Organises every graph transformation under one rule: changes written outside a function act on the output, vertically, and do what they say; changes written inside act on the input, horizontally, and do the opposite. Covers shifts, stretches, compressions and reflections, the order in which combined transformations are applied, and the even and odd symmetry tests.
Subject: Precalculus · 71 slides · symbolic lesson
Open the interactive version of this deck
Title
Precalculus · Section 1.5
Move it, flip it, stretch it: one graph becomes a whole family
Objectives
You already know the shapes of the toolkit functions by heart. This lesson turns each of those single graphs into an endless family, by changing the formula in four ways and learning to read each change straight off the page.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 100-130 — learning objectives 1.5.1 to 1.5.5
Everything here is built on graphs you already know: the line, the parabola, the cubic, the square root, the cube root, the absolute value and the reciprocal. Nothing new gets drawn from scratch. Instead you learn to take a shape you have already met and move it, flip it, stretch it or squash it, and to do all of that straight from the formula without plotting a single point.
The payoff is enormous and it starts immediately. Every quadratic in vertex form, every sine wave with an amplitude and a phase, every exponential growth model you meet later is one of these toolkit shapes with four numbers attached to it. Learn to read those four numbers now and the rest of the course stops feeling like a list of unrelated graph types.
There is one genuine difficulty, and it is worth naming up front. Half of the transformations behave exactly as the formula reads, and the other half behave in reverse. Knowing which half you are looking at, and why it reverses, is the real content of this lesson.
Warm-up
Discussion prompt
You know the graph of the squaring function. Predict where the graph sits when three is added to the whole thing, and then where it sits when three is added to x before the squaring happens. Which of the two answers surprises you?
Write both predictions down before you reveal anything. The point of committing on paper is that the second answer catches nearly everyone, and it only teaches you something if you can see that it caught you.
The first expression adds three after the squaring is finished, so it adds three to the answer. Every height goes up by three and the parabola rises. The second expression adds three to the input before the squaring happens, which is a completely different kind of change, and it turns out to slide the parabola sideways rather than up.
If you also predicted the direction of that sideways slide, check it: the plus sign moves the graph to the left, not the right. That is not a printing error and it is not arbitrary. By the end of Part 1 you will be able to explain it in one sentence.
Section
Part 1
Concept
There are a dozen separate rules in this section if you count them one by one, and two if you understand them. Every change you can make to a formula happens at one of exactly two moments.
Figure (svg): A flow diagram: the input x enters a box marked b times x minus h, which is the inside change; the result goes into the function f; the output goes into a box marked times a, plus k, which is the outside change; and then out as y.
transformation — A change to a function's formula that produces a predictable change to its graph. Changes written outside the function act on the output and are vertical; changes written inside act on the input and are horizontal.
Vertical changes do precisely what the formula says. Horizontal changes do the opposite. That second fact costs most students more marks than anything else in the chapter, so the next two slides are spent on it.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 100-101 — the section introduction
Follow the diagram left to right, because it is really a description of what happens when you evaluate the formula. You start with a number. Anything written inside the brackets happens to that number first, before the function has done anything at all. Then the function runs. Then anything written outside the brackets happens to the answer.
That order is the whole explanation of the lesson. A change that happens after the function has finished is acting on a height, so it moves the graph up and down, and it does precisely what it is written to do. A change that happens before the function runs is acting on an input, so it moves the graph left and right, and because it changes which input gets handed over, it works backwards.
Do not try to hold twelve separate rules in your head. Hold this diagram, and ask of every number you meet: does it happen before the function, or after it?
Notation
Annotate
On: \( g(x) = a\,f\bigl(b(x-h)\bigr) + k \)
Four letters, four moves. Learning to spot which letter is which in a printed formula is most of the skill, and the last part of this lesson is about formulas that hide their letters.
This one line is the destination of the whole lesson, so read each annotation slowly and match it to a picture in your head. The letters a and k live outside and control vertical behaviour. The letters b and h live inside and control horizontal behaviour.
Notice the shape of the inside part. It is a factor multiplying a bracket, and the bracket is written as a subtraction. Both of those choices are deliberate. The subtraction means the number h reads off as the shift itself, including its sign. The factoring means the scaling and the shifting are cleanly separated, which they are not when the inside is written as two terms added together.
You will meet formulas that are not in this form, and getting them into it is the last skill in the lesson. When a formula already looks like this line, reading off all four transformations takes about five seconds.
Picture it
Figure (svg): Two parabolas on one set of axes: the squaring function with its vertex at the origin, and the squaring of x minus two with its vertex two units to the right. A dashed line at height four meets the first at x equal to two and the second at x equal to four.
\[ g(x) = f(x-2) \;\Longrightarrow\; g(4) = f(2) \]
The new rule subtracts two before handing the input over, so it needs an input two larger to feed the old rule the same number. Every feature of the graph therefore appears two units later along the axis, and later along a left-to-right axis means further right.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 104 — the discussion before the Horizontal Shift box
This argument is worth following once, properly, because after that the reversal stops being something you memorise. Look at the dashed line at height four. The original parabola reaches that height at the input two. The new one reaches it at the input four.
Why? Because the new rule subtracts two from whatever it is given before squaring. To make it square the number two, you must hand it four. So every single thing the old graph did at some input, the new graph does two inputs later. Doing everything later means the whole picture appears further along the axis, and further along a left-to-right axis means further right.
Run the same sentence with a plus sign and you get the opposite: the rule adds before squaring, so it needs a smaller input, and everything happens sooner, which is further left. Say the sentence out loud a couple of times. It replaces the rule entirely.
Section
Part 2
Concept
Figure (svg): The cube root curve and the same curve raised one unit, drawn on one set of axes, with a vertical arrow of length one joining them.
\[ g(x) = f(x) + k \]
Think of the height of the graph as the letter y. Adding k to the function is the same as replacing y by y plus k, so every point rises by k. A positive k lifts the graph and a negative k lowers it, and nothing at all happens to the inputs.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 101 — the Vertical Shift box and Figure 1.71
Look at the picture and check that the gap between the two curves is the same everywhere. That is the signature of a vertical shift: not a tilt, not a stretch, the same one unit at the far left, at the origin and at the far right.
There is a neat way to see why. The height of a graph is the number you call y, and adding a constant to the function is the same as replacing y by y plus that constant. Every height, wherever it is, goes up by the same amount, so the whole curve is carried upward without deforming.
The other half of the statement matters just as much: nothing happens to the inputs. If you are ever unsure whether a change was vertical, look at where the graph crosses a particular height rather than at the height itself, and check whether those crossing points moved sideways. For a vertical shift they do not.
Worked example
Airflow vents near the roof of a green building open and close through the day. The area open is recorded in square feet against hours after midnight. The manager wants twenty more square feet open at every moment of the day and night.
| t, hours after midnight | 0 | 8 | 10 | 17 | 19 | 24 |
|---|---|---|---|---|---|---|
| V(t), square feet | 0 | 0 | 220 | 220 | 0 | 0 |
Decide which row the change touches
Why: Twenty more square feet is twenty more output, at every input. That is an outside change.
\[ S(t) = V(t) + 20 \]
Add twenty to each output
Why: The input row is copied across untouched.
\[ S(0) = 0+20 = 20, \qquad S(10) = 220+20 = 240 \]
Figure (svg): Two step-shaped graphs of open vent area against hours after midnight: the original plan rising from zero to 220 square feet between 8 and 10 and falling back between 17 and 19, and the new plan exactly twenty square feet higher everywhere.
Write the whole new row
Why: Six readings, each one twenty higher.
\[ S: \; 20, \; 20, \; 240, \; 240, \; 20, \; 20 \]
Check that nothing moved sideways
Why: The vents still begin to open at eight hours and are shut again by nineteen, exactly as before.
\[ S(8) = V(8)+20 = 20, \qquad S(19) = V(19)+20 = 20 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 101-102 — Example 1.49 and Table 1.25
Start by reading the situation, not the algebra. The manager is not changing when the vents open, only how much they open. That sentence alone tells you the change is to the output, so it is outside, so it is vertical, and so it does exactly what it says.
The table makes this concrete. The top row is untouched, entry for entry, and the bottom row is the old bottom row with twenty added to each number. When a question gives you a table rather than a formula, that pattern is how you recognise a vertical shift: one row changes and the other does not.
The check at the end is the one you should copy. Vents that used to start opening at eight hours after midnight still start opening at eight hours after midnight. If the times had moved, you would have made a horizontal change by mistake, and the numbers would tell you so immediately.
Comparison
\[ g(x) = f(x) - 3 \]
Comparison matrix
| x | f(x) | g(x) = f(x) − 3 |
|---|---|---|
| 2 | 1 | -2 |
| 4 | 3 | 0 |
| 6 | 7 | 4 |
| 8 | 11 | 8 |
Fill the blanks before checking. Notice which column never changes: the inputs are identical all the way down, which is the fingerprint of a vertical shift.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 103 — Example 1.50 and Tables 1.26 to 1.27
Fill in every blank before you look at the answers. The arithmetic is deliberately trivial, because the thing being tested is not arithmetic.
The first blank is the one worth pausing on. Taking three from one gives a negative answer, and a surprising number of people write two instead, simply because the table looked like a table of positive numbers. Subtracting a constant can take outputs below zero, and that is exactly what a downward shift of a graph looks like.
When you finish, run your eye down the first column rather than the last. It is identical to the original. That is the test that tells you this was a vertical shift and not something else, and it is worth writing as a sentence in any answer you hand in.
