4.10 Writing Quadratic Functions from Their Graphs and Data

Building a quadratic function from a vertex and a point, from two x-intercepts and a point, and from three arbitrary points using a system of equations; choosing the right form for the information given; and finding a best-fitting quadratic model from data.

Subject: Algebra 2 · 64 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Write Quadratic Functions and Models

Title

Algebra 2 · Lesson 4.10

The whole chapter, run backwards: from a picture or a table to the function

2. What this lesson gives you

Objectives

For nine lessons you have been handed a quadratic function and asked what its graph looks like. This lesson turns that round. Somebody shows you a graph, or a handful of points, or a table of measurements, and asks you for the function.

Nothing new has to be learned to do it. You already know the three forms from Lesson 4.2, and you already know how to solve a system of three equations from Lesson 3.4. What is new is choosing which of those tools the given information calls for.

The habit worth building from the first slide is this: decide the form before you write anything down. Almost every wasted page in this lesson comes from starting to substitute into a form the information does not suit.

The second habit is checking. Every function you write here must be tested against all of the points you were given, not just the last one you used, because a slip in a sign will still satisfy one of them.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 309 — Before, Now, Why and the Key Vocabulary

3. First, recall the three forms

Warm-up

Before any new work, get the three forms back in front of you. Write them down from memory and then read on; committing to paper first is what makes the comparison worth anything.

Discussion prompt

Write out the three forms of a quadratic function you met in Lessons 4.1 and 4.2, and beside each one say what you can read off it at a glance.

Look at what those three sentences have in common. Each form was built so that the numbers you most want to read are sitting right there in the formula, not hidden behind a calculation.

That is the fact this whole lesson exploits. If somebody tells you the vertex, then the form with the vertex written into it is the one that needs the least work, because two of its numbers are already known before you start.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.2, p. 245 — the three forms, as Lesson 4.2 introduced them

4. Which form the picture hands you

Section

Part 1

5. Running the chapter backwards

Concept

Everything in Chapter 4 so far has pointed one way. You were given a function, and you worked out the vertex, the crossings, the direction and the width. Reading a function is what lessons 4.1 and 4.2 taught.

\[ y = 2x^2 + x - 4 \;\longrightarrow\; \text{vertex, intercepts, shape} \]

Now the arrow reverses. You are handed some of the vertex, the crossings and a few points, and asked to produce the function that has them.

\[ \text{vertex, intercepts, points} \;\longrightarrow\; y = 2x^2 + x - 4 \]

This is a genuinely different kind of question, and it is the kind that real work asks. Nobody measuring a satellite dish or timing a tossed pumpkin is handed a formula; they are handed measurements, and the formula is what they have to build.

The route back is always the same in outline. Pick the form that already contains what you were told, substitute the known numbers into it, then use one further point to settle whatever is left unknown.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 309 — In this lesson you will write quadratic functions given information about their graphs

6. Three pictures, three routes

Picture it

Here are the three situations the lesson deals with, side by side. Look at them before reading the labels and ask yourself what is marked on each one, because that is the only thing that decides which route you take.

Figure (svg): Three small graphs. The first shows a parabola with only its vertex marked and one more point. The second shows a parabola with both of its x-intercepts marked and one more point. The third shows three scattered points with no curve drawn at all.

Read left to right: each picture hands you different information, and each one makes a different form of the quadratic the cheap one to use.

In the first picture the lowest point is marked, so vertex form already knows two of its three numbers. In the second the two crossings are marked, so intercept form already knows two of its three. In the third nothing special is marked at all.

That third picture is the expensive one. With three ordinary points you know none of the vertex, none of the crossings, and so you have to find all three coefficients of standard form at once, which takes a system of equations rather than a single substitution.

So the choice is not about taste. It is about how much the form already knows before you start, and the picture is what tells you.

7. Vertex form, letter by letter

Notation

Take vertex form apart before using it backwards. When you substitute into a formula you have decoded, a misplaced sign becomes obvious instead of invisible.

Annotate

On: \( y = a(x - h)^2 + k \)

  • The vertex is at the point with x equal to h and y equal to k. Given a vertex, these two are known immediately and never have to be solved for.
  • It is always written as x minus h. A vertex at x equal to negative 3 therefore appears as x minus negative 3, which is x plus 3. This single sign is the most common error on the whole topic.
  • The only unknown left once the vertex is in place. Its sign says which way the curve opens and its size says how narrow the curve is.
  • One unknown needs one equation, and one extra point on the curve gives exactly one equation. That is the arithmetic behind the whole method.

Read the last note again, because it generalises. Count the unknowns left after you substitute what you know, and you have counted the number of extra points you need.

Vertex form leaves one unknown, so it needs one extra point. Intercept form leaves one unknown, so it needs one extra point. Standard form with nothing known leaves three, so it needs three points, and that is where Lesson 3.4 comes back.

\[ y = a(x - h)^2 + k \;\xrightarrow{\;h, k \text{ known}\;} \; \text{one unknown, one point} \]

It is worth being clear about what the squaring does here as well. Because the bracket is squared, two inputs the same distance either side of h give the same output, and that is the symmetry every parabola has. Vertex form puts the mirror line at x equal to h, in plain sight.

One last warning about the sign of a. Its size does not tell you how steep the curve is at a particular place; it tells you how the curve grows relative to the vertex. A value of a below one gives a wide curve and a value above one a narrow one, which is exactly what you will see in the next picture.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.2, p. 245 — vertex form and what h and k mean

8. A vertex does not fix a parabola

Concept

It is tempting to think the vertex is the whole story, since it is the most striking feature of a parabola. It is not, and this picture is the fastest way to be convinced.

Figure (svg): Three parabolas sharing the vertex one, minus two, with different widths, and a single gold point at three, two that only the middle one passes through.

Every curve here has the same vertex, so the vertex alone cannot say which parabola you want. Only one of them reaches the gold point.

All three curves have their lowest point at the same place. They are still different functions, because a is different in each, and a is what decides how fast the curve climbs away from that lowest point.

Only one of them passes through the gold point. That is what the extra point is for: it is not decoration, it is the piece of information that selects one curve out of infinitely many.

\[ y = a(x - 1)^2 - 2 \qquad a = 0.4, \; 1, \; 2.2 \]

The same argument works for intercept form. Two crossings fix where a parabola meets the axis but say nothing about how deep the curve dips between them, so again one more point is needed.

9. Which form does each clue set call for?

Sorting

Six sets of information. For each one, ask which form already contains what you have been told, and sort it there. Do not solve anything yet; this is the decision that comes before the algebra.

Sort into buckets

Sort each set of given information by the form it makes cheapest.

Vertex form
vertex (4, −5); passes through (2, −1); lowest point (3, 2); passes through (5, 6)
Intercept form
x-intercepts −2 and 5; passes through (6, 2); crosses the axis at −3 and 3; passes through (1, −4)
Standard form, by a system
passes through (−1, 5), (0, −1) and (2, 11); passes through (−4, −6), (0, −2) and (2, 6)
v
A vertex, or a lowest or highest point, is exactly what vertex form has written into it; substituting it leaves only a to find.
i
Both crossings named, in any wording, is what intercept form has written into it; substituting them leaves only a to find.
s
Three ordinary points name neither the vertex nor the crossings, so all three coefficients are unknown and a system is the only way.

Notice how the wording varies. The lowest point, the highest point and the vertex are the same thing, and crosses the axis at, the x-intercepts are and the zeros are also all the same thing.

Questions rarely use the textbook word. Learning to hear the vertex behind the phrase the lowest point is a real part of this lesson, and it is the part that makes exam questions feel unfamiliar when it is missing.

Count the points as well, because that is the other giveaway. A vertex and one point is two pieces of information, and two crossings and one point is three; neither needs a system. Three loose points always does.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 1 to 6, whose wording this sorting copies

10. Before any algebra: what would you write first?

Step zero

Step zero on any of these problems is the same. Do not reach for a formula; read the information and name the form. Try it here, and write your first line before you look at the answer.

\[ \text{vertex } (5, -3); \quad \text{passes through } (1, 5) \]

Discussion prompt

Name the form you would use, write the very first line you would put on the page, and say what would still be unknown on that line.

Look at how little work that was. The vertex went straight into the formula without any solving at all, because vertex form was built to hold it.

Now imagine having started in standard form instead. You would be setting up three unknowns and hunting for three points, when two of the three numbers were handed to you at the start. The decision at step zero is worth more than any of the algebra after it.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 312 — Exercise 15

11. From a vertex and a point

Section

Part 2

12. Substitute the vertex, then solve for a

Concept

The method has exactly three moves, and it is worth naming them before seeing them work, so that the example reads as a procedure rather than a trick.

  1. Write vertex form and put the vertex straight in, as h and k
  2. Substitute the extra point for x and y, which leaves one equation in a alone
  3. Solve for a and write the finished function, then test it on the point you used

The second move is the one to watch. You are substituting a point into an equation that already has the vertex in it, so the only letter left standing is a, and what looks like an equation in two unknowns is really an equation in one.

\[ y = a(x - h)^2 + k \;\longrightarrow\; \underbrace{y_1 = a(x_1 - h)^2 + k}_{\text{only } a \text{ is unknown}} \]

It is also worth saying what you do not do. You never expand the bracket before finding a; expanding turns one tidy equation into a longer one with the same single unknown, and costs you two lines for nothing.

