-2. Equality needs x+2=±3, so x=-5 or x=1. Only x=-5 is in the domain.The location must list every accepted input that produces the extreme value.
Answer
  • Absolute maximum: 7, at x = -2.
  • Absolute minimum: -2, at x = -5.
Double-check Test interior points: f(-4) = -(-2)2 + 7 = 3, f(-3) = -1 + 7 = 6, f(-1) = -1 + 7 = 6. Every value lies between -2 and 7. The right endpoint value f(0) = 3 is larger than -2, so it is not the minimum.

Work to write

  1. Domain -5 ≤ x ≤ 0, both endpoints included (filled dots)
  2. Vertex (-2, 7) is in the domain, and f(x) ≤ 7 for all x
  3. f(-5) = -2 and f(0) = 3
  4. Absolute maximum: 7
  5. Occurs at x = -2
  6. Absolute minimum: -2
  7. Occurs at x = -5

Part 12: Hollow endpoints and unattained heights

A hollow dot is like a closed entrance: you can approach that address, but it is not accepted. The height at that excluded point cannot count as a reached record. A bound alone is not enough; an absolute extremum must happen.

  • A hollow dot is excluded from the graph.
  • A limiting height reached only at an excluded input cannot supply an absolute extremum.
  • A bounded function can still have no absolute maximum or minimum if the relevant bound is unattained.
02[0, 2)
The domain [0, 2) includes 0 and excludes 2.

Worked example

A graph stops before its apparent highest height

For f(x) = x on domain [0, 2), find the absolute maximum and minimum.

Show the steps and the answer
02[0, 2)
The same picture again: [0, 2) on a number line.
included minacceptedaccepted212
y = x.
  1. f(0)=0 and every accepted output is at least 0.Every accepted input is nonnegative, and the output equals the input.
  2. No accepted input produces output 2.The right endpoint is excluded.
  3. For any accepted input b, the input b+22 is still below 2 and is greater than b.It is halfway from b toward 2, so it gives a larger accepted output.
  4. There is no absolute maximum; the absolute minimum is 0 at 0.Every candidate high output is beaten, while the lower bound 0 is reached.
Answer
  • Absolute maximum: none.
  • Absolute minimum: 0 at x = 0.
Double-check Input 1.5 is accepted and gives output 1.5; input 1.75 is also accepted and gives a greater output. The same halfway argument works for every b < 2.

Part 13: Explain local versus absolute in words

A nearby record and an overall record answer different questions. Think of the tallest tree near your house compared with the tallest tree in the entire forest. Either comparison can tie, and either kind of record can fail to exist.

  • Local compares nearby domain points on both sides; absolute compares every domain point.
  • A local extremum can be an isolated turn or a nonstrict flat comparison.
  • An absolute extremum can occur at an included endpoint.
  • Both values can occur at more than one input, or fail to exist.
input comparisonoutput local / absolutewhat it beatsnearby both sides / whole domainwhere it sitsinterior turn or flat / interior or included endpointcan be absentyes / yescan occur at several inputsyes / yes
The four columns draw the same local-versus-absolute comparisons as the retained comparison table.

Worked example

A turn without an overall record

The given full curve f(x) = x3 − 12x rises to (−2, 16), falls to (2, −16), then rises again. Its left end continues downward without bound and its right end upward without bound. Explain its local and absolute extrema.

Show the steps and the answer
below local minlocal maxlocal minabove local max−4−224−32−161632
y = 100.
  1. There is local maximum 16 at −2 and local minimum −16 at 2.The stated direction changes make the nearby two-sided comparisons hold.
  2. There is no absolute maximum or minimum.The given end behavior beats every upper or lower candidate bound.
  3. State that local describes a neighborhood while absolute describes the whole domain.The turning values win their nearby comparisons but lose the whole-domain comparison.
Answer
  • Local maximum: 16 at x = −2.
  • Local minimum: −16 at x = 2.
  • Absolute maximum: none.
  • Absolute minimum: none.
Double-check f(−4.5) = −37.125 < −16 and f(4.5) = 37.125 > 16 already beat the local values; the stated unbounded ends rule out every possible absolute record.

Part 14: Absolute records of a finite data function

A table can be a complete function when its domain is exactly the listed inputs, like a list of measured prices on eight specific dates. Then you compare every listed output to find the highest and lowest recorded values.