Worked example
A ball is thrown straight up from the ground. Its height in metres after t seconds is given below. Now suppose the same throw is made from the top of a ten-metre building instead.
\[ h(t) = -4.9t^2 + 30t \]
Decide inside or outside
Why: Every height is ten metres greater than before; the timing of the throw is unchanged. That is an output change.
\[ b(t) = h(t) + 10 \]
Write the formula
Why: Add ten on the outside of the whole rule.
\[ b(t) = -4.9t^2 + 30t + 10 \]
Test the start of the flight
Why: At time zero the ball leaves the roof, not the ground.
\[ b(0) = 0 + 0 + 10 = 10 \]
Check the highest point
Why: The turning time is unchanged and the peak height is exactly ten metres higher, which is what a purely vertical shift must do.
\[ t = \frac{30}{9.8} \approx 3.061: \quad h \approx 45.918, \quad b \approx 55.918 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 103 — Try It 1.32
Before touching the formula, decide what physically changed. The throw is the same throw: the same push, the same upward speed, the same gravity. The only difference is that the ball starts ten metres higher. So every height is ten greater and no time has moved.
That reasoning tells you the ten goes outside, and outside for a formula written out in full just means at the end. You do not touch the squared term and you do not touch the thirty times t. You add ten on the end.
The final check is the one that proves you got the direction right. The ball still reaches its highest point at the same moment, because the timing did not change, and the peak height is exactly ten metres higher than before. If your formula had put the ten inside, the turning time would have shifted, which would mean you had accidentally claimed that standing on a roof changes when the ball stops rising.
Concept
Figure (svg): The cube root curve and the curve of the cube root of x plus one on one set of axes, with a horizontal arrow one unit long pointing left.
\[ g(x) = f(x-h) \]
The standard form is written with a minus sign on purpose. In that form the number h is the shift itself: positive h moves the graph right, negative h moves it left. Written the other way, as a plus, the direction is the opposite of the sign you can see.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 103-104 — the Horizontal Shift box and Figure 1.74
Compare this picture with the vertical one two slides back. The gap between the curves is again constant, but now it is measured horizontally, and the two curves sit at the same heights as each other rather than at the same inputs.
The formula is written with a minus sign, and that is a convention chosen for your convenience rather than a fact of nature. In that form, whatever number you see after the minus is the shift, with its sign. A minus two means two to the right. A minus with a negative two, which simplifies to a plus two, means two to the left.
So when you meet a plus inside, your first move is either to rewrite it as a subtraction of a negative, or to ask which input makes the inside equal zero. Both give the same answer, and the second is faster once you trust it.
Worked example
Back to the building. In autumn the manager decides the venting programme starts too late and wants the whole plan to begin two hours earlier.
Match one event on the two plans
Why: The old plan first moves air at eight; the new plan first moves air at six.
\[ F(6) = V(8), \qquad F(8) = V(10) \]
Turn the matching into a rule
Why: Whatever the new plan does at time t, the old plan did two hours later.
\[ F(t) = V(t+2) \]
Put it in the standard form
Why: The standard form subtracts h, so a plus two means h is negative two.
\[ F(t) = V\bigl(t-(-2)\bigr), \qquad h = -2 \]
Check the direction against common sense
Why: A negative h means a shift to the left, and starting everything earlier does move every event leftwards on a time axis.
\[ h = -2 < 0 \;\Longrightarrow\; \text{shift left } 2 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 104-105 — Example 1.51 and Figure 1.75
This example is the one that makes the reversal feel reasonable rather than perverse, so it is worth reading twice. Starting the programme earlier moves everything to the left on a time axis, which everybody agrees with. The question is what the formula looks like.
Work from one matched event. The old plan first moved air at eight hours; the new plan does it at six. So the new plan at six is doing what the old plan did at eight. Two hours have to be added to the input before the old rule will recognise it, which gives a plus two inside.
Now line that up with the standard form, which subtracts h. A plus two inside means h is negative two, and a negative h means a shift to the left. Everything agrees: earlier, leftwards, plus inside, negative h. If you can reconstruct that chain, the reversal is no longer something you can forget.
Worked example
\[ g(x) = f(x-3) \]
| x | 2 | 4 | 6 | 8 |
|---|---|---|---|---|
| f(x) | 1 | 3 | 7 | 11 |
Ask what input g needs to reach the first entry
Why: The rule takes three away before calling f, so the input handed to g must be three larger.
\[ g(5) = f(5-3) = f(2) = 1 \]
Repeat for the other three
Why: Every input grows by three; every output is copied straight across.
\[ g(7) = f(4) = 3, \quad g(9) = f(6) = 7, \quad g(11) = f(8) = 11 \]
| x | 5 | 7 | 9 | 11 |
|---|---|---|---|---|
| g(x) | 1 | 3 | 7 | 11 |
Check which row actually moved
Why: The outputs are unchanged and the inputs each rose by three, so the picture slid three units to the right.
\[ \{2,4,6,8\} \;\longmapsto\; \{5,7,9,11\} \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 105-106 — Example 1.52 and Tables 1.28 to 1.29
Here the reversal appears in a table instead of a picture, and tables are often where it becomes clearest. The rule takes three away from its input before it calls the original function, so to make the original function see the number two, you must give the new rule a five.
Work entry by entry rather than trying to see the pattern all at once. Five gives two, seven gives four, nine gives six, eleven gives eight. Only when the whole row is done should you step back and notice that every input grew by three.
Compare this with the vertical table earlier. There the inputs were fixed and the outputs moved. Here the outputs are fixed and the inputs moved. Those two patterns are how you tell a horizontal change from a vertical one when all you have is a table of numbers.
Worked example
A graph has exactly the shape of the squaring function, but its lowest point sits at the place where x is two and y is zero.
Find the matching point on the parent
Why: The squaring function has its vertex at the origin.
\[ f(x) = x^2, \qquad f(0) = 0 \]
Say what the new rule must do
Why: It has to produce the output zero when the input is two.
\[ g(2) = 0 \]
Choose the sign that makes that happen
Why: Only subtracting two sends the input two back to the input zero that f needs.
\[ g(2) = f(2-2) = f(0) = 0 \;\Longrightarrow\; g(x) = f(x-2) \]
Write the formula out
Why: Substitute x minus two into the squaring rule.
\[ g(x) = (x-2)^2 \]
Check with a second point
Why: The parent passes through height one at input one; two units to the right of that is input three, and the new rule does give one there.
\[ g(3) = (3-2)^2 = 1^2 = 1 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 107 — Example 1.53 and Figure 1.77
Going from a picture to a formula is the reverse of everything you have done so far, and the trick is always the same: find one point whose position you know exactly on both the parent graph and the new one.
For a parabola the obvious choice is the vertex, because it is the only point you can locate by eye without measuring. The parent has its vertex at the origin; this one has its vertex two units to the right. So the shift is two to the right, and the standard form tells you to write that as a minus two inside.
The book's analysis is worth internalising as a sanity check. Ask yourself what input the new rule needs in order to feed zero to the squaring function, and the answer is two. If you had written a plus two by mistake, the new rule would need an input of negative two, and the vertex would be on the wrong side of the origin.
Prediction
The function G tells you how many gallons of fuel it takes to drive m miles.
\[ G(m) + 10 \qquad \text{against} \qquad G(m+10) \]
Predict first
Which reading of the two expressions is the correct one?
Correct: The first is ten extra gallons; the second is the fuel for a journey ten miles longer
Why: The ten written outside is added to the output, and the output is measured in gallons, so it is ten extra gallons and the graph moves up. The ten written inside is added to the input, and the input is measured in miles, so it is the fuel for a longer journey and the graph moves sideways. Checking the units of the thing you added is the fastest way to tell the two apart.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 108 — Example 1.54
Commit to an option before revealing. This is the cleanest test in the section of whether you have really absorbed the inside and outside idea, because there are no graphs to lean on at all.
The shortcut is to check the units of the thing you are adding to. Add ten to the output and you are adding ten gallons, because gallons is what the output measures. Add ten to the input and you are adding ten miles, because miles is what the input measures. The units tell you which one you have, and then the rules tell you which way the graph moves.
Applied questions almost always arrive in this form, with a sentence rather than a formula. Practise translating the sentence into inside or outside before you write anything down, because once you have chosen correctly the rest is mechanical.
Trap
A line that appears on almost every first attempt:
\[ y = (x+3)^2 \]
\[ \Rightarrow\; \text{right } 3? \]
Wrong. The plus is inside, so the move is left, not right.
Rewrite the inside in the standard form before reading anything, and the sign stops lying to you.
\[ (x+3)^2 = \bigl(x-(-3)\bigr)^2 \]
\[ h = -3 \;\Longrightarrow\; \text{left } 3 \]
If you would rather not rewrite it, ask instead which input makes the inside zero. Here the inside vanishes at negative three, and that is where the vertex goes.
This is the single most common error in the whole section, and it is common precisely because the correct answer feels wrong. A plus sign inside moves the graph to the left, and no amount of staring at the plus sign will make that feel natural.
The defence is to refuse to read the direction off the sign at all. Instead, ask which input makes the inside equal zero. For the squaring of x plus three, the inside vanishes at negative three, so the vertex is at negative three, so the graph moved left. That question works for every toolkit function and never lies to you.
If you prefer algebra to questions, rewrite the plus as a subtraction of a negative, so that the formula is genuinely in standard form before you read anything off it. Both methods cost about five seconds, which is cheap insurance on an exam.