The third move ends with a test, and it should test both facts rather than one. Substituting the vertex back must return the vertex height, and substituting the extra point back must return its height. A slip in the sign of h fails the first test while passing nothing at all, and a slip in a fails the second.

One more piece of advice about presentation. Write the form down first, with letters, before you put any numbers in. Students who substitute straight into a half-remembered formula are the ones who lose the sign on h, because they never see the subtraction written out.

13. Example 1: a vertex at (1, −2) and the point (3, 2)

Worked example

The graph shows a parabola whose lowest point is at one, negative two, and which also passes through three, two. Write a function for it. The vertex is named, so the form is chosen before any work starts.

\[ y = a(x - h)^2 + k \]

Substitute 1 for h and negative 2 for k

Why: the vertex goes in directly; this is the whole reason for choosing this form.

\[ y = a(x - 1)^2 - 2 \]

Two of the three numbers are now settled, and the line already describes every parabola with that vertex. The extra point is what picks one of them out.

Substitute 3 for x and 2 for y

Why: the other given point lies on the curve, so its coordinates must satisfy the equation.

\[ 2 = a(3 - 1)^2 - 2 \]

Simplify the coefficient of a

Why: three minus one is two, and two squared is four.

\[ 2 = 4a - 2 \]

Solve for a

Why: add 2 to both sides, then divide by 4.

\[ 4 = 4a \;\Longrightarrow\; a = 1 \]

\[ y = (x - 1)^2 - 2 \]

Figure (svg): The parabola y equals x minus one, all squared, minus two, drawn on axes with its vertex at one, minus two marked in gold and the second given point three, two marked in green.

The vertex fixes where the parabola sits; the green point fixes how wide it is. Two pieces of information, two unknowns settled.

Look at the finished curve against the two marked points. The vertex sits where it was told to, and the green point is on the curve, which is the only evidence that the value of a is right.

A value of a equal to one means this is the plain parabola, moved but not stretched: one step right of the vertex it rises one, two steps right it rises four. That is a useful benchmark to carry around, since every other value of a is read against it.

If a standard form is wanted, expanding takes one line: x squared minus two x plus one, minus two, which is x squared minus two x minus one. The constant term of negative one is the height at which the curve meets the vertical axis, and the picture confirms it.

Check both given facts

Why: put x equal to 1 into the answer and you get negative 2, the vertex height; put x equal to 3 and you get 2, the given point. Both facts the question gave are reproduced, so the function is right.

\[ (1 - 1)^2 - 2 = -2, \qquad (3 - 1)^2 - 2 = 4 - 2 = 2 \]

Checking both, rather than only the point you used last, is the habit to build now. A wrong sign on k would still satisfy the point and fail the vertex.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 309 — Example 1

14. Guided Practice 1: vertex (4, −5) through (2, −1)

Fill the middle

The same three moves on new numbers. Fill in the coefficient and then the finished function; the arithmetic is two lines, and the point of the exercise is the setup rather than the sums.

Fill in the blanks

-1 = a(2 - 4)^2 - 5 \;\Longrightarrow\; 4a = 4 \;\Longrightarrow\; a = 1

Why: Two minus four is negative two, and negative two squared is positive four, so the equation is negative one equals four a minus five. Adding five to both sides gives four a equal to four, so a is one and the function is y equal to the quantity x minus four, squared, minus five.

\[ y = (x - 4)^2 - 5 \]

The trap hiding in this one is the squaring. Two minus four is negative two, and squaring it makes it positive four; a student who writes negative four here gets a equal to negative one and a parabola opening the wrong way.

A picture settles it without any algebra. The given point sits above the vertex, at height negative one against negative five, so the curve must rise away from the vertex and a has to be positive.

It is worth converting the answer to standard form once, because many questions ask for that: expanding gives x squared minus eight x plus sixteen, minus five, which is x squared minus eight x plus eleven. Test it at x equal to two and you get four minus sixteen plus eleven, which is negative one.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 1

15. Guided Practice 2: vertex (−3, 1) through (0, −8)

Worked example

A negative number in the vertex, which is where the sign inside the bracket starts to matter. Read the form as x minus h and substitute negative 3 for h without shortcutting.

\[ y = a(x - h)^2 + k \]

Substitute negative 3 for h and 1 for k

Why: x minus negative three is x plus three; write the double negative out once and then simplify it, rather than guessing the sign.

\[ y = a(x - (-3))^2 + 1 = a(x + 3)^2 + 1 \]

That is the sign rule in action. A vertex sitting to the left of the origin produces a plus inside the bracket, which looks wrong to the eye and is right every time.

Substitute 0 for x and negative 8 for y

Why: the curve passes through the point where it crosses the vertical axis, and that is the extra fact you were given.

\[ -8 = a(0 + 3)^2 + 1 \]

Simplify the coefficient of a

Why: zero plus three is three, and three squared is nine.

\[ -8 = 9a + 1 \]

Solve for a

Why: subtract 1 from both sides, then divide by 9.

\[ -9 = 9a \;\Longrightarrow\; a = -1 \]

\[ y = -(x + 3)^2 + 1 \]

Figure (svg): The downward parabola y equals minus one times x plus three, all squared, plus one, with its vertex at minus three, one and the given point zero, minus eight marked.

The given point sits far below the vertex, so the curve must fall away from the vertex: a comes out negative, and the picture said so first.

The negative value of a was predictable from the picture. The given point lies nine units below the vertex, so the curve has to fall away from the vertex, and only a negative a does that.

Check both given facts

Why: at x equal to negative 3 the bracket is zero and y is 1, the vertex height; at x equal to 0 the bracket is 3, so y is negative nine plus one, which is negative eight.

\[ -(0 + 3)^2 + 1 = -9 + 1 = -8 \]

Notice this curve does cross the x-axis, at negative two and negative four, so it could also have been written in intercept form. Choosing vertex form was cheaper only because the vertex is what you were told.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 2

16. Find the error: a vertex poured into intercept form

Error analysis

This is the book's own error-analysis exercise, and the mistake in it is the one this part exists to prevent. Every line is arithmetically correct, and the answer is still wrong.

Annotate

On: \( y = a(x - 2)(x - 3) \;\Longrightarrow\; 5 = a(1 - 2)(1 - 3) \;\Longrightarrow\; a = \tfrac52 \)

  • This is intercept form, and the numbers 2 and 3 have been taken from the vertex (2, 3). But 2 and 3 are the x and y of one single point, not two x-intercepts, so they cannot both go in as crossings.
  • The form was chosen before the information was read. A vertex is one point; intercept form wants two points, and both of them on the x-axis.
  • Use vertex form: y equals a times the quantity x minus 2, squared, plus 3. Now the 2 is an x and the 3 is a y, which is what they actually are.
  • Substitute the given point, five when x is one: 5 equals a times one, plus 3, so a is 2 and the function is y equals 2 times the quantity x minus 2, squared, plus 3.

\[ y = 2(x - 2)^2 + 3 \]

Test the wrong answer and the right one against the vertex and you can see the difference immediately. The corrected function gives 3 at x equal to 2, which is the vertex height, and the wrong one gives zero there, because it was built to cross the axis at 2.

The general lesson is about reading, not algebra. Before substituting, say out loud what each number is: is this an x-coordinate, a y-coordinate, or an x-intercept? Numbers in a question do not carry labels, so you have to supply them.

It is also worth noticing which form the wrong work would have suited. Intercept form needs two numbers that are both x-intercepts, and here there were none, because a vertex above the axis on a curve opening up means no crossings at all.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 27

17. Exercise 5: reading a vertex straight off a grid

Worked example

Often nobody tells you the numbers; a graph is printed and you have to take them off it. The skill is choosing two points you can read exactly, and lattice points where the curve crosses a grid corner are the ones to take.

Figure (svg): An upward parabola with its lowest point at minus one, minus three and a second marked point at one, minus one, drawn on a grid.

Everything you need is already printed on the grid: the lowest point and one lattice point the curve visibly passes through.

Two readings are all you need here. The lowest point sits at negative one, negative three, and the curve visibly passes through the lattice point at one, negative one.

\[ y = a(x + 1)^2 - 3 \]

Substitute 1 for x and negative 1 for y

Why: using the second lattice point, the only other exact reading on the picture.

\[ -1 = a(1 + 1)^2 - 3 \]

Simplify the coefficient of a

Why: one plus one is two, and two squared is four.

\[ -1 = 4a - 3 \]

Solve for a

Why: add 3 to both sides, then divide by 4.

\[ 2 = 4a \;\Longrightarrow\; a = \tfrac12 \]

\[ y = \tfrac12 (x + 1)^2 - 3 \]

A value of a below one says the curve is wider than the plain parabola, which is what the drawing shows: it takes two steps sideways to climb one step up near the vertex.

That agreement between the number and the picture is the point of reading graphs at all. Before trusting any value of a you compute, glance at how steeply the drawn curve leaves its vertex, and ask whether a fraction, a one or a large number is what you should expect.

Notice also which reading was risky and which was safe. The vertex sits exactly on a grid corner and so does the second point, and both were taken from corners rather than estimated between them. A reading taken halfway between two grid lines would have put the error straight into a.

Check on a third reading

Why: the picture is symmetric about x equal to negative one, so the curve should also pass through negative three, negative one; substituting gives a half of four, minus three, which is negative one.

\[ \tfrac12(-3 + 1)^2 - 3 = \tfrac12(4) - 3 = -1 \]

Using a third point rather than the one you solved with is what makes this a real check. The point you used will fit whatever you did to it.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 312 — Exercise 5

18. Slide a and watch the parabola through the vertex

Tweak it

Here is the family from the earlier picture, live. The vertex is pinned at one, negative two, and a is yours to move. Watch what changes and, more importantly, what does not.