  • For domain exactly {2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012}, the table is the full finite function.
  • Absolute maximum: $3.68 at 2012. Absolute minimum: $2.31 at 2005.
  • A sampled peak or dip relative to adjacent years does not establish a continuous local extremum.
  • The finite table has no open domain neighborhood around a listed year, so this text's two-sided interval definition is not being applied to the sampled peak language.
input yearoutput price ($)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68
This table is complete only for the specified eight-input domain.

Worked example

Use all recorded columns

Let the gas-price function's domain be exactly the eight recorded years in the picture. Find its absolute maximum and minimum and describe the adjacent-year peak and dip honestly.

input yearoutput price ($)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68
The same picture again: A table of 8 values of price ($), for year from 2005 to 2012.
Show the steps and the answer
input yearoutput price ($)20052.3120062.6220072.8420083.320092.4120102.8420113.5820123.68↓ evaluate: input given, read the output below it
Highlight the two global records and the two adjacent-year sampled comparisons without filling in unknown…
  1. Scan every output column; 3.68 is greatest and 2.31 is least.All domain outputs appear in this finite table.
  2. The column under 2012 gives 3.68 and the column under 2005 gives 2.31.These are the input locations of the record outputs.
  3. The 2008 output 3.30 exceeds its adjacent recorded outputs 2.84 and 2.41; the 2009 output 2.41 is below adjacent 3.30 and 2.84.Those column comparisons describe an observed sampled peak and dip, rather than continuous behavior at unrecorded times.
Answer
  • Absolute maximum of the finite data function: $3.68 at 2012.
  • Absolute minimum of the finite data function: $2.31 at 2005.
  • Sampled adjacent-year peak: $3.30 in 2008.
  • Sampled adjacent-year dip: $2.41 in 2009.
  • Unrecorded prices and their extrema: undetermined.
Double-check Every listed output lies between 2.31 and 3.68, and both bounds occur in the table. The early and late endpoint years can hold the finite data function's absolute records.

5 How

  1. Review the nine toolkit shapes and their domain restrictions before using the summary table.
  2. Look at the whole stated domain, including included endpoint dots and any arrows or stated continuation.
  3. If outputs rise without bound, there is no absolute maximum; if they fall without bound, there is no absolute minimum.
  4. A height approached but never reached is not an absolute extremum. Check excluded hollow dots and unattained levels.
  5. Compare the remaining reached heights with every domain piece. Find the greatest and least reached outputs and every input producing each.
  6. Answer with the value and location on separate lines. A viewing-window edge does not establish a domain endpoint.

Check the whole domain and attainment

1
Is the proposed point excluded by a hollow dot or a domain restriction?
YesIts height is not reached there. It cannot supply an absolute extremum.
NoKeep it as a candidate.
↓
2
Do outputs continue beyond every upper or lower bound?
YesThere is no absolute maximum for unbounded-above outputs, or no absolute minimum for unbounded-below outputs.
NoCheck attainment instead of assuming a record exists.
↓
3
Is the best apparent height only approached, never reached?
YesThat bound is not an absolute extremum.
NoCompare the attained height with every domain piece.
↓
4
Does more than one input attain the same absolute height?
YesList every such input; there is still one maximum or minimum value.
NoGive the single value and location.
  1. Identify the exact domain and its included or excluded endpoints.
  2. Check every branch for outputs without an upper or lower bound.
  3. Check whether a proposed record height is actually reached.
  4. Compare all reached candidate heights, including included endpoints.
  5. List every input that reaches each absolute value.

6 Worked example

Worked example

Showing that a highest point exists

Find the absolute maximum and absolute minimum of f(x) = 4 − x2 on its whole real-number domain, if they exist.

Show the steps and the answer
(0, 4)absolute max 4−22−8−44
The same picture again: y = −x2 + 4.
input xoutput 4 − x²−20−1304132010−96
x:−2,−1,0,1,2,10;4−x2:0,3,4,3,0,−96.
  1. x2≥0 for every real input x.Positive and negative numbers have positive squares, and 02 = 0.
  2. Thus f(x)=4−x2≤4 for every input.Subtracting a nonnegative number leaves 4 or less.
  3. f(0)=4−02=4.Input 0 is accepted and is the only input making the square zero.
  4. The absolute maximum is 4 at x=0.No output is greater and this value is attained.
  5. f(10)=−96 and f(100)=−9,996; larger input sizes make the output fall without bound.x2 grows without bound and is subtracted from 4.
  6. There is no absolute minimum.Every candidate low height is eventually beaten by subtracting a still larger square.
Answer
  • Absolute maximum: 4 at x = 0.
  • Absolute minimum: none.
Double-check The fntable columns under −2 and 2 both show output 0, the columns under −1 and 1 show 3, and the column under 0 shows the vertex height 4. The column under 10 drops to −96. These samples support the picture; the square bound and unbounded-square argument prove the full-domain result.