Worked example
\[ f(x) = \lvert x \rvert, \qquad h(x) = f(x+1) - 3 \]
Split the formula into its two halves
Why: One number sits inside the function and one sits outside it.
\[ \text{inside: } x+1, \qquad \text{outside: } -3 \]
Read the inside change
Why: Inside changes reverse, so a plus one means a move of one to the left.
\[ f(x+1) = f\bigl(x-(-1)\bigr): \; \text{left } 1 \]
Read the outside change
Why: Outside changes are literal, so minus three means down three.
\[ -3: \; \text{down } 3 \]
Walk the corner point through both moves
Why: Start at the origin, go one left, then three down.
\[ (0,0) \;\longmapsto\; (-1,0) \;\longmapsto\; (-1,-3) \]
Figure (svg): The V-shaped absolute value graph with its corner at the origin, and a second V with its corner one unit left and three units down, with two arrows tracing the corner from one to the other.
Check the corner against the formula
Why: Substituting the predicted input should give the predicted height.
\[ h(-1) = \lvert -1+1 \rvert - 3 = 0 - 3 = -3 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 108-109 — Example 1.55 and Figures 1.78 to 1.79
Two transformations at once, and the safe way to handle them is to refuse to think about both at the same time. Split the formula in half first: one is inside and one is outside. Then read each half on its own.
Tracking a single well-chosen point is what makes the sketch quick. For the absolute value graph that point is the corner, because the two arms are just straight lines through it at fixed slopes. Move the corner and the rest of the graph follows automatically.
Notice that one of these moves is horizontal and one is vertical, so you could do them in either order and land in the same place. That is not a general licence to shuffle transformations around, though. It works here only because the two moves are in different directions, and Part 5 is about the cases where the order really does matter.
Worked example
A curve has exactly the shape of the square root function, but it begins at the place where x is one and y is two.
Locate the parent's starting point
Why: The square root graph begins at the origin.
\[ f(x) = \sqrt{x}, \qquad f(0) = 0 \]
Measure the move between the two starting points
Why: From the origin to the new start is one across and two up.
\[ h = 1, \qquad k = 2 \]
Write it in function notation
Why: Right one is a minus one inside; up two is a plus two outside.
\[ h(x) = f(x-1) + 2 \]
Substitute the square root rule
Why: Replace x by x minus one inside the radical.
\[ h(x) = \sqrt{x-1} + 2 \]
Figure (svg): The square root curve starting at the origin and a second square root curve of the same shape starting at the point where x is one and y is two, with an arrow joining the two starting points.
Check the start and one more point
Why: The radicand is zero at input one, giving height two; at input five the radicand is four.
\[ h(1) = \sqrt{0}+2 = 2, \qquad h(5) = \sqrt{4}+2 = 4 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 110 — Example 1.56 and Figure 1.80
The square root graph has a corner too, at its left-hand end, and that corner does the same job the vertex did for the parabola. Find where the parent's corner is, find where this one is, and the difference between them is the pair of shifts.
From the origin to the new starting point is one across and two up. Right one is written as a minus one inside, up two as a plus two outside, and you now have the formula in function notation. Substituting the actual square root rule is the last and easiest step.
Do not skip the domain and range line at the end. A square root only accepts inputs that make the radicand zero or positive, so the domain starts at one rather than zero, and the outputs all sit at two or above. Those two intervals are a free check on whether you moved the graph the right way, and examiners often ask for them explicitly.
Concept
A horizontal shift slides the set of allowed inputs; a vertical shift slides the set of achieved outputs. Nothing else about either set changes, because a shift never bends or scales the picture.
\[ h(x) = \sqrt{x-1}+2: \quad \text{domain } [1,\infty), \quad \text{range } [2,\infty) \]
| what you write | domain | range |
|---|---|---|
| f(x) + k | unchanged | every value moves by k |
| f(x − h) | every value moves by h | unchanged |
| f(x − h) + k | moves by h | moves by k |
This is worth a line of your written answer whenever a question asks you to describe a transformation, because the two intervals are the quickest check that you read the directions the right way round.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 110 — the Analysis after Example 1.56
It is easy to treat domain and range as a separate topic that gets bolted onto the end of a question. Here they are genuinely useful, because they give you a numerical grip on a transformation that you might otherwise only have a picture of.
The pattern is exactly the inside and outside pattern again. A horizontal shift changes which inputs are allowed, so it moves the domain. A vertical shift changes which outputs are achieved, so it moves the range. Neither one touches the other, and neither one changes the shape of the interval, only where it starts.
Use this as a check whenever you are unsure of a direction. If you claim the square root graph moved right one, then the smallest allowed input must have gone up by one. If your stated domain still starts at zero, you have written the shift the wrong way round.
Fill the middle
Both of these start from the reciprocal function. Fill each blank with the number that belongs there, remembering that the horizontal one is written in reverse.
Fill in the blanks
g(x) = \dfrac11}} + 3, \qquad p(x) = \dfrac-4___}} + ___
Why: The first is the reciprocal moved one right and one up, so a right shift of one is written as minus one inside and the up shift as plus one outside. The second is moved three right and four down, so the inside carries a minus three and the outside a minus four. Both vertical asymptotes move with the inside number and both horizontal asymptotes move with the outside number.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 110-131 — Try It 1.35 and Exercise 306
Fill all four blanks before checking. The reciprocal function is a good test case because it has two asymptotes, and asymptotes move with the graph just like everything else.
The first pair is deliberately gentle: both shifts are one unit, so the only thing being tested is whether you put the minus sign inside and the plus sign outside. The second pair is the real question, because down four has to become a negative number on the outside while right three becomes a positive number after the minus inside.
When you have finished, picture both graphs. The vertical asymptote of the reciprocal sits wherever the denominator vanishes, so it moves with the inside number. The horizontal asymptote sits at whatever the function settles down to for huge inputs, so it moves with the outside number. Being able to say where both asymptotes went is the mark of having really understood the shift rather than just placed the symbols.
Sorting
Sort into buckets
Sort each transformation of f by the direction it moves the graph.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 131 — Exercises 307 to 314
Sort all six before checking any of them. The two large numbers, forty-nine and forty-three, are in there on purpose: they make it harder to picture the graph and force you to work from the structure of the formula instead.
Sort by asking two questions in order. Is the number inside the brackets or outside them? That decides horizontal against vertical. Then, what sign does it carry? For the outside ones the sign is literal, so plus goes up. For the inside ones the sign reverses, so plus goes left.
If you got one wrong, it was almost certainly one of the inside pair, and almost certainly because you read the sign directly. Go back to the trick of asking which input makes the inside zero, and check the answer that way instead.
Tweak it
Parameter explorer
The curve is the absolute value function with a horizontal shift and a vertical shift. Move each slider on its own first. Which one reverses the direction of the number, and where does the corner sit for any pair of values?
\[ y = \lvert x - ({h}) \rvert + ({k}) \]
Move one slider at a time, at least to begin with. Push the inside parameter up and watch the V slide to the right, then pull it below zero and watch it go left. Push the outside parameter and watch the whole V rise without moving sideways at all.
Once you have seen each on its own, set both to something awkward, like an inside value of negative three and an outside value of two, and predict where the corner will be before you look. The corner always sits at the point whose first coordinate is the inside parameter and whose second is the outside parameter, and that is the single most useful fact about this family of graphs.
Notice something the sliders make obvious that a static picture cannot: the two arms of the V keep exactly the same steepness no matter what you do. Shifts move a graph without deforming it, which is why they are the easiest transformations to draw and the safest to do in either order.
Section
Part 3
Concept
Figure (svg): Three square-root curves on one set of axes: the original rising to the right from the origin, its mirror image below the axis rising to the right, and its mirror image to the left of the vertical axis.
\[ g(x) = -f(x) \qquad \text{against} \qquad g(x) = f(-x) \]
vertical reflection — Negating the output, which produces the mirror image of the graph in the horizontal axis.
horizontal reflection — Negating the input, which produces the mirror image of the graph in the vertical axis.
The same inside-outside rule is running here. Outside, the minus acts on heights and flips the picture top to bottom. Inside, it acts on inputs and flips the picture left to right.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 110-111 — the Reflections box and Figure 1.81
Reflections are the easiest transformations to draw and the easiest to write down the wrong way, because both versions are just a minus sign and the two look almost identical on the page. The only difference is whether the minus is in front of the function or in front of the input.
In the picture, follow the original square root curve out to the right. The lower curve is its mirror image in the horizontal axis: same inputs, opposite heights. The left-hand curve is its mirror image in the vertical axis: same heights, opposite inputs. Each one is a fold along a different line.
The inside and outside rule is still running underneath. The outside minus acts on heights, so it flips top to bottom and changes the range. The inside minus acts on inputs, so it flips left to right and changes the domain. If you can say which of the two sets moved, you have identified the reflection correctly.
Worked example
\[ s(t) = \sqrt{t} \]
Reflect vertically by negating the output
Why: Every height becomes its own opposite, which is a flip across the horizontal axis.
\[ V(t) = -s(t) = -\sqrt{t} \]
Say what that did to the range
Why: The heights used to be zero and above; now they are zero and below.
\[ \text{range: } [0,\infty) \;\longmapsto\; (-\infty, 0] \]
Reflect horizontally by negating the input
Why: Every input becomes its own opposite, which is a flip across the vertical axis.
\[ H(t) = s(-t) = \sqrt{-t} \]
Say what that did to the domain
Why: A square root needs a radicand that is not negative, so now the input must be zero or less.
\[ \text{domain: } [0,\infty) \;\longmapsto\; (-\infty, 0] \]
Check one point in each
Why: Four is in the original domain; negative four is in the reflected one.
\[ V(4) = -\sqrt{4} = -2, \qquad H(-4) = \sqrt{4} = 2 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 111-113 — Example 1.57 and Figures 1.82 to 1.83
Work each reflection all the way through, including the sentence about the domain or the range, because that sentence is what turns a guess into an argument. It is also very often where the marks are.