Parameter explorer

Move a and find the value that sends the curve through the point three, two. Then say what stays fixed no matter what a does.

\[ y = {a}(x - 1)^2 - 2 \]

  • a — from -2 to 3: a

The vertex never moves, whatever you do to a. That is the point of vertex form: h and k own the position, and a owns nothing but the shape.

Push a through zero and watch the curve flip. Positive values open the curve upward, negative values downward, and at a equal to zero the parabola collapses to a horizontal line, which is a warning about the collinear case waiting in Part 4.

Try to hit the point three, two by eye and you will find only one value works. That is the picture of what solving for a does: the extra point is a single condition, and a single condition on one unknown has one answer.

19. Which other point is on this parabola?

Prediction

A vertex, one point and a list of candidates. Write the function first; guessing from the picture in your head is exactly what this question is designed to punish.

\[ \text{vertex } (5, -3), \quad \text{through } (1, 5) \]

Predict first

The vertex is at (5, -3) and the parabola passes through (1, 5). Which of these points is also on it?

  • (0, 3)
  • (-1, 9)
  • (-1, 15)
  • (7, 7)

Correct: (-1, 15)

Why: Substituting the given point gives 5 equal to 16a minus 3, so a is one half and the function is y equal to one half times the quantity x minus 5, squared, minus 3. Testing each candidate: at x equal to 0 the value is 9.5, at x equal to negative 1 it is 15, and at x equal to 7 it is negative 1. Only the candidate at x equal to negative 1 with height 15 fits.

\[ y = \tfrac12 (x - 5)^2 - 3 \]

Two of the candidates share an x-coordinate, which is the trap. It forces you to compute a height rather than recognise a shape, and a height is exactly what the function is for.

There is a shortcut worth knowing, though it is a check rather than a method. The given point at x equal to one is four units left of the vertex, and the candidate at x equal to negative one is six units left, so its height above the vertex should be nine quarters of the given one: nine quarters of eight is eighteen, and negative three plus eighteen is fifteen.

That shortcut is the squared in vertex form showing itself. Double the distance from the vertex and the height above the vertex goes up four times, which is worth having as a mental picture even when you do the algebra properly.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 312 — Exercise 15

20. Trap: the sign inside the bracket

Trap

The trap

Vertex at negative four, one. A very common first line:

\[ y = a(x - 4)^2 + 1 \]

The negative sign has quietly been dropped, because x minus negative four looks wrong written out.

Everything after this line will be arithmetically perfect and the answer will still be a parabola in the wrong place.

The fix

The form is x minus h, always. With h equal to negative four, that reads x minus negative four.

\[ y = a(x + 4)^2 + 1 \]

One test catches it in a second: put x equal to negative four into your line and see whether you get the vertex height of one.

The wrong line gives sixty-four a plus one there, not one, so it does not even have the vertex it was built from.

The same trap works in reverse. A written bracket of x plus 4 always means a vertex at negative 4, so reading a finished answer is the mirror image of writing one.

Make that substitution every time you write a vertex form, before you go looking for a. It costs one line and it is the only check that can catch this particular mistake.

21. From two x-intercepts and a point

Section

Part 3

22. What p and q really are

Concept

Intercept form is the second of the three, and it is the one whose numbers are easiest to misread, so start with what the letters mean rather than with the method.

\[ y = a(x - p)(x - q) \]

Put x equal to p and the first bracket is zero, so y is zero. Put x equal to q and the second bracket is zero, so y is zero again. That is the entire content of the form: p and q are the two inputs that send the output to zero.

Those inputs have several names, and questions use all of them. The x-intercepts, the zeros, the roots and the places where the graph crosses the horizontal axis are four phrases for the same pair of numbers.

\[ x = p \;\Longrightarrow\; y = 0, \qquad x = q \;\Longrightarrow\; y = 0 \]

As with vertex form, substituting the two crossings settles two of the three numbers and leaves a alone, so one extra point off the axis finishes the job. The extra point must be off the axis: a third point with y equal to zero would tell you nothing new.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.2, p. 248 — intercept form and the meaning of p and q

23. Intercept form, bracket by bracket

Notation

The same decoding pass as for vertex form, because the same sign problem lives in these brackets. Take one minute over it now and save a wrong answer later.

Annotate

On: \( y = a(x - p)(x - q) \)

  • The two x-intercepts. They are interchangeable: writing the brackets in the other order gives exactly the same function, because multiplication does not care about order.
  • Both brackets are written as x minus something. A crossing at negative 2 appears as x minus negative 2, that is x plus 2. A bracket reading x plus 2 always means a crossing at negative 2.
  • Still the only unknown after the crossings go in, and still the number that says which way the curve opens and how deep it dips between the two crossings.
  • The vertex sits exactly halfway between p and q, at x equal to p plus q over 2. So intercept form quietly tells you the vertex as well, which is useful once a is known.

The last note is worth keeping. Intercept form is not a worse form than vertex form; it simply stores the same parabola using different landmarks, and each landmark can be recovered from the other.

\[ x_{\text{vertex}} = \dfrac{p + q}{2} \]

The interchangeability in the first note matters too. If two students write the brackets in opposite orders, both are right, and expanding will show that the two expressions are identical. Nothing in the method depends on which crossing you call p.

The book prints an Avoid Errors note right beside this form, and it is worth taking seriously: be sure you substitute the x-intercepts and the coordinates of the given point for the correct letters. The numbers in a problem do not announce which is which, and putting a point where a crossing belongs is the single most common way this form goes wrong.

One more thing this form makes easy. Because y is a product of a, one bracket and the other bracket, you can read off the sign of y anywhere just by counting how many of the three parts are negative, which is how you know the curve is above the axis between the crossings when a is negative.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 309 — the Avoid Errors note beside Example 2

24. Example 2: crossings at −1 and 4, through (3, 2)

Worked example

The graph shows a parabola meeting the horizontal axis at negative one and at four, and passing through three, two. The crossings are given, so intercept form is the cheap route.

\[ y = a(x - p)(x - q) \]

Substitute negative 1 for p and 4 for q

Why: x minus negative one is x plus one; the other bracket keeps its minus sign because 4 is positive.

\[ y = a(x + 1)(x - 4) \]

Two of the three numbers are in place. This line describes every parabola crossing at those two points, wide or narrow, opening up or down.

Substitute 3 for x and 2 for y

Why: the extra given point is on the curve, so its coordinates satisfy the equation.

\[ 2 = a(3 + 1)(3 - 4) \]

Simplify the coefficient of a

Why: three plus one is four and three minus four is negative one, so the product is negative four.

\[ 2 = -4a \]

Solve for a

Why: divide both sides by negative 4.

\[ a = -\tfrac12 \]

\[ y = -\tfrac12 (x + 1)(x - 4) \]

Figure (svg): The downward parabola through the x-axis at minus one and at four, with the point three, two marked on it, labelled y equals minus one half times x plus one times x minus four.

The two gold points are where the curve meets the axis, and they are exactly the numbers that go into the brackets. The green point then fixes a.

The negative a matches the picture: the point at three, two sits above the axis and lies between the two crossings, so the curve must arch upward between them and come back down.

Notice how little arithmetic the second bracket needed. Three minus four is negative one, and a single negative factor is what turned the whole coefficient negative, so the sign of the answer was settled by one subtraction.

The vertex is available for free once you have this, because it sits halfway between the crossings, at x equal to one and a half. Its height is negative a half times two and a half times negative two and a half, which is about three point one two five, and that is the highest the pumpkin-shaped arch gets.

Check all three given facts

Why: at x equal to negative 1 the first bracket is zero so y is zero; at x equal to 4 the second bracket is zero so y is zero; and at x equal to 3 the product is negative four, times negative a half, which is 2.

\[ -\tfrac12(3 + 1)(3 - 4) = -\tfrac12(-4) = 2 \]

The two crossings check themselves in this form, which is one of its quiet advantages: you only have to work at verifying the third point.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 309 — Example 2

25. Guided Practice 3: crossings at −2 and 5, through (6, 2)

Worked example

The same three moves with the extra point outside the crossings rather than between them. Predict the sign of a before starting, since that prediction is your first check.

\[ y = a(x + 2)(x - 5) \]

The point at six, two lies to the right of both crossings and above the axis. Outside the crossings a parabola sits on the same side of the axis as its opening direction, so a should come out positive.

Substitute 6 for x and 2 for y

Why: the given point lies on the curve.

\[ 2 = a(6 + 2)(6 - 5) \]

Simplify the coefficient of a

Why: six plus two is eight and six minus five is one, so the product is eight.

\[ 2 = 8a \]

Solve for a

Why: divide both sides by 8 and reduce the fraction.

\[ a = \tfrac28 = \tfrac14 \]

\[ y = \tfrac14 (x + 2)(x - 5) \]

The prediction was right, and a below one also tells you the curve is a wide one. Between the crossings it dips to a low point at x equal to one and a half, where its value is a quarter of three and a half times negative three and a half.