Ladder: from easy to exam-hard

Try each rung before you open its steps.

Rung 1 of 6

Absolute extrema of a restricted parabola with an open endpoint

The function f(x) = -(x - 1)2 + 4 is defined only on the domain -1 ≤ x < 4. Its graph is a downward-opening parabola with vertex (1, 4). There is a closed (filled) dot at (-1, 0) and an open (hollow) dot at (4, -5). Find the absolute maximum and the absolute minimum of f, if they exist. For each one that exists, give the value and every input where it occurs.

(-1, 0) closed(1, 4) vertex(4, -5) open−2246−44
The parabola f(x) = -(x - 1)2 + 4.
Show the steps and the answer
  1. Identify the toolkit shape: f is the square function y=x2 reflected across the x-axis, shifted 1 right and 4 up. It opens downward with vertex (1, 4).Knowing the toolkit shape tells us the graph rises up to the vertex and falls on both sides of it.
  2. Read the whole stated domain: -1≤x<4. The left endpoint x=-1 is included (closed dot). The right endpoint x=4 is excluded (hollow dot). There are no arrows, so the graph does not continue past either end.Absolute extrema are judged only over the accepted inputs, so we must know exactly which endpoints count.
  3. Check for unbounded behavior. The domain is a bounded interval and the graph stops at both ends, so the outputs neither rise nor fall without bound.Unbounded outputs would rule out an extremum, but that does not happen here.
  4. Check the hollow dot. As x approaches 4 from the left, f(x) approaches f(4)=-(3)2+4=-5. But x=4 is not in the domain, so the height -5 is never reached. For example, f(3.9)=-(2.9)2+4=-4.41, which is still above -5.A height that is approached but never attained cannot be an absolute extremum.
  5. Compare the reached heights. For -1≤x<4, the quantity x-1 lies in [-2,3), so (x-1)2 lies in [0, 9). That makes f(x) lie in (-5,4]. The greatest reached output is 4, at x=1 only. The closed endpoint gives f(-1)=0, which is less than 4. No least reached output exists, because the outputs get arbitrarily close to -5 without ever reaching it.The absolute maximum is the greatest attained output. The absolute minimum would have to be a least attained output, and the set (-5, 4] has no least element.
  6. State the results with value and location on separate lines.The lesson asks for the value and the location to be reported separately.
Answer
  • Absolute maximum value: 4
  • Location: x = 1
  • Absolute minimum: none (the outputs approach -5 but never reach it, because x = 4 is excluded)
Double-check f(1) = -(0)2 + 4 = 4, and every domain input gives (x - 1)2 ≥ 0, so f(x) ≤ 4. For a minimum, test values near the open end: f(3.99) = -(2.99)2 + 4 = -4.9401, which is above -5. Any proposed minimum m > -5 is beaten by inputs closer to 4, and -5 itself is never attained. So there is no absolute minimum.

Work to write

  1. Domain: -1 ≤ x < 4; x = 4 excluded (hollow dot)
  2. Vertex (1, 4) is the highest reached point
  3. Absolute maximum value: 4
  4. Location: x = 1
  5. f(x) → -5 as x → 4⁻, but f never equals -5
  6. No absolute minimum

Rung 2 of 6

Showing that a highest point exists

Find the absolute maximum and absolute minimum of f(x) = 4 − x2 on its whole real-number domain, if they exist.

Show the steps and the answer
(0, 4)absolute max 4−22−8−44
The same picture a third time: y = −x2 + 4.
input xoutput 4 − x²−20−1304132010−96
The same picture again: x: −2, −1, 0, 1, 2, 10; 4 − x2: 0, 3, 4, 3, 0, −96.
  1. x2≥0 for every real input x.Positive and negative numbers have positive squares, and 02 = 0.
  2. Thus f(x)=4−x2≤4 for every input.Subtracting a nonnegative number leaves 4 or less.
  3. f(0)=4−02=4.Input 0 is accepted and is the only input making the square zero.
  4. The absolute maximum is 4 at x=0.No output is greater and this value is attained.
  5. f(10)=−96 and f(100)=−9,996; larger input sizes make the output fall without bound.x2 grows without bound and is subtracted from 4.
  6. There is no absolute minimum.Every candidate low height is eventually beaten by subtracting a still larger square.
Answer
  • Absolute maximum: 4 at x = 0.
  • Absolute minimum: none.
Double-check The fntable columns under −2 and 2 both show output 0, the columns under −1 and 1 show 3, and the column under 0 shows the vertex height 4. The column under 10 drops to −96. These samples support the picture; the square bound and unbounded-square argument prove the full-domain result.