For the vertical reflection the minus sits outside the radical, and it is the heights that change sign. The outputs used to run from zero upward; now they run from zero downward. The set of allowed inputs is untouched.
For the horizontal reflection the minus goes under the radical, and now the inputs change sign. This one has a consequence you should watch for: a square root needs a radicand that is zero or positive, so the allowed inputs become zero and everything below it. Reflections can move a restricted domain to a completely different part of the axis, and a graph that used to live on the right now lives on the left.
Comparison
\[ g(x) = -f(x), \qquad h(x) = f(-x) \]
Comparison matrix
| x | f(x) | g(x) = −f(x) | input for h giving the same output |
|---|---|---|---|
| 2 | 1 | -1 | -2 |
| 4 | 3 | −3 | −4 |
| 6 | 7 | -7 | −6 |
| 8 | 11 | −11 | -8 |
The third column keeps the inputs and flips the outputs. The fourth column keeps the outputs and flips the inputs, which is why the table for h has to be written with a new set of inputs.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 113-114 — Example 1.58 and Tables 1.30 to 1.32
Fill the blanks before revealing. The two right-hand columns are doing genuinely different jobs, and lining them up side by side is the fastest way to see it.
The third column keeps the inputs exactly as they are and flips every output. That is what negating the output means, and it is why that column is easy: one minus sign per row and nothing else changes.
The fourth column is the awkward one. To find the function that uses the negated input, you cannot keep the old inputs, because the new rule evaluates the original function at the opposite number. So the output eleven, which used to belong to the input eight, now belongs to the input negative eight. When a question asks you to tabulate this kind of reflection, the header row has to change, and forgetting that is the usual mistake.
Worked example
A common model for learning says that the percentage of mastery after t practice sessions is given below. It is built from the doubling function by three transformations at once.
\[ k(t) = -2^{-t} + 1 \]
Peel the formula apart from the inside out
Why: First the exponent is negated, then the whole power is negated, then one is added.
\[ 2^{t} \;\to\; 2^{-t} \;\to\; -2^{-t} \;\to\; -2^{-t}+1 \]
Name the three moves
Why: A reflection in the vertical axis, a reflection in the horizontal axis, and a lift of one.
\[ t \mapsto -t, \quad y \mapsto -y, \quad y \mapsto y+1 \]
Track the parent's crossing of the vertical axis
Why: It sits still under both reflections, then rises by one.
\[ (0,1) \to (0,1) \to (0,-1) \to (0,0) \]
Track a second point
Why: The point at input one goes left, then down, then up one.
\[ (1,2) \to (-1,2) \to (-1,-2) \to (-1,-1) \]
Figure (svg): The doubling curve rising steeply to the right, and the learning-model curve which starts low on the left, passes through the origin and flattens toward the dashed horizontal line at height one.
Check both points against the formula
Why: The printed book drops the minus signs on the second tracked point, listing it as the point with both coordinates positive one; substituting shows the correct landing place.
\[ k(0) = -1+1 = 0, \qquad k(-1) = -2^{1}+1 = -1 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 114-116 — Example 1.59 and Figures 1.84 to 1.85
Three transformations at once, and the safest way in is to peel the formula from the inside out rather than reading it left to right. The exponent is negated first, then the whole power is negated, then one is added. That order is the order of the three moves.
Follow the two tracked points through all three stages and write each stage down. The point on the vertical axis is easy: both reflections leave it where it is horizontally, and the final lift moves it from one down to negative one and then up to zero. The second point is the instructive one, because the horizontal reflection sends it to the other side of the axis.
One warning about the printed book here. It lists the second tracked point as ending at the place where both coordinates are positive one, and that is a misprint: the minus signs have been lost somewhere in typesetting. Substituting into the formula shows the correct landing place has both coordinates equal to negative one, which is what the picture on this slide draws. When a worked answer disagrees with a substitution you have checked, trust the substitution.
Prediction
\[ f(x) = x^2: \quad g(x) = -f(x), \quad h(x) = f(-x) \]
Predict first
Form both reflections of the squaring function. What do the two new graphs look like?
Correct: The output reflection turns it upside down; the input reflection leaves the graph unchanged
Why: Negating the output sends every height to its opposite, so the parabola opens downward. Negating the input squares a negative instead, and the square of a negative equals the square of the positive, so absolutely nothing moves. That second surprise has a name, and the rest of this part is about it.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 116 — Try It 1.38
Predict before you reveal, and predict both halves. Most people get the first half right immediately and are genuinely surprised by the second.
Negating the output turns every height into its opposite, so the parabola opens downward. Nothing subtle there. Negating the input, though, hands the squaring function a negative number instead of a positive one, and squaring destroys the difference. The reflected graph lands exactly on top of the original.
That is the observation the book wants you to notice, and it is the doorway into the next few slides. A graph that its own mirror image lands on has a name, and the name is even. Keep this parabola in mind as the picture of what even means.
Concept
Some graphs are unchanged by a reflection. When the mirror in the vertical axis puts the graph back on top of itself, the function is called even. When reflecting in both axes puts it back, the function is called odd.
\[ \text{even: } f(-x) = f(x) \qquad \text{odd: } f(-x) = -f(x) \]
even function — One whose graph is symmetric about the vertical axis: opposite inputs give the same output.
odd function — One whose graph is symmetric about the origin, so that a half turn about the origin leaves it unchanged: opposite inputs give opposite outputs.
Most functions are neither, and neither is a complete and correct answer. The only function that is both is the one that is zero everywhere.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 116-117 — the Even and Odd Functions box and Figure 1.86
These two words are not describing a transformation you perform; they describe a property a function either has or does not have. That distinction matters, because the test looks like a transformation and is easy to mistake for one.
Even means the mirror in the vertical axis sends the graph to itself, so opposite inputs give equal outputs. Odd means the half turn about the origin sends the graph to itself, so opposite inputs give opposite outputs. Both can be checked with one substitution and no drawing at all, which is the whole point of the algebraic version.
The most important sentence on this slide is the last one. Neither is not a failure or a dodge: it is the ordinary case, and the honest answer for most functions you will ever meet. Write it with confidence when it is true.
Picture it
Figure (svg): Two panels. On the left a parabola with two marked points at the same height on either side of the vertical axis. On the right a cubic with two marked points at opposite heights on either side of the origin, joined by a dashed line through the origin.
On the left, pick any input and its negative: the two heights are equal, so folding the page along the vertical axis makes the two halves coincide. On the right, the two heights are opposites, so folding does not work, but spinning the page half a turn about the origin does.
Pick a number on the horizontal axis of the left panel, find its height, then find the height at the negative of that number. The two are equal, and that is true no matter which number you pick. Equal heights on both sides is what a fold along the vertical axis looks like.
Now do the same on the right panel. The two heights are equal in size but opposite in sign, so folding along the vertical axis does not work. What does work is turning the whole page upside down about the origin, which sends every point to the point with both coordinates negated, and the cubic lands exactly on itself.
Carry the two pictures rather than the two equations. When you later meet a graph in a question with no formula attached, these are the two things you look for, and you can usually see either one in a second.
Notation
Annotate
On: \( f(-x) = f(x) \quad\text{or}\quad f(-x) = -f(x) \)
Work the substitution once and compare it with two things: the original, and the negative of the original. Two comparisons, three possible verdicts.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 117 — the how-to for determining even, odd or neither
There is one substitution and two comparisons, and running them in that order stops most of the mistakes. Replace every x by its negative, simplify carefully, and then stop. Do not negate anything yet.
Now compare what you got with two things. If it equals the original function, the function is even. If it equals the negative of the original, it is odd. The second comparison is easier if you write the negative of the original out in full rather than trying to do it in your head.
The fourth annotation is the one that costs marks. Failing the first comparison tells you only that the function is not even. It is not evidence for oddness, and the next slide but two has a worked case where the leap from one to the other gives the wrong answer.
Worked example
\[ f(x) = x^3 + 2x \]
Substitute minus x
Why: Cube the negative and double the negative.
\[ f(-x) = (-x)^3 + 2(-x) = -x^3 - 2x \]
Compare with f itself
Why: The two expressions differ in every sign, so the even test fails.
\[ -x^3-2x \ne x^3+2x \]
Negate that result
Why: Change the sign of both terms.
\[ -f(-x) = -(-x^3-2x) = x^3+2x \]
Compare again
Why: This is f exactly, so the odd test passes.
\[ -f(-x) = f(x) \;\Longrightarrow\; f \text{ is odd} \]
Figure (svg): The curve of x cubed plus two x, rising through the origin, with the point at x equal to one and height three marked and the point at x equal to negative one and height negative three marked, joined by a dashed line through the origin.
Check a pair of points
Why: Whatever height the graph has at input one, it has the opposite height at input negative one.
\[ f(1) = 1+2 = 3, \qquad f(-1) = -1-2 = -3 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 117-118 — Example 1.60 and Figure 1.87
Watch the signs in the very first line. Cubing a negative gives a negative, and doubling a negative gives a negative, so both terms flip. Writing the substitution out with the brackets still in place, before simplifying, is worth the extra second.