Check all three given facts

Why: the two crossings give zero automatically; at x equal to 6 the brackets give eight times one, and a quarter of eight is two, the height that was given.

\[ \tfrac14(6 + 2)(6 - 5) = \tfrac14(8) = 2 \]

Expanding is worth one more line if a standard form is wanted: a quarter of x squared minus three x minus ten, which is a quarter x squared minus three quarters x minus two and a half.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 3

26. Find the error: the point and the crossings swapped

Error analysis

The book's other error-analysis exercise, and the mistake is not a slip of arithmetic but a mix-up of which numbers play which part. Read the work before reading the notes.

Annotate

On: \( y = a(x - 5)(x + 5) \;\Longrightarrow\; -3 = a(4 - 5)(4 + 5) \;\Longrightarrow\; a = \tfrac13 \)

  • The x-intercepts are 4 and negative 3, and the curve passes through the point (5, negative 5). Those are the only facts available.
  • The brackets have been built from the point (5, negative 5) instead of from the crossings. The number 5 is an x-coordinate of an ordinary point, and negative 5 is its height; neither is an x-intercept.
  • Then the crossing 4 has been substituted as if it were an x-coordinate of an ordinary point, with negative 3 as its height. The two roles have been swapped completely.
  • Crossings first: y equals a times the quantity x minus 4, times the quantity x plus 3. Then substitute the point, negative 5 when x is 5: negative 5 equals a times 1 times 8, so a is negative five eighths.

\[ y = -\tfrac58 (x - 4)(x + 3) \]

The check that would have caught it is the free one. Put x equal to 4 into the wrong answer and you get a third of negative one times nine, which is negative three, not zero, so 4 is not a crossing of that curve at all.

The wider habit is the one from the previous error-analysis slide. Label every number before you use it. Crossings go in brackets; ordinary points go in as an x and a y, and they never trade places.

Notice that the correct value of a is negative, which the picture would also have told you. The given point is below the axis and lies outside both crossings, and a parabola opening upward is above the axis out there.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 26

27. Exercise 19: reading both crossings off a grid

Worked example

A printed graph again, but this time the crossings are the exact readings. When a curve meets the horizontal axis at whole numbers, intercept form is almost always the fastest route.

Figure (svg): An upward parabola crossing the x-axis at minus three and three, passing through the point one, minus four, with its lowest point at zero, minus four and a half.

Both crossings are whole numbers you can read straight off the grid, which is what makes intercept form the cheap route on a picture like this.

Three readings come off this grid without any guessing: the curve crosses at negative three, crosses again at three, and passes through the lattice point at one, negative four.

\[ y = a(x + 3)(x - 3) \]

Substitute 1 for x and negative 4 for y

Why: the lattice point is the only reading that is not on the axis, so it is the one that can fix a.

\[ -4 = a(1 + 3)(1 - 3) \]

Simplify the coefficient of a

Why: one plus three is four and one minus three is negative two, so the product is negative eight.

\[ -4 = -8a \]

Solve for a

Why: divide both sides by negative 8 and reduce.

\[ a = \tfrac{-4}{-8} = \tfrac12 \]

\[ y = \tfrac12 (x + 3)(x - 3) \]

The two crossings are symmetric about the vertical axis here, so the vertex sits at x equal to zero, and its height is a half of three times negative three, which is negative four and a half. The drawing agrees.

Reading a graph is a skill with its own discipline. Take only readings you are certain of, which means points where the curve passes visibly through a grid corner, and ignore places where it passes between two corners however tempting the estimate looks.

It is also worth asking whether vertex form would have been quicker here. The lowest point sits at zero, negative four and a half, which is only half a square down from a grid line, so reading it exactly is a judgement call. The two crossings are unmistakable, and that is what settles the choice.

Check on the symmetric partner

Why: the point at one, negative four should have a mirror image at negative one, negative four; substituting gives a half of two times negative four, which is negative four.

\[ \tfrac12(-1 + 3)(-1 - 3) = \tfrac12(-8) = -4 \]

Expanding gives a half x squared minus four and a half, which has no x term at all, exactly as a curve symmetric about the vertical axis should.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 312 — Exercise 19

28. Exercise 25: crossings at −6 and 3, through (0, −9)

Fill the middle

Here the extra point sits on the vertical axis, so its x is zero. Fill in the product of the brackets and then the value of a, and watch the signs as you go.

Fill in the blanks

-9 = a(0 + 6)(0 - 3) = -18a \;\Longrightarrow\; a = 1/2

Why: Zero plus six is six and zero minus three is negative three, so the product of the brackets is negative eighteen. Then negative nine equals negative eighteen a, and dividing both sides by negative eighteen gives a equal to one half. The function is y equal to one half times the quantity x plus six, times the quantity x minus three.

\[ y = \tfrac12 (x + 6)(x - 3) \]

Both a minus sign and a negative value were in play and the answer still came out positive, because a negative divided by a negative is positive. Doing that division in one step, rather than carrying the signs separately, is what keeps it straight.

The picture confirms it. The given point at zero, negative nine is between the two crossings and below the axis, so the curve dips under the axis there and comes back up, which is a curve opening upward.

The vertex sits halfway between negative six and three, at x equal to negative one and a half, and its height is a half of four and a half times negative four and a half, which is about negative ten and an eighth. That is the deepest the curve goes.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 25

29. When intercept form is simply not available

Counterexample

Every method in this lesson has a situation it cannot handle, and knowing which is part of choosing well. Intercept form fails in a way that is easy to see and easy to forget.

Figure (svg): Two upward parabolas of the same width: one crosses the x-axis at one and three, the other is lifted four units higher and never touches the axis.

The lifted curve has a vertex and countless points, but no x-intercepts, so intercept form has nothing to put in its brackets.

Discussion prompt

The higher curve in the picture has vertex (2, 3) and never meets the horizontal axis. Explain why no function of the form y equals a times the quantity x minus p, times the quantity x minus q, can describe it, and say which form you would use instead.

This is the discriminant of Lesson 4.8 arriving from a new direction. A negative discriminant means no real crossings, and no real crossings means intercept form has nothing to work with.

Vertex form, by contrast, never fails. Every parabola has exactly one vertex, so you can always write one, which is why it is the form to fall back on when a problem gives you a choice.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.8, p. 292 — the discriminant, and what a negative one means for crossings

30. Vertex form or intercept form?

Discrimination

Six descriptions, all of them workable, but each one makes a different form cheap. Decide on the wording alone, then say to yourself which phrase settled it.

Sort into buckets

Sort each description by the form it makes cheapest.

Vertex form
highest point (4.5, 7.25); also through (7, −3); turns around at (−1, −4); also through (2, −1); minimum value 0 reached at x = 5; through (2, −27)
Intercept form
zeros at 3 and 7; also through (6, −9); crosses the axis only at −11 and 3; through (1, −192); roots −5 and −1; also through (−7, −24)
v
The highest point, the point where the curve turns around, and the place where a minimum value is reached are all names for the vertex, so vertex form already holds two of its three numbers.
i
Zeros, roots and places where the curve crosses the axis are all names for the x-intercepts, so intercept form already holds two of its three numbers.

The last description is the sly one. A minimum value of zero reached at x equal to five means the vertex is at five, zero, and that vertex happens to sit on the axis, so the curve has one crossing rather than two.

Intercept form can actually cope with that case, by taking p and q both equal to five, but it is an awkward way to say something vertex form says plainly. When both forms work, take the one that reads more naturally.

Notice how much of this exercise is vocabulary. Zeros, roots, x-intercepts and crossings are one idea; vertex, turning point, highest point, lowest point, maximum and minimum are another. Half of choosing well is knowing those two lists.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 312 — Exercises 8 to 14 and 20 to 25, whose wording this copies

31. From intercept form to standard form, and why you would

Worked example

Answers are often wanted in standard form even when intercept form was the sensible route. Converting is one multiplication, and it is worth doing carefully rather than quickly.

\[ y = -\tfrac12 (x + 1)(x - 4) \]

Multiply the two brackets first

Why: leave the a outside for now; handling one thing at a time is what keeps the signs honest.

\[ (x + 1)(x - 4) = x^2 - 4x + x - 4 = x^2 - 3x - 4 \]

Multiply through by a

Why: distribute negative one half across all three terms, not just the first.

\[ y = -\tfrac12 x^2 + \tfrac32 x + 2 \]

Two terms changed sign because a is negative, and the term that was already negative became positive. Distributing to every term is the move most often done to only the first one.

Standard form now hands you something the other form did not display: the value of c, which is 2, is the height at which the curve crosses the vertical axis.

What it hides, in exchange, is the pair of crossings. Nothing in the expression minus a half x squared plus three halves x plus two announces that the curve meets the axis at negative one and at four; you would have to factor or use the quadratic formula to recover them.

That trade is the honest summary of all three forms. Each one displays some features and hides others, and none is better than the rest in general. The form to write is the one whose visible features are the ones the question is about.

Check against the original at two points

Why: at x equal to 0 the standard form gives 2 and the intercept form gives negative a half times one times negative four, which is also 2; at x equal to 3 the standard form gives negative four and a half plus four and a half plus two, which is 2, matching the given point.

\[ -\tfrac12(9) + \tfrac32(3) + 2 = -4.5 + 4.5 + 2 = 2 \]

Checking a conversion at two different values of x is enough to catch any single arithmetic slip, and it takes two lines.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.2, p. 249 — converting between the forms

32. From three points: a system

Section

Part 4

33. Three points, three equations

Concept

Now the hard case. You are given three points, none of them the vertex and none of them a crossing, so neither of the two cheap forms has anything to hold. Standard form is the only one left.

\[ y = ax^2 + bx + c \]

All three coefficients are unknown, which is the problem and also the clue. Three unknowns need three equations, and each given point supplies exactly one: substitute its x and its y and you get a statement about a, b and c.

\[ (x_1, y_1) \;\longrightarrow\; y_1 = ax_1^2 + bx_1 + c \]

Look closely at that equation. The x and y are now numbers, so the only letters left are a, b and c, and they appear only to the first power. It is a linear equation in three unknowns, exactly the kind Lesson 3.4 dealt with.