Rung 3 of 6

Absolute extrema of a parabola on a closed interval with a tied maximum

The function f(x) = (x - 2)2 - 1 is defined only on the domain -1 ≤ x ≤ 5. Its graph is an upward-opening parabola with vertex (2, -1). Both ends are included. There is a closed (filled) dot at (-1, 8) and a closed (filled) dot at (5, 8). Find the absolute maximum and the absolute minimum of f, if they exist. Give each value and every input where it occurs.

(-1, 8)(2, -1)(5, 8)−224648
Graph of f(x) = (x - 2)2 - 1.
Show the steps and the answer
  1. Identify the toolkit shape. f(x)=(x-2)2-1 is the squaring function shifted right 2 and down 1. It opens upward and its vertex is (2,-1).The squaring toolkit function x2 has a lowest point at its vertex and rises on both sides. A shift moves the vertex but does not change that behavior.
  2. Read the whole stated domain: -1≤x≤5. Both endpoints use ≤, so x=-1 and x=5 are accepted inputs. The graph has no arrows.The domain is fixed by the problem statement. It is not set by any viewing window. Included ends shown with filled dots are candidates for extrema.
  3. Check for unbounded behavior. The domain is a closed, bounded interval, so the outputs cannot rise or fall without bound.Unbounded outputs would rule out an extremum. Here the graph stops at two filled dots.
  4. Check for heights that are approached but never reached. There are no hollow dots, and every height on the curve from x=-1 to x=5 is reached.Only attained values can be absolute extrema. With both ends closed, no endpoint height is excluded.
  5. List the candidates. The interior vertex is at x=2, with f(2)=(0)2-1=-1. The left end is at x=-1, with f(-1)=(-3)2-1=9-1=8. The right end is at x=5, with f(5)=(3)2-1=9-1=8.On a closed interval, an absolute extremum of this parabola can occur only at the vertex or at an included endpoint.
  6. Compare the reached heights: -1, 8 and 8. The greatest is 8, reached at both x=-1 and x=5. The least is -1, reached only at x=2.The absolute maximum must be at least as large as every output, and the absolute minimum must be at most as small as every output. Every input that produces the extreme value must be reported.
Answer
  • Absolute maximum: 8, at x = -1 and x = 5.
  • Absolute minimum: -1, at x = 2.
Double-check Test an interior input such as x = 0: f(0) = 4 - 1 = 3, which lies between -1 and 8. Test x = 4: f(4) = 4 - 1 = 3, which also lies between them. The two ends are each 3 units from the vertex at x = 2, so by symmetry they share the height 8. No point in the domain is lower than the vertex, because (x - 2)2 ≥ 0 means f(x) ≥ -1.

Work to write

  1. f(2) = -1 (vertex, interior)
  2. f(-1) = 8 (included left end)
  3. f(5) = 8 (included right end)
  4. Absolute maximum: 8
  5. occurs at x = -1 and x = 5
  6. Absolute minimum: -1
  7. occurs at x = 2

Rung 4 of 6

A graph stops before its apparent highest height

For f(x) = x on domain [0, 2), find the absolute maximum and minimum.

Show the steps and the answer
02[0, 2)
The same picture a third time: [0, 2) on a number line.
included minacceptedaccepted212
The same picture again: y = x.
  1. f(0)=0 and every accepted output is at least 0.Every accepted input is nonnegative, and the output equals the input.
  2. No accepted input produces output 2.The right endpoint is excluded.
  3. For any accepted input b, the input b+22 is still below 2 and is greater than b.It is halfway from b toward 2, so it gives a larger accepted output.
  4. There is no absolute maximum; the absolute minimum is 0 at 0.Every candidate high output is beaten, while the lower bound 0 is reached.
Answer
  • Absolute maximum: none.
  • Absolute minimum: 0 at x = 0.
Double-check Input 1.5 is accepted and gives output 1.5; input 1.75 is also accepted and gives a greater output. The same halfway argument works for every b < 2.