The first comparison fails, and notice that the write-up says so explicitly and then keeps going. Do not stop there and do not conclude anything from it. The second comparison is a fresh question: negate the result and see whether you land back on the original.
The picture confirms it in a way the algebra cannot. The curve passes through height three at input one and through height negative three at input negative one, and the dashed line joining those two points runs through the origin. Every pair of opposite inputs on this graph does the same thing, which is exactly what half-turn symmetry means.
Worked example
\[ f(s) = s^4 + 3s^2 + 7 \]
Substitute minus s
Why: A negative raised to an even power loses its sign, and the constant is untouched.
\[ f(-s) = (-s)^4 + 3(-s)^2 + 7 = s^4 + 3s^2 + 7 \]
Compare with f
Why: Identical term by term, so the even test passes.
\[ f(-s) = f(s) \;\Longrightarrow\; f \text{ is even} \]
Rule out odd as well
Why: The odd test would need f to equal its own negative, which a function that is never zero cannot do.
\[ -f(-s) = -s^4-3s^2-7 \ne f(s) \]
Check one pair numerically
Why: Two and negative two give the same height, as an even function must.
\[ f(2) = 16+12+7 = 35, \qquad f(-2) = 16+12+7 = 35 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 118 — Try It 1.39
This one tests whether you handle the constant term correctly. A constant has no x in it at all, so the substitution leaves it completely alone, and that is fine: a constant behaves like an even power, because seven is seven whichever input you feed in.
Both other terms survive the substitution unchanged too, because a negative raised to an even power loses its sign. So the substituted expression is identical to the original, term for term, and the even test passes on the first comparison.
The third step is worth keeping in your written answers even though the verdict is already settled. Ruling out odd takes one line and shows you know the two tests are separate. Here the reasoning is quick: this function is never zero, and a function that is never zero cannot possibly equal its own negative.
Sorting
Sort into buckets
Substitute minus x in each rule and sort it by the verdict.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 134 — Exercises 340 to 345
Sort all six before checking. The fast route is to look at the powers rather than substituting, but do the substitution for at least two of them to make sure the shortcut is one you have earned.
The shortcut is this: a polynomial with only even powers is even, one with only odd powers is odd, and a mixture is neither. A constant term counts as an even power, which is why adding a constant to an odd polynomial wrecks its symmetry.
Two of the six are not polynomials, and they are the interesting ones. The reciprocal added to three x is built from two odd pieces and comes out odd. The square root is the trap: it is not defined for negative inputs at all, so the value at the negated input does not exist, and a test you cannot even perform certainly does not pass. Neither is the answer.
Error analysis
Annotate
On: \( (x-2)^2: \; f(-x) \ne f(x) \;\Rightarrow\; \text{odd} \)
Read the argument and try to find the fault before revealing the annotations. That is harder than it sounds here, because the first half of the reasoning is genuinely correct.
The substitution really does show that this function is not even, and nothing is wrong with that line. The fault is the arrow: failing one test is being treated as evidence for the other, as though every function had to be one or the other.
Doing the odd test properly settles it. The negation of the substituted expression is negative everywhere, while the original is positive everywhere, so they cannot be equal and the odd test fails as well. The answer is neither, and the wider lesson is to distrust any argument that concludes a property from the failure of a different property.
Counterexample
Discussion prompt
A classmate says: a polynomial is even exactly when every exponent appearing in it is even. Find a polynomial that breaks this claim, say what goes wrong, and state the repaired version.
Write your counterexample before revealing. Almost any mixture works, and the simplest mixture is a square plus a first power.
Then read the second half of the question carefully, because finding the counterexample is only half the exercise. A good counterexample tells you what the repaired claim should say, and here the repair is the word only. A polynomial is even exactly when only even powers appear, and odd exactly when only odd powers appear.
The last line is the part people trip on later in the course. A constant term is an even power in disguise, so adding a number to an odd polynomial breaks its symmetry completely. Keep the pair of examples at the bottom of the slide in mind the first time you meet a cubic with a constant term.
Section
Part 4
Concept
Figure (svg): Three parabolas on one set of axes: the squaring function, a narrower one that is twice as tall at every input, and a wider one that is half as tall at every input.
\[ g(x) = a\,f(x), \qquad a > 0 \]
Shifting moved the graph without changing its shape. Scaling changes the shape, by pulling every point away from one of the axes or pushing it toward that axis.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 119 — the Vertical Stretches and Compressions box and Figure 1.88
Shifts moved a graph around without changing its shape. Scaling is the first transformation in this lesson that genuinely deforms the picture, and the way to see it is to look at what happens to the distance from each point to the horizontal axis.
Multiplying every output by a number bigger than one pulls every point further from that axis, so the graph gets taller and, for a parabola, looks narrower. Multiplying by a number between zero and one pushes every point toward the axis and the graph flattens. A negative multiplier does the scaling and a flip at the same time.
The last bullet is the one to remember when you are sketching. Anything already sitting at height zero cannot move, because multiplying zero by anything leaves zero. So the points where a graph crosses the horizontal axis are pinned in place by a vertical scaling, and they are the natural anchors for your sketch.
Worked example
A fruit-fly population P is recorded against time in hours. A scientist compares it with a second colony Q that follows the same pattern but is always twice as large.
Decide inside or outside
Why: The counts double; the timing of every rise and fall is unchanged. That is an output change.
\[ Q(t) = 2P(t) \]
Double the height of each reference point
Why: The first coordinates stay exactly where they are.
\[ (0,1) \to (0,2), \quad (3,3) \to (3,6), \quad (6,2) \to (6,4) \]
Watch the point on the axis
Why: Twice zero is zero, so a crossing of the horizontal axis does not move at all.
\[ (7,0) \to (7,0) \]
Figure (svg): Two piecewise-straight population graphs against time in hours: the original rising to three, easing to two and falling to zero at seven hours, and a second graph with exactly double the height at every time.
Check the factor at one time
Why: At three hours the second colony should be twice the first.
\[ Q(3) = 2P(3) = 2 \cdot 3 = 6 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 119-121 — Example 1.61 and Figures 1.89 to 1.90
Begin with the sentence, not the symbols. The second colony follows the same pattern but is always twice as large, and always twice as large means at every time, which tells you the counts are doubled and the timings are untouched. Output change, so outside, so vertical.
Doubling the four reference points is straightforward arithmetic, but look carefully at the last one. The population that hit zero at seven hours still hits zero at seven hours, because twice zero is zero. That point is pinned, exactly as the previous slide warned.
In the picture the two graphs have their corners at the same times, and only the heights differ. Get into the habit of checking that: if the times of the peaks have moved, you have made a horizontal change by accident, and the model you have written is describing a different biological claim from the one in the question.
Worked example
A curve has exactly the shape of the cubing function, and one point on it can be read off cleanly: the height is two when the input is two.
Evaluate the parent at the same input
Why: Two cubed is eight.
\[ f(x) = x^3, \qquad f(2) = 8 \]
Compare the two heights
Why: The new height is a quarter of the old one at the same input.
\[ g(2) = 2 = \tfrac14 \cdot 8 = \tfrac14 f(2) \]
Assume the same factor everywhere
Why: A vertical scaling multiplies every output by the same number, so one clean point fixes it.
\[ g(x) = \tfrac14 f(x) \]
Write the formula
Why: Substitute the cubing rule.
\[ g(x) = \tfrac14 x^3 \]
Check at a different input
Why: At input four the parent gives sixty-four, and a quarter of that is sixteen.
\[ g(4) = \tfrac14 (4)^3 = \tfrac14 (64) = 16 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 122 — Example 1.63 and Figure 1.91
Going from a picture to a scale factor needs one point you can read off cleanly, and the phrase relatively clear in the book is doing real work. Choose a point where the curve passes through a grid intersection, not one where you are guessing at a decimal.
Then the method is a single division. Work out what the parent function does at that same input, and divide the new height by the old one. Here two divided by eight is one quarter, so every output has been quartered.
The final check is what makes this legitimate rather than a guess. One point can only suggest a factor; testing a second point is what confirms that the same factor works everywhere, which is what a vertical scaling claims. If the second point disagreed, you would be looking at something other than a pure vertical scaling.
Concept
Figure (svg): Three parabolas on one set of axes: the squaring function, a narrow one produced by doubling the input before squaring, and a wide one produced by halving the input before squaring.
\[ g(x) = f(bx), \qquad b > 0 \]
Once again the inside behaves in reverse, and this time it is worse than a sign: the number you see is the reciprocal of the factor you get.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 123 — the Horizontal Stretches and Compressions box and Figure 1.92
This is the fourth and last of the four transformations, and it is the hardest of the four, so slow down here. Everything about it is reversed twice over: it is a horizontal change, so it acts on the input, and the number you see is not the size of the change but its reciprocal.
Look along a horizontal line in the picture, say at height four. The three parabolas cross that line at three different distances from the vertical axis, and the distances are in the ratio of one to two to four. The narrow one carries a factor of two inside and is half the width; the wide one carries a half and is twice the width.
The pinned points here are the ones on the vertical axis rather than the horizontal one, because multiplying zero by any factor still gives zero. That is the mirror image of the vertical case, and it is a useful anchor when you sketch.
Notation
Annotate
On: \( g(x) = f(bx), \qquad b > 0 \)
Read the first annotation as a sentence about time: whatever the original function does at some input, the new one does it at an input b times smaller. Everything happens sooner when b is large, and later when b is small.