So this is not a new method at all. It is Lesson 3.4's machinery pointed at Chapter 4's problem, and the only novelty is that the three unknowns are called a, b and c instead of x, y and z.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Example 3, and the Review note pointing back to systems of equations

34. The system a set of three points builds

Notation

Here is the system the three points of Example 3 produce. Read each row against the point that made it, so that the pattern becomes obvious rather than memorised.

Annotate

On: \( \begin{cases} a - b + c = -3 \\ c = -4 \\ 4a + 2b + c = 6 \end{cases} \)

  • It came from the point (negative 1, negative 3). Squaring negative one gives one, so a is multiplied by 1; the middle term is b times negative one, which is minus b.
  • It came from (0, negative 4). Zero squared is zero and zero times b is zero, so both of those terms vanish and the row says c on its own. A point with x equal to zero is a gift.
  • It came from (2, 6). Two squared is four and two times b is 2b, so the row reads four a plus two b plus c.
  • Use the easy row to remove c from the other two, leaving two equations in a and b, then eliminate again. That is the Lesson 3.4 procedure, unchanged.

Notice how the coefficients in each row are built. The number multiplying a is the point's x squared, the number multiplying b is the point's x, and the number multiplying c is always one.

Once you see that, writing the system takes no thought at all. You are reading the x of each point twice, once squared and once as it stands, and copying the y to the other side.

\[ (x_1, y_1) \;\longrightarrow\; x_1^2 \, a + x_1 \, b + 1 \, c = y_1 \]

Notice which letters are the unknowns and which are the numbers, because that swap is the only conceptual jump in this whole part. In Chapter 4 x has always been the variable and a, b and c the fixed coefficients. Here it is the other way round: the points supply the numbers, and a, b and c are what you are solving for.

That is also why the system is linear even though the function is quadratic. Nothing in these equations is squared except numbers you already know, so no unknown ever appears to a power above one, and Lesson 3.4's elimination applies without any modification at all.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §3.4, p. 178 — solving a linear system in three variables

35. Example 3: a parabola through (−1, −3), (0, −4) and (2, 6)

Worked example

Three ordinary points and nothing else. Write the standard form, turn each point into an equation, and then solve the system exactly as Lesson 3.4 taught.

\[ y = ax^2 + bx + c \]

Substitute the point (−1, −3)

Why: negative one squared is one, so the a term is just a, and b times negative one is minus b.

\[ -3 = a(-1)^2 + b(-1) + c \;\Longrightarrow\; a - b + c = -3 \]

Substitute the point (0, −4)

Why: zero squared is zero and zero times b is zero, so both terms disappear and c is handed to you outright.

\[ -4 = a(0)^2 + b(0) + c \;\Longrightarrow\; c = -4 \]

Substitute the point (2, 6)

Why: two squared is four and two times b is 2b.

\[ 6 = a(2)^2 + b(2) + c \;\Longrightarrow\; 4a + 2b + c = 6 \]

The middle equation is the lucky one, and it is worth always looking for it. A point whose x is zero gives c immediately and turns a three-unknown problem into a two-unknown one for free.

Substitute c equal to −4 into the other two

Why: replacing c everywhere leaves a system of two equations in a and b alone.

\[ a - b - 4 = -3 \;\Longrightarrow\; a - b = 1 \]

Do the same to the third equation

Why: four a plus two b, minus four, equals six, so four a plus two b is ten.

\[ 4a + 2b - 4 = 6 \;\Longrightarrow\; 4a + 2b = 10 \]

Eliminate b

Why: double the first equation so the b terms cancel when the two are added.

\[ 2a - 2b = 2 \;\; + \;\; 4a + 2b = 10 \;\Longrightarrow\; 6a = 12 \]

Solve for a, then back-substitute for b

Why: six a equals twelve gives a equal to two; then two minus b equals one gives b equal to one.

\[ a = 2, \quad b = 1, \quad c = -4 \]

\[ y = 2x^2 + x - 4 \]

Figure (svg): The parabola y equals two x squared plus x minus four drawn through the three given points at minus one, minus three, at zero, minus four and at two, six.

Three points, three equations, one curve. Notice that none of the three is the vertex or a crossing, which is exactly why a system was needed.

The drawn curve passes through all three marked points, which is the only visual evidence that the system was solved correctly. One wrong sign and one of the three dots would sit off the curve.

Look back at how the three unknowns fell. The point on the vertical axis gave c outright, substituting it reduced a three-unknown system to a two-unknown one, and a single elimination then finished it. Three lines of real work for three unknowns is about as cheap as this route gets.

Compare that with Example 1, where a vertex and one point were settled in a single division. The extra cost here is the price of being told nothing special about the three points you were given, which is the whole argument for choosing the form carefully.

Check all three given points

Why: at x equal to negative 1 the value is two minus one minus four, which is negative 3; at x equal to 0 it is negative 4; at x equal to 2 it is eight plus two minus four, which is 6.

\[ 2(-1)^2 + (-1) - 4 = -3, \quad 2(2)^2 + 2 - 4 = 6 \]

All three have to be tested, not just the last one used, because an error made early in the elimination can still leave the final equation satisfied.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Example 3

36. Put the three-point method in order

Ranking

The steps are not interchangeable, and doing them out of order is what makes this method feel long. Arrange them as you would actually carry them out.

Put in order

Order the steps of writing a quadratic through three given points.

  1. Write y = ax² + bx + c, with all three coefficients unknown
  2. Substitute each point's x and y to get three linear equations
  3. Use the simplest equation to remove one unknown from the others
  4. Eliminate a second unknown and solve the pair that remains
  5. Back-substitute to recover the remaining coefficients
  6. Test the finished function on all three original points

Why: Standard form comes first because the other two forms have nothing to hold when no vertex and no crossings are given. Each point then becomes one linear equation. From there it is Lesson 3.4: use the simplest equation, often the one from a point with x equal to zero, to remove one unknown, eliminate a second, solve the remaining pair and back-substitute. The last step is testing on all three points, because an error in the elimination can still leave the final equation satisfied.

The step people leave out is the last one. Solving a system of three equations takes several lines, and a slip anywhere in them produces a function that fits no points at all, which the test catches in three substitutions.

The step people rush is the third. Choosing which equation to work from is worth a few seconds: an equation with a missing term, or one whose numbers are small, saves arithmetic all the way down.

Notice also that there is no picture in this list. You can carry the whole method out without ever drawing the parabola, which is exactly what makes it work when no graph is available.

37. Why a point on the vertical axis is a gift

Socratic

Example 3 got c in a single line because one of its points had x equal to zero. That is worth understanding rather than just enjoying, because you can often arrange it.

Discussion prompt

Explain, using the standard form, why a point whose x-coordinate is zero gives you c straight away. Then say what it means graphically, and why it makes the rest of the work shorter.

This is why modelling problems so often measure time from the start of the experiment rather than from a calendar date. Choosing the origin sensibly is a decision that makes the algebra easier before any algebra happens.

It is also why the colour television data later in this lesson lists years since 1970 rather than the years themselves. Had it listed 1970 through 2000, every x would be a four-digit number and every coefficient a horror.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Example 3, Step 1, Equation 2

38. Guided Practice 4: through (−1, 5), (0, −1) and (2, 11)

Worked example

The same shape of problem, including the useful point with x equal to zero. Work it through without looking back at Example 3 if you can.

\[ y = ax^2 + bx + c \]

Substitute each point in turn

Why: the three points become three linear equations in a, b and c.

\[ a - b + c = 5, \quad c = -1, \quad 4a + 2b + c = 11 \]

Use the middle equation to remove c

Why: put negative one in place of c in the other two and tidy each.

\[ a - b = 6, \quad 4a + 2b = 12 \]

The second of those simplifies further: every term is even, so halving it gives two a plus b equals six, and small numbers are always worth taking.

Eliminate b

Why: halve the second equation, then add it to the first so the b terms cancel.

\[ 2a + b = 6 \;\; + \;\; a - b = 6 \;\Longrightarrow\; 3a = 12 \]

Solve for a, then back-substitute

Why: three a equals twelve gives a equal to four; then four minus b equals six gives b equal to negative two.

\[ a = 4, \quad b = -2, \quad c = -1 \]

\[ y = 4x^2 - 2x - 1 \]

Before checking, look at whether the answer is plausible. A positive a means the curve opens upward, and the three given heights of five, negative one and eleven do dip and rise again, which fits.

It is worth noticing which equation did the real work. The one from the point with x equal to zero cost nothing to write and removed an entire unknown, which is why it is always the first one to look for in a list of points.

If you would rather avoid halving, you can eliminate a instead: multiply the first equation by four and subtract it from the second to get six b equal to negative twelve, so b is negative two and then a is four. Either order works, and choosing the one with smaller numbers is a habit worth having.