Rung 5 of 6

A turn without an overall record

The given full curve f(x) = x3 − 12x rises to (−2, 16), falls to (2, −16), then rises again. Its left end continues downward without bound and its right end upward without bound. Explain its local and absolute extrema.

Show the steps and the answer
below local minlocal maxlocal minabove local max−4−224−32−161632
The same picture again: y = 100.
  1. There is local maximum 16 at −2 and local minimum −16 at 2.The stated direction changes make the nearby two-sided comparisons hold.
  2. There is no absolute maximum or minimum.The given end behavior beats every upper or lower candidate bound.
  3. State that local describes a neighborhood while absolute describes the whole domain.The turning values win their nearby comparisons but lose the whole-domain comparison.
Answer
  • Local maximum: 16 at x = −2.
  • Local minimum: −16 at x = 2.
  • Absolute maximum: none.
  • Absolute minimum: none.
Double-check f(−4.5) = −37.125 < −16 and f(4.5) = 37.125 > 16 already beat the local values; the stated unbounded ends rule out every possible absolute record.

Rung 6 of 6

Absolute maximum at two places

You are asked for the highest and lowest heights reached on the complete pictured domain [−2.5, 3], including both filled endpoint dots. The dots sample the original smooth curve f(x) = 16 − 1.04(x2 − 4)2 on this restricted domain; it reproduces the extreme values in the textbook illustration. Read all inputs where each absolute extreme value occurs.

Show the steps and the answer
included left endpointmaximum 16−22−12−661218
The dots sample one smooth curve on the complete restricted domain.
−2.53left endright end[−2.5, 3]
Both ends of the allowed-input interval are included, so both endpoint outputs must be checked.
  1. Scan the pictured heights over the entire stated domain. The two tallest dots have height 16, at x=−2 and x=2.These are the tops of the pictured curve. Algebra checks the upper bound: (x2 − 4)2 ≥ 0, so subtracting 1.04 times that square can never produce an output above 16.
  2. It is reached at two inputs, x=−2 and x=2.Substitute x = −2 or x = 2: x2 = 4 and (x2 − 4)2 = 0, so both outputs equal 16. No other input makes x2 − 4 zero on this domain.
  3. The least y-value anywhere is −10, reached at x=3.The curve is lowest at the included right endpoint. On [−2.5, 3], 0 ≤ x2 ≤ 9, so −4 ≤ x2 − 4 ≤ 5. Thus the distance of x2 − 4 from 0 is at most 5, and its square is at most 25. Multiplying by positive 1.04 preserves that upper bound; subtracting from 16 gives f(x) ≥ 16 − 1.04 · 25 = −10. Finally f(3) = −10 reaches the bound.
Answer
  • Absolute maximum: 16 at x = −2 and x = 2.
  • Absolute minimum: −10 at x = 3.
Double-check Each answer has two parts: the value (a y-value) and where it occurs (x-values). There is one absolute maximum value, 16, even though it happens at two places. The left endpoint has f(−2.5) = 10.735, above −10. The local valley at x = 0 has height −0.64, also above −10, so a nearby valley need not be the absolute minimum.

Counterexamples

What it is not, and the tempting wrong moves.

✗ Not this: The tallest visible sample must be the absolute maximum.
A continuing curve can reach greater heights beyond the window; finite samples do not bound the whole domain.
✓ Instead: Check the full stated domain and end behavior.
✗ Not this: If outputs approach 0, then 0 is the absolute minimum.
1x2 stays positive at every allowed input and never reaches 0.
✓ Instead: Require the candidate height to be attained.
✗ Not this: A domain endpoint cannot be an extremum.
For x2 on [−3, 2], the included endpoint −3 gives the absolute maximum 9.
✓ Instead: Check included endpoints for absolute extrema; this text's local rule still needs both sides.
✗ Not this: An absolute maximum occurring at two inputs means two different maximum values.
Two peaks of height 16 share the same greatest output.
✓ Instead: Give one maximum value and list every input that reaches it.

Tips and tricks

  • Absolute here means over the whole domain; it does not mean the absolute value operation |x|.
  • Write both the output record and every input where it is attained.
  • A complete graph stops at its stated ends. Check its included end dots, but do not treat every picture-window edge as an endpoint.
  • Bounded is not enough. Check whether the best height is reached.
  • The toolkit summary's reciprocal intervals describe each branch separately.

7 The trap

Calling the highest point visible in a picture the absolute maximum when an arrow shows the graph keeps rising beyond the window. Check how the graph behaves at its ends before you name an absolute maximum or minimum.