Happening sooner means the graph is pulled in toward the vertical axis, which is a compression. Happening later means it is pushed outward, which is a stretch. Notice that you can get from the sentence to the direction without remembering any rule.
The last annotation is the one to write on your hand. The number inside is never the size of the stretch. If a question asks you to stretch by four, the thing you write is a quarter, and if you write four instead you have squeezed the graph to a sixteenth of the width you wanted.
Intuition
Take a function that reaches its peak at an input of six, and squeeze it so the peak arrives at an input of two. Ask what the squeezed rule has to do with the number two.
\[ g(2) = f(6) \;\Longrightarrow\; g(x) = f(3x) \]
It has to triple it, because that is the only way to hand the original rule the six it needs. So a graph squeezed to one third of its width carries the factor three, and the factor you see is three while the scaling you get is one third.
\[ \text{factor } b = 3 \;\Longrightarrow\; \text{width scaled by } \tfrac1b = \tfrac13 \]
The same argument runs in reverse for widening. Remember the sentence rather than the rule: the inside factor is what the input has to be multiplied by before the old rule will recognise it.
This is the same argument as the one for the inside shift, run with multiplication instead of subtraction, and it is worth doing once so that the reciprocal stops being arbitrary.
Start from what you want: the new graph should peak at an input of two, and the old one peaked at six. So the new rule, when handed a two, has to make the old rule see a six. The only way to do that is to multiply by three.
So a graph squeezed to a third of its width carries the factor three. The number you write and the change you get are reciprocals, and the reason is that the inside number is not describing the picture at all. It is describing what has to be done to the input before the old rule will recognise it.
Worked example
\[ g(x) = f\left(\tfrac12 x\right) \]
| x | 2 | 4 | 6 | 8 |
|---|---|---|---|---|
| f(x) | 1 | 3 | 7 | 11 |
Test the smallest input first
Why: Halving two gives one, and the table has no entry for f at one, so this input cannot be used.
\[ g(2) = f(1): \; \text{not known} \]
Find the inputs that do work
Why: The input to g must be twice an input that f already knows.
\[ g(4) = f\left(\tfrac12 \cdot 4\right) = f(2) = 1 \]
Finish the row
Why: Double each of f's inputs and copy the outputs straight across.
\[ g(8) = f(4) = 3, \quad g(12) = f(6) = 7, \quad g(16) = f(8) = 11 \]
| x | 4 | 8 | 12 | 16 |
|---|---|---|---|---|
| g(x) | 1 | 3 | 7 | 11 |
Check the direction of the change
Why: Every input doubled, so the picture is twice as wide, which is the reciprocal of the one half in the formula.
\[ b = \tfrac12 \;\Longrightarrow\; \text{stretch by } \tfrac{1}{b} = 2 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 124-125 — Example 1.65 and Tables 1.37 to 1.38
The first step here is one most write-ups leave out, and it is a good habit. Test the smallest input first and discover that it does not work: halving two gives one, and the table simply has no entry for the original function at one.
Saying that out loud stops you inventing data. The inputs you can use are the ones that halve to something the table already knows, so they must be double the original inputs: four, eight, twelve, sixteen.
When the row is complete, compare the two input rows. Every input doubled, so the picture is twice as wide, and yet the number written in the formula was a half. That is the reciprocal in the flesh, and seeing it happen in a table of real numbers is more convincing than any rule.
Worked example
Two curves of the same shape are drawn on one set of axes. The first reaches its last peak at the place where x is six and y is four; the second reaches the same height at x equal to two.
Compare the two end inputs
Why: Two is a third of six.
\[ 6 \cdot \tfrac13 = 2 \]
Check a second matching pair
Why: The two curves agree at inputs in the same ratio of one to three.
\[ g(2) = f(6), \qquad g(1) = f(3) \]
Read the rule off the pattern
Why: The new rule at x must do what the old one did at three x.
\[ g(x) = f(3x) \]
Name the transformation
Why: An inside factor of three squeezes the picture to a third of its width.
\[ b = 3 \;\Longrightarrow\; \text{horizontal compression by } \tfrac13 \]
Figure (svg): Two curves of the same shape starting at the origin: a wide one whose last peak is at x equal to six and height four, and a narrow one whose last peak is at x equal to two and the same height four.
Check against the reciprocal rule
Why: To widen by four you would write a quarter inside, so to narrow to a third you write three inside, which is what the picture shows.
\[ f\left(\tfrac14 x\right): \; \text{four times as wide} \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 126 — Example 1.66 and Figure 1.95
Recognising a horizontal scaling on a graph is a matter of finding matching features and comparing their inputs. The two curves here have identical heights and identical shapes, so the only thing that can have changed is the horizontal scale.
Pick the clearest matching feature, which is the final peak, and compare the two inputs: six and two. The ratio is one third, so every input of the original has been multiplied by a third to get the new one. Then check a second pair to make sure the ratio is constant, because a single pair could be a coincidence.
Now translate the ratio into a formula, and this is where the reversal bites. The picture is a third of the width, so the factor written inside is three, not a third. Reading the ratio off the graph and then writing it straight into the formula is the mistake to guard against.
Fill the middle
The first is the square root function stretched horizontally by a factor of three. The second is the square root reflected over the horizontal axis and stretched horizontally by a factor of two.
Fill in the blanks
g(x) = \sqrt32}}, \qquad q(x) = -\sqrt______}}
Why: A horizontal stretch by three needs an inside factor of one third, which is the same as dividing the input by three. A stretch by two needs division by two. The minus sign on the second one sits outside the radical, because reflecting over the horizontal axis negates the output, not the input.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 126-134 — Try It 1.42 and Exercise 362
Fill both blanks before checking. Both are horizontal stretches, so both need the reciprocal, and writing the division rather than a fractional coefficient often makes that easier to see.
A stretch by three means the picture is three times as wide, so every input has to be three times larger to produce what the original produced. Dividing the input by three is exactly that, and the same reasoning with a two gives the second answer.
Look at where the minus sign sits in the second formula. It is outside the radical, because reflecting over the horizontal axis negates the output. If you had put it under the radical you would have reflected over the vertical axis instead, which is a different graph on a different side of the plane.
Trap
The commonest slip once scaling starts:
\[ y = f(4x) \]
\[ \Rightarrow\; \text{stretch by } 4? \]
Wrong. Four inside squeezes the graph, to a quarter of its width.
Test it on a point. If the old graph peaked at input eight, where does the new one peak?
\[ g(2) = f(8) \]
\[ \text{peak at } 8 \to 2: \; \text{width } \tfrac14 \]
The peak arrives four times sooner, so the picture is four times narrower. Write the reciprocal down first and read the verdict off that.
This is the scaling version of the sign trap from Part 2, and it is more expensive, because getting it backwards changes the width by a factor of sixteen rather than merely pointing the graph the wrong way.
The cure is the same: refuse to read the answer off the number, and test it on a point instead. If the original graph peaked at an input of eight, ask where the new one peaks. The new rule needs its input multiplied by four to reach the eight the old rule wants, so the peak arrives at two.
A peak that used to be at eight and is now at two means the picture is a quarter as wide. Do that one-point test every time until the reciprocal becomes automatic, and write the reciprocal down before you write the word stretch or compression.
Tweak it
Parameter explorer
The parent here is a cubic with one hump and one dip. Move the outside factor a alone, then the inside factor b alone, and watch what each does to the height of the hump and to how far apart the hump and the dip sit.
\[ y = {a}\,f({b}x), \qquad f(u) = u^3 - 3u \]
The parent here has a hump and a dip, which is exactly what you need to tell the two scalings apart. On a parabola a vertical stretch and a horizontal squeeze look almost identical, and that similarity hides the difference you are trying to learn.
Move the outside factor on its own. The hump gets taller and the dip gets deeper, but the places where they occur do not budge, and neither do the three points where the curve crosses the horizontal axis. Now move the inside factor on its own. The heights of the hump and the dip stay fixed while the whole picture is dragged in toward the vertical axis.
Set them both away from one and try to say which is doing what before you look at the readout. Being able to separate the two effects by eye is the skill this slide is for, and it is the skill that makes graph-to-formula questions quick.
Edge cases
Parameter explorer
Slide the outside factor down through one, then through zero, and into the negatives. At which value does the parabola stop being a curve at all, and what happens to it on the far side of that value?
\[ y = {a}x^2 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 119 — the third bullet of the Vertical Stretches box
Start at one and come down slowly. The parabola flattens, and the flattening is smooth: there is nothing special about a half or a quarter, they are just progressively flatter curves.
Then reach zero. Every output becomes zero and the graph collapses onto the horizontal axis, becoming a straight line. That is the one value of the factor at which the shape is destroyed rather than changed, and it is worth pausing on, because it is also the reason the definition of a stretch excludes it.
Keep going into the negatives and the parabola reappears, opening downward. A negative factor is doing two things at once, a scaling by its size and a reflection over the horizontal axis, and the slider lets you see that those two are genuinely one continuous motion rather than two separate rules.
Matching
Match the pairs
Why: The same two numbers appear in all four rows, so the only thing deciding the answer is where each one is written. Outside the function it scales heights and acts literally; inside it scales widths and acts as a reciprocal.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 134 — Exercises 348 to 355
The four formulas use only two numbers, four and a quarter, and each appears twice. So nothing about the numbers themselves can decide the answers; the only information is where each one is written.
Work through them in pairs. The two with the multiplier outside are vertical, and they behave literally: four stretches the heights, a quarter flattens them. The two with the multiplier inside are horizontal, and they behave as reciprocals: four squeezes the width to a quarter, and a quarter widens it fourfold.