Check all three given points

Why: at x equal to negative 1 the value is four plus two minus one, which is 5; at x equal to 0 it is negative 1; at x equal to 2 it is sixteen minus four minus one, which is 11.

\[ 4(-1)^2 - 2(-1) - 1 = 5, \quad 4(2)^2 - 2(2) - 1 = 11 \]

Two of those checks involve a negative substituted into a squared term, which is the step worth slowing down for: negative one squared is positive one.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 4

39. Guided Practice 6: through (−1, 0), (1, −2) and (2, −15)

Worked example

This time no point has x equal to zero, so c does not fall out for free. The elimination takes one more step, and choosing which unknown to remove first is the decision that saves work.

\[ y = ax^2 + bx + c \]

Substitute each point in turn

Why: three points, three linear equations, exactly as before.

\[ a - b + c = 0, \quad a + b + c = -2, \quad 4a + 2b + c = -15 \]

Look at the first two equations. They differ only in the sign of b, so subtracting one from the other kills both a and c at once and hands you b in a single line.

Subtract the first equation from the second

Why: a and c cancel completely, leaving two b equal to negative two.

\[ 2b = -2 \;\Longrightarrow\; b = -1 \]

Put b back into the second equation

Why: a minus one plus c equals negative two, so a plus c is negative one.

\[ a + c = -1 \]

Put b into the third equation

Why: four a minus two plus c equals negative fifteen, so four a plus c is negative thirteen.

\[ 4a + c = -13 \]

Eliminate c and solve

Why: subtract the smaller equation from the larger; three a equals negative twelve, so a is negative four and then c is 3.

\[ 3a = -12 \;\Longrightarrow\; a = -4, \quad c = 3 \]

\[ y = -4x^2 - x + 3 \]

A negative a means this parabola opens downward, which the given heights support: they fall steeply from zero to negative two to negative fifteen as x increases.

Check all three given points

Why: at x equal to negative 1 the value is negative four plus one plus three, which is 0; at x equal to 1 it is negative four minus one plus three, which is negative 2; at x equal to 2 it is negative sixteen minus two plus three, which is negative 15.

\[ -4(2)^2 - 2 + 3 = -16 - 2 + 3 = -15 \]

The point at negative one, zero means the curve crosses the axis there, so this parabola could also be written in intercept form once its second crossing is found; the system found it without needing to know that.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 6

40. Guided Practice 5: when the answer is not a whole number

Fill the middle

Three points again, but the coefficients come out as fractions. That is normal and not a sign of an error; fill in the two missing values and see for yourself.

\[ \text{through } (-2, -1), \; (0, 3), \; (4, 1) \]

Fill in the blanks

c = 3, \quad 2a - b = -2, \quad 8a + 2b = -1 \;\Longrightarrow\; a = -\tfrac12___}

Why: The point with x equal to zero gives c equal to 3 at once. Substituting that into the other two and dividing each by 2 gives two a minus b equal to negative two and eight a plus two b equal to negative one. Rearranging the first to b equal to two a plus two and substituting gives twelve a equal to negative five, so a is negative five twelfths and b is seven sixths.

\[ y = -\tfrac{5}{12}x^2 + \tfrac76 x + 3 \]

Fractional coefficients feel wrong the first few times, but there is nothing special about whole numbers. A parabola through three arbitrary points has no reason to have tidy coefficients, and the textbook examples are tidy only because they were chosen to be.

Checking is where fractions earn their keep. At x equal to four the first term is negative five twelfths of sixteen, which is negative twenty thirds; the second is seven sixths of four, which is fourteen thirds; and negative twenty thirds plus fourteen thirds is negative two, which with the 3 gives 1, the given height.

If you would rather avoid the fractions, multiply the whole equation by twelve and work with twelve y equal to negative five x squared plus fourteen x plus thirty-six. The curve is the same; only the bookkeeping changes.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 310 — Guided Practice 5

41. Exercise 30: three points read off a printed graph

Worked example

A graph with no vertex and no crossings marked, and three lattice points the curve visibly passes through. Read them exactly, then it is the system again.

Figure (svg): The parabola y equals a half x squared plus three x minus two drawn through the points minus four, minus six, zero, minus two and two, six.

Reading three lattice points off a grid is the whole of the work; after that it is Lesson 3.4's elimination, with a, b and c in place of x, y and z.

The three readings are negative four, negative six; zero, negative two; and two, six. The middle one is on the vertical axis, so it will hand over c immediately.

\[ y = ax^2 + bx + c \]

Substitute each point in turn

Why: the point with x equal to zero gives c straight away, and the other two give equations in a and b.

\[ 16a - 4b + c = -6, \quad c = -2, \quad 4a + 2b + c = 6 \]

Remove c and simplify

Why: put negative two in place of c, then divide each equation by 4 and by 2 to shrink the numbers.

\[ 4a - b = -1, \quad 2a + b = 4 \]

Eliminate b and solve

Why: add the two equations so the b terms cancel; six a equals three, so a is a half, and then b is 3.

\[ 6a = 3 \;\Longrightarrow\; a = \tfrac12, \quad b = 3 \]

\[ y = \tfrac12 x^2 + 3x - 2 \]

Dividing each equation by its common factor before eliminating is a small habit with a large payoff. It turned sixteen a minus four b into four a minus b, and the whole elimination then fitted on one line.

Choosing which three points to read matters too. Take readings that are far apart and land exactly on grid corners, because a point read a little wrong distorts the whole curve, and points bunched together make the system badly behaved.

It is worth asking why this graph could not have been done more cheaply. The curve does cross the axis, but not at whole numbers, and its lowest point is at negative three with height negative six and a half, which is also not a lattice point. Nothing exact is readable except three ordinary points, so the system is the only honest route.

Check all three readings

Why: at x equal to negative 4 the value is eight minus twelve minus two, which is negative 6; at x equal to 0 it is negative 2; at x equal to 2 it is two plus six minus two, which is 6.

\[ \tfrac12(-4)^2 + 3(-4) - 2 = 8 - 12 - 2 = -6 \]

The vertex, if you want it, sits at x equal to negative three with height negative six and a half, which matches the lowest point of the printed curve.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 30

42. Three points that refuse to be a parabola

Counterexample

The method never breaks down, but it can return an answer you did not expect. This is the book's challenge exercise, and the surprise in it is worth more than several routine ones.

\[ \text{through } (1, -4), \; (-3, -16), \; (7, 14) \]

Discussion prompt

Set up and solve the system for these three points. Say what value of a you get, and explain what that tells you about the three points.

Figure (svg): Three points at one, minus four, at minus three, minus sixteen and at seven, fourteen, all lying exactly on the straight line y equals three x minus seven.

The system is perfectly happy; it simply reports that the leading coefficient is zero, which is the honest answer for three points in a row.

The picture confirms it: the three points sit exactly on a line of slope three. Three points determine at most one parabola, and when they happen to be collinear that parabola is flat, which is to say it is a line.

So the honest statement of the rule is this. Three points with different x-coordinates determine exactly one function of the form a x squared plus b x plus c, and that function is a genuine parabola precisely when a comes out non-zero.

If you ever get a equal to zero on a question that promised a parabola, do not hunt for an arithmetic slip first. Check whether the points are in a straight line, because the algebra may be right and the question wrong.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 45

43. What if two points share an x-coordinate?

Edge cases

The rule just stated came with a condition attached: the three x-coordinates must be different. Here is why, and it is not a technicality.

Figure (svg): Two points at x equal to two, one at height three and one at height minus one, with a vertical dashed line through them, and a parabola that can only reach one of them.

A function gives one output per input, so the dashed line may meet the curve once at most. Two points on it is a contradiction, not a hard problem.

Discussion prompt

Try to set up the system for a curve through (2, 3) and (2, -1). Write the two equations those points give and say what goes wrong.

This is the vertical line test from Chapter 2 arriving as an inconsistent system. The same fact shows up as a picture and as a contradiction, and the two ways of seeing it support each other.

It also tells you something practical about data. If a table of measurements lists two different results for the same input, no function of any kind will fit both, and you have a measurement problem rather than a modelling problem.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §2.1, p. 72 — the vertical line test for a function

44. The antenna dish, from three surveyed points

Real world

This is what the three-point method is actually for. Nobody hands an engineer a vertex; a dish gets measured at whatever places a tape measure reaches, and the curve has to be reconstructed from those.

Figure (svg): The cross section of an antenna dish drawn as an upward curve from zero, four down to its lowest point at four, three and back up, with the three measured points marked.

Three surveyed points on the dish are enough to pin the whole cross section, and the model then hands you the lowest point for free.

Discussion prompt

Three points on the cross section of an antenna dish are (0, 4), (2, 3.25) and (5, 3.0625), with all measurements in feet. Write a quadratic model for the cross section and find the lowest point of the dish.

Notice how the vertex was recovered even though no vertex was measured. Once the three coefficients are known, every other feature of the parabola follows, which is why standard form is worth the extra work when it is the only route available.

The check is the same as always and matters more here, because these numbers are measurements. At x equal to 5 the model gives twenty-five sixteenths minus two and a half, plus four, which is 3.0625, exactly the surveyed value.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 314 — Exercise 46

45. The three routes, side by side

Comparison

Fill in the blanks from what the three parts have shown. Reading across a row tells you everything about one route; reading down a column compares them.