If you mismatched one, it was almost certainly an inside one, and almost certainly because you carried the number straight across. Say the reciprocal out loud before choosing, every time.
Section
Part 5
Concept
Shifting up three and then doubling is not the same as doubling and then shifting up three. When you shift first, the doubling doubles the shift as well.
\[ 2\bigl(f(x)+3\bigr) = 2f(x) + 6 \;\ne\; 2f(x) + 3 \]
So which order does a printed formula mean? The same order as the arithmetic. Given an output, you multiply before you add, so the stretch happens first and the shift afterwards.
\[ a\,f(x) + k: \quad \text{stretch by } a, \text{ then shift by } k \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 126-127 — the opening of Performing a Sequence of Transformations
Up to now you have been able to do transformations in any order you liked, because the combinations were one horizontal and one vertical. As soon as two of them act in the same direction, the order changes the answer.
The algebra on this slide shows why in one line. Shift first and the stretch that follows is applied to the shift as well, so the shift comes out doubled. Stretch first and the shift is added afterwards, untouched.
So which does a printed formula mean? Exactly what the order of operations says. Given an output you multiply before you add, so a formula that multiplies and then adds is telling you to stretch and then shift. That is not a convention you have to memorise separately; it is the same arithmetic rule you have used since primary school.
Picture it
Figure (svg): Three parabolas on one set of axes: the squaring function with its lowest point at the origin, a doubled one lifted by three, and a doubled one lifted by six, showing that shifting before stretching lifts the graph twice as far.
Both curves came from the same two instructions, doubling and lifting by three. The middle curve doubled first and then lifted, so its lowest point is at three. The outer curve lifted first, and the doubling that followed doubled the lift as well, putting its lowest point at six.
Both of the transformed curves came from the same two instructions: double it, and lift it by three. Only the order changed, and the two results are genuinely different graphs.
Look at where each one has its lowest point. The middle curve doubled first, so the three was added at the end and its vertex is at three. The outer curve lifted first, so the doubling doubled the lift as well and its vertex is at six. The little arrow is measuring the extra three that the stretch manufactured.
Notice also that the two curves have exactly the same width. Both were stretched by the same factor, so the shape is identical; the only difference is how high it sits. When two orders of the same instructions disagree, they usually disagree in one number like this, and it is worth finding which number before you decide you have made an error.
Intuition
The two reversals you have learned can feel like a lot to hold at once. Here is the relief: a horizontal move and a vertical move never affect each other, so you can do all the horizontal work first, or all the vertical work first, and get the same picture.
\[ 2f(x+1) - 3: \quad \text{left } 1 \text{ first, or stretch and drop first} \]
What you must not do is mix the order inside a single direction. Two vertical instructions have an order, and so do two horizontal ones. Across the two directions, order does not matter at all.
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 127 — the Combining Transformations box
After the last two slides it would be reasonable to conclude that order always matters and that you must be careful all the time. Half of that is wrong, and this slide is the relief.
A horizontal move changes where a point is along one axis, and a vertical move changes where it is along the other. Neither one can affect the other's coordinate, so it genuinely does not matter which block of work you do first. You may do all the horizontal work then all the vertical work, or the other way round.
What you must not do is shuffle the order within one direction. Two vertical instructions have a required order, and so do two horizontal ones. In practice this means you can treat a problem as two small problems, which is much less to hold in your head than one big one.
Concept
Figure (svg): Three parabolas: the squaring function with vertex at the origin, the narrow parabola of two x plus four all squared with vertex at x equal to negative two, and a dashed wrong guess with vertex at x equal to negative four.
\[ f(bx+p) = f\left(b\left(x + \frac{p}{b}\right)\right) \]
Written with two terms inside, the formula is very hard to read: the number added inside is not the shift, because it is about to be divided by the factor. Take the factor out and the two horizontal instructions separate cleanly.
\[ (2x+4)^2 = \bigl(2(x+2)\bigr)^2: \quad \text{squeeze to half, then left } 2 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 127 — the factoring discussion and the Combining Transformations box
This is the last skill of the lesson and the one that separates a formula you can read from a formula you cannot. When the inside of a function has two terms added together, neither of the two horizontal transformations can be read off directly.
The picture shows what goes wrong. The dashed curve is what you would draw if you took the four inside as a shift of four to the left. The real graph is only two to the left, because the four is sharing the bracket with a factor of two and is about to be divided by it.
Factoring fixes it in one line. Take the factor out and what is left inside the bracket is the honest shift, so the two horizontal instructions separate cleanly and can be applied in the order the previous slides established. Make factoring the reflex you reach for whenever you see two terms inside.
Worked example
\[ g(x) = 2f(3x) + 1 \]
| x | 6 | 12 | 18 | 24 |
|---|---|---|---|---|
| f(x) | 10 | 14 | 15 | 17 |
Start inside: the horizontal factor
Why: The rule f(3x) reaches each of f's values at an input three times smaller, so divide each input by three.
\[ 6,12,18,24 \;\longmapsto\; 2,4,6,8 \]
Apply the outside multiplier
Why: Double every output; the inputs are now fixed.
\[ 2f(3x): \; 20, \; 28, \; 30, \; 34 \]
Apply the outside addition last
Why: Multiplication before addition, so the plus one comes after the doubling.
\[ g(x): \; 21, \; 29, \; 31, \; 35 \]
| x | 2 | 4 | 6 | 8 |
|---|---|---|---|---|
| g(x) = 2f(3x) + 1 | 21 | 29 | 31 | 35 |
Check one entry straight from the definition
Why: Three times four is twelve, f of twelve is fourteen, doubled is twenty-eight, plus one is twenty-nine.
\[ g(4) = 2f(12)+1 = 2(14)+1 = 29 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 127-128 — Example 1.67 and Tables 1.39 to 1.42
Three transformations and a table, which is actually easier than three transformations and a graph, because every stage is a row of numbers you can check.
Work inside out. The inside factor is handled first, and it acts only on the input row: each input is divided by three so that the original function sees the numbers it knows. The outputs are carried across unchanged at this stage, which is the step people most often get wrong by doubling too early.
Then the outside work, in the order of operations. Multiply every output by two, then add one to every output. Doing the addition first would add one and then double it, which would land you two too high on every row. The final check, working one entry straight from the definition, is quick and catches exactly that error.
Worked example
The graph of f is the upper half of a circle of radius two centred at the origin, so it runs from the point at negative two on the axis, up to height two, and back down at positive two.
\[ k(x) = f\left(\tfrac12 x + 1\right) - 3 \]
Factor the inside
Why: Take the one half out of both terms.
\[ \tfrac12 x + 1 = \tfrac12 (x+2) \]
Rewrite the whole rule
Why: Now the inside reads as a factor followed by a shift.
\[ k(x) = f\left(\tfrac12 (x+2)\right) - 3 \]
Stretch horizontally first
Why: An inside factor of one half doubles every horizontal distance from the vertical axis, and twice zero is still zero.
\[ (-2,0) \to (-4,0), \quad (0,2) \to (0,2), \quad (2,0) \to (4,0) \]
Shift left two
Why: The plus two inside moves the whole widened arc two units to the left.
\[ (-4,0) \to (-6,0), \quad (0,2) \to (-2,2), \quad (4,0) \to (2,0) \]
Drop three
Why: The minus three outside lowers every point, and this is the last move.
\[ (-6,-3), \quad (-2,-1), \quad (2,-3) \]
Figure (svg): A half-circle arc of radius two above the origin, the same arc stretched to twice its width, and the final arc moved two to the left and three units down so that it sits below the horizontal axis.
Check the top of the arc from the formula
Why: At input negative two the inside is zero, f of zero is two, and two minus three is negative one.
\[ k(-2) = f(0) - 3 = 2 - 3 = -1 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, pp. 128-130 — Example 1.68 and Figures 1.96 to 1.99
This is the whole lesson in one problem, and the very first move is the one from two slides back: factor the inside. Until you do, the plus one looks like a shift of one, and it is really a shift of two.
Then work the stages in order and track three points rather than trying to redraw the curve each time. The two ends of the arc and its top are enough to place it. During the stretch, watch the top: it sits on the vertical axis, so twice its horizontal position is still zero and it does not move sideways, while both ends move outward to double the distance.
The dashed curve in the picture is the halfway stage, after the stretch but before either shift, and it is worth looking at on its own. The arc is now eight units across instead of four. Only after that does the whole widened shape slide two left and three down, so that it ends up entirely below the horizontal axis.
Worked example
Compress the reciprocal-square function vertically by one third, then move it two units left and three units down. Apply one instruction at a time and write the rule after each.
\[ f(x) = \frac{1}{x^2} \]
Apply the vertical compression
Why: A factor of one third multiplies the output, so it goes in front of the whole fraction.
\[ \tfrac13 f(x) = \frac{1}{3x^2} \]
Apply the horizontal shift
Why: Left two is a plus two inside, so every x becomes x plus two.
\[ \tfrac13 f(x+2) = \frac{1}{3(x+2)^2} \]
Apply the vertical shift
Why: Down three is a minus three outside, and it comes last.
\[ g(x) = \frac{1}{3(x+2)^2} - 3 \]
Check the two asymptotes
Why: The denominator vanishes at negative two, and for very large inputs the fraction dies away and only the minus three is left.
\[ g(-1) = \tfrac13 - 3 = -\tfrac83 \approx -2.667 \]
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 135 — Exercise 363
A description in words is the friendliest kind of question, because the instructions arrive already in order. Your job is to write down the formula after each one rather than trying to assemble the whole thing in a single line.
Take them literally and in order. Compress vertically by a third, so a third multiplies the whole fraction, which is the same as putting a three in the denominator. Move left two, so every x becomes x plus two, remembering that left is the plus. Drop three, so subtract three at the very end, outside everything.
The check at the end uses the two asymptotes, and it is the right check for this family. The vertical asymptote sits where the denominator vanishes, which the horizontal shift moved from zero to negative two. The horizontal asymptote is whatever the graph settles down to for enormous inputs, which the vertical shift moved from zero to negative three. If both asymptotes are where you expect, the shifts went the right way.
Ranking
Put in order
Order the steps of a complete sketch of a transformed toolkit function.
Why: Factoring comes before any drawing, because until the inside reads as a factor times a shifted input the shift you read off has the wrong size. After that each direction follows the order of operations: scale first, then shift. The two directions are independent, so the two horizontal steps and the two vertical steps may be swapped as blocks.
Drag all six into place before checking. The two steps whose position is genuinely worth arguing about are the first two, and they are both about preparation rather than drawing.
Naming the parent comes first because it tells you what shape you are about to move, and factoring comes second because until the inside is factored the shift you would read off is the wrong size. Do those two on paper before the pencil touches the axes.
After that each direction runs scale first, then shift, following the order of operations. The horizontal pair and the vertical pair can be swapped as blocks, so if you prefer to do all the vertical work first, the order you are checked against here is not the only correct one. What is not negotiable is the order inside each pair.
Step zero
\[ y = (3x-6)^2 + 1 \]
Discussion prompt
Before you draw anything at all, what is the single first move here, and why can you not read the horizontal shift straight off the formula as written?
OpenStax, Precalculus, §1.5 Transformation of Functions §1.5, p. 127 — the worked factoring discussion
Write your answer before revealing, and resist the pull to start sketching. The temptation with a formula like this is to see the six, call it a shift of six, and draw a parabola in completely the wrong place.
The six is not the shift, because it is sharing the bracket with a factor of three. Factoring the three out leaves a two behind, and two to the right is the real move. The three itself then squeezes the picture to a third of its width.
There is a fast way to confirm it that avoids the factoring altogether: ask which input makes the inside zero. Three x minus six vanishes at two, so the vertex is at two. Use that as a check on your factored form every time, because the two methods failing to agree is the clearest possible signal that something has gone wrong.
Real world
Discussion prompt
Your phone's battery level, as a percentage, t hours after you unplug it at full charge, is given by a function B. Two things change today: a screen-dimming app makes the battery last exactly twice as long, and you unplug at eighty percent instead of a hundred. Write each change as a transformation of B, and say which is inside and which is outside.
Write down both formulas before revealing. The question is chosen because the two changes are of opposite kinds and the words in the problem tell you which is which if you read them carefully.
Lasting twice as long is a statement about time, and time is the input, so it is an inside change. Then the reciprocal bites: the graph is twice as wide, so the factor written inside is a half, not a two. Starting at eighty rather than a hundred is a statement about the battery level, which is the output, so it is an outside change and it behaves literally.
The last line is worth noticing. Because one change is horizontal and the other vertical, you can apply them in either order and the final graph is identical. That is why real models can pile several transformations together without the modeller having to be careful about sequencing, as long as they are not stacking two changes in the same direction.
Section
Part 6
Pattern
Figure (svg): A flow diagram: name the parent function, factor the inside, then split into a horizontal branch that scales by one over b and shifts by h and a vertical branch that scales by a and shifts by k, both arriving at the finished graph.
This is the order to work in for any transformed toolkit function, and it is deliberately front-loaded with two steps that involve no drawing at all. Most sketching errors are committed before the pencil moves.
Naming the parent and factoring the inside are the preparation. Then the horizontal work, both parts of it reversed, and then the vertical work in the order of operations. Because the two directions are independent, the diagram splits into two branches that can be done in either order.
The last step is the one that catches the errors the first five let through. Pick one recognisable point on the parent, a corner or a vertex or an intercept, follow it all the way through your sequence, and then substitute its final input into the original formula. If the height you predicted and the height the formula gives are different, one of your directions is reversed, and you will know before you hand the work in.
Check
Check your understanding
The graph of a function f is moved 4 units to the LEFT. Which formula describes the new graph?
Answer: C
Why: A sideways move is written inside the function, and inside changes run in reverse: a plus 4 inside moves the graph 4 units to the left, because every feature now arrives at an input 4 smaller than before.
Two of these four options move the graph sideways and two move it up and down, so the first decision costs nothing: a sideways move must be written inside the function, which eliminates half the list immediately.
Then the reversal decides between the two survivors. Plus four inside means the rule subtracts nothing and adds four to the input before running, so every feature arrives at an input four smaller, which is four to the left. If you picked the minus, you read the sign directly instead of reversing it.
Check
Check your understanding
Which formula stretches the graph of f horizontally by a factor of 5?
Answer: C
Why: A sideways scaling is an inside multiplier, and the size of the stretch is its reciprocal. To widen by 5 the inside factor must be one fifth, so the input is divided by 5 and every feature arrives 5 times later.
Again start by splitting the options by position. A horizontal scaling must be an inside multiplier, so the two options with the five outside the function are describing vertical changes and can go.
Between the two that remain, the question is whether five inside stretches or squeezes. It squeezes, to a fifth of the width, so the one that stretches by five must divide the input by five instead. Writing the reciprocal down before choosing is the habit that makes this automatic.
Check
Check your understanding
Which of these functions is odd?
Answer: B
Why: Substituting minus x gives negative two x plus x cubed, which is exactly the negative of the original, so the odd test passes and the graph has half-turn symmetry about the origin.
Scan the powers before doing any algebra. Only even powers means even, only odd powers means odd, and a mixture means neither, and three of these four can be settled that way in a couple of seconds.
The fourth is the one that needs a different kind of thought. The square root is not defined for negative inputs, so the value at the negated input does not exist and neither test can even be attempted. A test you cannot perform is not a test you pass; the answer is neither, and that reasoning is worth writing out if a question asks.
Explain it to yourself
Discussion prompt
Two of the four parameters behave in reverse. In two or three sentences, explain why a change written inside the function reverses, without using the words rule or memorise.
Write your two or three sentences before revealing, and try to write them without using the word rule. If you can only justify the reversal by quoting a rule, you have not yet got the explanation, and the explanation is what survives an exam under pressure.
The core of it is a question of timing. An outside change happens after the function has finished, so it acts on the answer and does exactly what it says. An inside change happens before the function runs, so it changes which input the original rule is fed, and to make the original rule see the number it needs, the input handed over must be adjusted the other way.
The same sentence explains the reciprocal for stretches, with multiplication in place of subtraction. If you can say it once for shifts and once for scalings, you never have to remember four separate rules again.
Exit ticket
\[ g(x) = -2\sqrt{4-x} + 1 \]
Discussion prompt
Name every transformation of the square root function in this rule, in the order you would apply them, and give the coordinates of the point that the corner of the parent graph ends up at.
This puts the whole lesson into one formula, and the first thing to notice is that the inside is written back to front: the x has a minus in front of it. That is a reflection hiding in plain sight, and you cannot read the shift until you have pulled the minus out.
Factoring the negative one out of the inside turns four minus x into the negative of x minus four, so there is a reflection over the vertical axis and a shift of four to the right. Outside there are three instructions in the order of operations: stretch by two, flip over the horizontal axis, then lift by one.
The corner is the point to track, and it lands at the place where x is four and y is one. Check it by substituting: at input four the radicand is zero, so the whole first term vanishes and only the plus one is left. Finish by saying which way the curve opens, because the reflections are what decide that, and a correct corner with the curve drawn on the wrong side is still the wrong graph.
Recap
\[ g(x) = a\,f\bigl(b(x-h)\bigr) + k \]
| what you write | where it acts | what the graph does |
|---|---|---|
| f(x) + k | outside, on the output | moves up k, or down if k is negative |
| f(x − h) | inside, on the input | moves right h, or left if h is negative |
| a f(x) | outside, on the output | scales heights by a; flips over the x-axis if a is negative |
| f(bx) | inside, on the input | scales widths by 1/b; flips over the y-axis if b is negative |
| f(−x) = f(x) | a test, not a change | even: mirror symmetry in the y-axis |
| f(−x) = −f(x) | a test, not a change | odd: half-turn symmetry about the origin |
Two sentences carry the whole section. Outside changes act on the output, are vertical, and do what they say. Inside changes act on the input, are horizontal, and do the opposite: signs flip and factors invert.
Next, Section 1.6 takes the absolute value function and runs this whole machinery on it, which is where these four parameters first start earning their keep.
OpenStax Algebra and Trigonometry 2e, §3.5 Transformation of Functions §3.5 — the same material in Algebra and Trigonometry
Everything in this section is in the one line at the top of the slide. Four letters, four transformations, and two of them reversed. If you can look at a formula, match it against that line, and name what each number does, you can graph anything built from a toolkit function.
The table is your reference while the rules are still settling. Read the middle column first, because where a number is written decides everything else about it: outside means the output, which means vertical, which means literal; inside means the input, which means horizontal, which means reversed.
The two symmetry tests at the bottom are a different kind of statement, and it is worth keeping them mentally separate. They are not things you do to a graph; they are questions you ask about one. Most functions answer neither, and that is a complete answer.
Want this taught 1-on-1? Alexander tutors Precalculus — $55/session, free consultation.