Comparison matrix

routewhat you must be givenunknowns left after substitutingequations to solve
vertex formthe vertex, plus one more point1one
intercept formboth x-intercepts, plus one more point1, namely aone
standard formthree points3, namely a, b and cthree, as a system

The table makes the cost of each route visible. Two of them cost one equation and one of them costs three, and which one you are in was decided entirely by what the question told you.

There is a fourth row the table does not need. If you are given the vertex and both crossings, any of the three works, and you should take vertex form or intercept form because a single equation beats a system every time.

There is also a column the table deliberately leaves out, because it would read the same in every row. Whatever route you took, the last thing you do is test the finished function on every fact you were given, and that step never changes.

46. Best-fitting quadratic models

Section

Part 5

47. Data does not sit on a curve

Concept

Everything so far has been exact. Three points were given, and the parabola went through all three of them precisely. Real measurements never behave like that.

Figure (svg): A scatter plot of six points: launch angles twenty to seventy degrees along the bottom and distances between three hundred and twenty and five hundred and ten feet up the side. The points rise, level off and fall.

Plot the data before fitting anything. These points rise, flatten and fall, which is a parabola's shape and not a line's.

These are six measured distances for a pumpkin launched from a catapult at six different angles. They are measurements, so each carries a little error, and no parabola will pass through all six exactly.

But look at the shape. The distances rise, level off around forty or fifty degrees and then fall away, which is the outline of a parabola and not of a line. A line would have to keep rising or keep falling.

best-fitting quadratic model — The quadratic function produced by quadratic regression on a data set: the one that comes closest to all the data points at once, rather than passing through any particular three of them.

Quadratic regression is the machine that finds it, and a graphing calculator does the arithmetic. What it hands back is a, b and c, and what you must supply is the judgement about whether a quadratic was the right thing to ask for.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 311 — the definition of a best-fitting quadratic model

48. Example 4: the best-fitting model for the pumpkin data

Worked example

Pumpkin tossing is a real contest in Morton, Illinois, where catapults compete on distance. This catapult launches from twenty-five feet up at one hundred and twenty-five feet per second, and the table records how far the pumpkin travels at six launch angles.

angle in degrees203040506070
distance in feet372462509501437323

Enter the two lists and make a scatter plot

Why: angles in the first list and distances in the second; plotting first is what tells you a quadratic is worth fitting at all.

\[ x: 20, 30, 40, 50, 60, 70 \qquad y: 372, 462, 509, 501, 437, 323 \]

The plot rises, flattens and falls, which no straight line does, so a quadratic is the right family to ask for. Skipping this step is how people end up fitting curves to data that has no curve in it.

Run quadratic regression

Why: the calculator returns the three coefficients of the best-fitting quadratic, which the book rounds to three significant figures.

\[ y = -0.261x^2 + 22.6x + 23.0 \]

Figure (svg): The same six pumpkin measurements with the best-fitting parabola drawn through them, its highest point marked near forty-three degrees, and the curve continuing past seventy degrees to cross below zero.

The solid stretch is where measurements exist; the dashed stretch is invention. The model predicts a negative distance at ninety degrees.

The curve is drawn through the data, not through any three chosen points. Every measurement sits close to it and none sits on it, which is exactly what a fit to real data should look like.

Check the model against the measurements

Why: at 20 degrees the model gives about 370.3 feet against a measured 372; at 40 degrees about 508.4 against 509; at 70 degrees about 323.4 against 323. The largest miss anywhere is about three and a half feet in a distance of over four hundred.

\[ -0.261(40)^2 + 22.6(40) + 23.0 \approx 509 \]

Under one percent error at every point, with some misses above and some below, is the signature of a good fit. A model that missed every point on the same side would be the wrong shape, however small the misses.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 311 — Example 4

49. How far the model misses, point by point

Picture it

The fit is easier to judge when you look at the misses alone rather than at the curve. Each bar here is one measurement minus what the model predicts for it.

Figure (svg): Six vertical bars measuring how far the pumpkin model misses each measurement, drawn from a zero line: two feet too low at twenty degrees, three and a half feet too high at thirty, and under half a foot at sixty and seventy.

Some measurements sit above the model and some below, and the largest miss is three and a half feet in a throw of 462. That mixture, with no drift, is what a good fit looks like.

Two things are worth noticing. The misses are tiny: the largest is about three and a half feet against distances in the hundreds, which is well under one percent.

And the misses change sign. Some measurements are above the curve and some below, with no run of same-signed misses at one end, which says the curve is bending the way the data bends.

Had every miss at the ends been negative and every miss in the middle positive, that pattern would have told you the curve was too flat, and a different model was needed. Reading the misses is how you tell a fit that is good from one that is merely close.

50. Guided Practice 7: the angle that throws a pumpkin farthest

Worked example

The model is not the answer to anything by itself; it is a tool for answering questions. The competitor's question is which angle to use, and that is asking for the vertex.

\[ y = -0.261x^2 + 22.6x + 23.0 \]

Identify a and b in the model

Why: a is negative, so the parabola opens downward and its vertex is the highest point, which is what farthest means here.

\[ a = -0.261, \quad b = 22.6 \]

This is Lesson 4.1's formula being used on a model rather than on a made-up function, which is the whole point of building models: old tools keep working on them.

Use the vertex formula

Why: the x-coordinate of the vertex of a parabola in standard form is negative b over two a.

\[ x = -\dfrac{b}{2a} = -\dfrac{22.6}{2(-0.261)} \approx 43.3 \]

With the calculator's unrounded coefficients the same formula gives about 43.2 degrees, so the rounding in the printed model costs you a tenth of a degree and nothing more. That is worth knowing in general: rounding a model before using it moves the answer a little, and you should say how much.

The answer is also worth sanity-checking against physics you may already know. Ignoring air resistance and launching from ground level, the best angle is forty-five degrees; this catapult launches from twenty-five feet up, which shifts the best angle slightly below forty-five, and the model says forty-three.

Find the distance at that angle

Why: substitute the angle back into the model to see how far the pumpkin goes.

\[ y \approx -0.261(43.2)^2 + 22.6(43.2) + 23.0 \approx 511 \]

Check against the measurements

Why: the two measured angles nearest 43 degrees are 40 and 50 degrees, which gave 509 and 501 feet, and the model's maximum of about 511 feet sits just above both, exactly where a peak between them should be.

\[ 509 < 511 > 501 \]

A vertex that fell outside the range of the data, or below the highest measurement, would be a sign that something had gone wrong. This one passes both tests.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 311 — Guided Practice 7

51. How far would it go at ninety degrees?

Prediction

A model is a summary of the data it was built from. Ask it a question far outside that data and it will still answer, confidently and wrongly. Predict what this one says before you compute it.

\[ y = -0.261x^2 + 22.6x + 23.0, \qquad x = 90 \]

Predict first

The pumpkin data covers launch angles from 20 to 70 degrees. What does the model predict for a launch angle of 90 degrees?

  • About 511 feet, the same as at the best angle
  • About 61 feet, a short but sensible throw
  • About negative 61 feet, which is impossible
  • Zero feet, since the pumpkin goes straight up

Correct: About negative 61 feet, which is impossible

Why: Substituting 90 gives negative 0.261 times 8100, plus 22.6 times 90, plus 23.0, which is about negative 61. A distance cannot be negative, so the model has left the range where it means anything. Physically a pumpkin launched straight up lands where it started, so the true answer is near zero, and the model is out by about sixty feet.

Notice that the honest answer, near zero, is not the model's answer. The parabola has to keep falling once past its vertex, because that is what parabolas do, while the real distance flattens towards zero instead.

So the model is an excellent description between twenty and seventy degrees and a bad one outside. That is not a flaw in the fit; it is the nature of fitting a simple curve to a limited stretch of data.

The discipline this demands is easy to state and easy to forget: say, out loud, over what range your model was fitted, and refuse to use it far outside that range. Every answer you report should come with that range attached.

52. Exercise 49: wind and a two hundred metre time

Worked example

A second set of real measurements, of a different kind. A tailwind makes a two hundred metre runner faster and a headwind makes them slower, and the table records the change in finishing time against wind speed.

wind speed, metres per second−6−4−20246
change in time, seconds2.281.420.670−0.57−1.05−1.42

Positive wind speeds are tailwinds and negative ones are headwinds; the change is measured against the runner's own no-wind time, which is why the middle entry is exactly zero.

Plot before fitting

Why: a headwind of six costs 2.28 seconds while a tailwind of six saves only 1.42, so the effect is not symmetric and a straight line cannot describe it.

\[ 2.28 \ne 1.42 \]

Figure (svg): Seven measurements of how wind speed changes a two hundred metre time, from a six metre per second headwind to a six metre per second tailwind, with a gently curved best-fitting model drawn through them.

The points are not on a straight line: the gain from a tailwind shrinks as the wind strengthens, and only a curve can say that.

Run quadratic regression on the seven pairs

Why: the calculator returns the best-fitting quadratic for this data.

\[ t \approx 0.0119s^2 - 0.3086s - 0.0005 \]

The small positive leading coefficient is what makes the curve bend gently upward, and that bend is precisely the asymmetry the plot showed: headwinds hurt more than tailwinds help.

The constant term is almost exactly zero, which is a good sign rather than a coincidence. The measurement at no wind was defined to be zero change, so any decent model has to pass close to the origin, and this one misses it by half a thousandth of a second.

The middle coefficient is the one an athlete would care about. Near zero wind the curve is nearly straight with a slope of about negative three tenths, so the first metre per second of tailwind is worth roughly three tenths of a second, and the next one a little less. That shrinking return is exactly what the squared term contributes.

Predict the change at ten metres per second

Why: substitute ten for s in the model, which is beyond the measured range, so treat the answer as an estimate rather than a fact.

\[ t \approx 0.0119(100) - 0.3086(10) - 0.0005 \approx -1.90 \]

Check the model on the measured values

Why: at a wind speed of negative 6 it gives 2.28 seconds against a measured 2.28; at negative 2 it gives 0.66 against 0.67; at 4 it gives negative 1.04 against negative 1.05. Every prediction is within one hundredth of a second.

\[ 0.0119(-6)^2 - 0.3086(-6) - 0.0005 \approx 2.28 \]

The prediction at ten metres per second is an extrapolation of four more units past the data, so report it as about one point nine seconds faster and say that it assumes the trend continues.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 314 — Exercise 49

53. Which claims about a best-fitting model survive?

Elimination

Four statements about the pumpkin model. Three of them are things students genuinely believe after meeting regression for the first time. Strike those out and keep the one that holds.

Eliminate the wrong options

Only one of these claims about the best-fitting quadratic model for the pumpkin data is true.

  • a. The curve passes exactly through all six measured points
  • b. It comes closer to all six points at once than any other quadratic
  • c. It can be trusted at any launch angle at all
  • d. Since it fits well, the data must really be quadratic

Survives elimination: b

Why: Regression does not interpolate; it balances. The surviving claim is the definition of a best-fitting model: among all quadratics, it is the one that comes closest to the whole data set at once. The claim about passing through every point confuses regression with the three-point method of Part 4, the claim about trusting it everywhere ignores what happens past seventy degrees, and the claim that the data must be quadratic mistakes a good description for a proof.

The distinction in that last one is worth dwelling on. A model that fits is a claim about the data you have, never a claim about how the world works, and the same data can often be fitted almost as well by several different families of curve.

The difference from Part 4 is the sharpest point here. There you had three points and demanded that the curve hit all of them; here you have six and ask only that it come close to all of them, and those are genuinely different requests.

With exactly three data points the two ideas coincide, because a quadratic can hit all three, and the best fit is the exact fit. From the fourth point onward they part company for good.

54. Colour televisions, and a model built by hand

Real world

Regression needs a calculator, but you can build a decent model by hand from a scatter plot using the three-point method, which is what this exercise asks for.

years since 1970051015202530
households with colour televisions, millions214763789094101

Notice the choice of variable. Counting years since 1970 rather than using the years themselves keeps the numbers small and, better, puts a data point at x equal to zero, which hands you c for nothing.

Discussion prompt

Pick the first, middle and last readings, write the quadratic through those three points, and then compare its values with the whole table. Say where it fits well and where it does not.

The comparison is the lesson. Three well-spread points give a model that is surprisingly good, and regression improves on it by refusing to privilege any three readings over the rest.

One more piece of honesty is needed before using either model. Both turn downward and predict a fall in the number of households after about the year 2002, which is the parabola's shape talking rather than the data, since no reading past year 30 was ever taken.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 314 — Exercise 50

55. Putting it together

Section

Part 6

56. One decision, then one of three routes

Pattern

The whole lesson is one decision followed by work you already knew how to do. Make the decision explicitly, in writing, before starting, and the work that follows is almost mechanical.

Figure (svg): A decision diagram: the question of what you were given branches to three routes, each naming the form to use, what the unknown is and how many equations you must solve.

Read down a column and you have the whole method for that case. The only genuine decision is the one at the top.
  1. Read what you were given and name it: a vertex, two crossings, or ordinary points.
  2. A vertex and one point: write vertex form with h and k in place, substitute the point, solve for a.
  3. Both crossings and one point: write intercept form with p and q in place, substitute the point, solve for a.
  4. Three ordinary points: write standard form, turn each point into a linear equation, and solve the system as in Lesson 3.4.
  5. A table of measurements: plot it first, then run quadratic regression, and state the range over which the model may be used.
  6. Always finish by testing the function on every fact the question gave you, not just the one you used last.

The last line is the one that gets skipped and the only one that catches errors. Testing all the given facts takes three substitutions and rescues an entire question.

Notice also what is common to the first three routes. In each one you substitute what you know into the form that already holds it, and then use a point to settle what is left; only the amount left over changes.

The diagram is worth reading upward as well as downward. If you find yourself setting up a system of three equations, look back up the column and ask whether the question really gave you three ordinary points, or whether it named a vertex or a pair of crossings that you read past.

Nothing in this procedure is special to Chapter 4. The same counting argument decides how many points a cubic needs, or a line, and Chapter 5 uses it again. Count the unknowns in the form, count the facts you were given, and the difference is the number of extra points to ask for.

57. Match the given information to the first line you would write

Matching

Four questions, four opening lines. No solving is needed: decide from the wording alone, then say which phrase in each description settled it.

Match the pairs

  • a. vertex (3, 5); passes through (7, −3)
  • b. x-intercepts −11 and 3; passes through (1, −192)
  • c. through (−2, −13), (2, 3) and (4, 5)
  • d. a table of six measured pairs
  • w. y = a(x − 3)² + 5, then substitute the point
  • x. y = a(x + 11)(x − 3), then substitute the point
  • y. y = ax² + bx + c, then build three equations
  • z. a scatter plot, then quadratic regression

Why: A named vertex goes straight into vertex form, with the signs written as x minus h. Named x-intercepts go straight into intercept form, so a crossing at negative eleven appears as x plus eleven. Three ordinary points name neither, so standard form and a system are the only route. A table of measurements is different in kind: no curve passes through all of it, so you plot it and fit a best model instead of solving anything exactly.

The first two openings are one substitution away from being finished, and the third is a page of elimination. That difference in cost is the reason the decision comes first.

The fourth is different in kind rather than in cost. The first three ask for the curve through some points; the last asks for the curve near many points, and no amount of algebra will make it pass through them all.

If you can write these four opening lines quickly, you have the lesson. Everything after the opening line is Chapter 4 and Lesson 3.4 arithmetic that you have already done many times.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercises 37, 41 and 42

58. One parabola, written three ways

Worked example

To finish the algebra, here is a single curve described in three different ways, so that you can see the forms are three spellings of one function rather than three different objects.

\[ \text{vertex } (1, 4), \quad \text{crossings at } -1 \text{ and } 3, \quad \text{through } (-2, -5) \]

Route one: use the vertex and the point

Why: substitute the vertex, then the point, and solve for a.

\[ -5 = a(-2 - 1)^2 + 4 \;\Longrightarrow\; -9 = 9a \;\Longrightarrow\; a = -1 \]

\[ y = -(x - 1)^2 + 4 \]

Route two: use the crossings and the point

Why: substitute the crossings, then the same point, and solve for a again.

\[ -5 = a(-2 + 1)(-2 - 3) \;\Longrightarrow\; -5 = 5a \;\Longrightarrow\; a = -1 \]

\[ y = -(x + 1)(x - 3) \]

Both routes produced the same value of a, which is not luck. There is only one parabola with those features, so any correct route must arrive at it.

The two routes even used the same amount of work: one substitution, one simplification and one division each. Given both a vertex and a pair of crossings, you may take whichever you find easier to read, and the exercise the book sets here asks for all three forms precisely so that you notice that.

Check the two descriptions against each other before expanding. The crossings at negative one and three have midpoint one, which is the x of the vertex you were given, so the two facts agree; had they disagreed, no parabola would have both.

Expand both to standard form

Why: expanding the vertex form gives minus the quantity x squared minus two x plus one, plus four; expanding the intercept form gives minus the quantity x squared minus two x minus three.

\[ y = -x^2 + 2x + 3 \]

Three expressions, one function. Vertex form displays the turning point, intercept form displays the crossings and standard form displays the height at which the curve meets the vertical axis, which here is 3.

Check all four given facts

Why: at x equal to 1 the value is negative one plus two plus three, which is 4, the vertex height; at x equal to negative 1 and at x equal to 3 the value is zero, the two crossings; and at x equal to negative 2 it is negative four minus four plus three, which is negative 5.

\[ -(-2)^2 + 2(-2) + 3 = -4 - 4 + 3 = -5 \]

Testing every fact rather than the one you used is what turns three separate pieces of work into one confirmed answer.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 43

59. Check: a vertex and one point

Check

Decide the form from the wording first, then substitute. Watch the sign inside the bracket and watch the sign of a, which the relative heights of the two points will tell you.

Check your understanding

A parabola has vertex (4, −5) and passes through (2, −1). Which function is it?

  • A. y = (x − 4)² − 5 (correct)
  • B. y = −(x − 4)² − 5
  • C. y = (x + 4)² − 5
  • D. y = (x − 4)² + 5

Answer: A

Why: Vertex form with h equal to 4 and k equal to negative 5 gives y equal to a times the quantity x minus 4, squared, minus 5. Substituting the point gives negative 1 equal to 4a minus 5, so a is 1. The given point sits above the vertex, so a had to come out positive.

Why B tempts people
A negative leading coefficient opens the curve downward, which would put the second point below the vertex rather than above it.
Why C tempts people
A plus sign inside the bracket means a vertex at x equal to negative 4, not at 4; substituting x equal to 4 into it gives 59, not the vertex height.
Why D tempts people
This has the vertex at height positive 5. Substituting x equal to 4 must return negative 5, and this returns positive 5.

Every wrong option here is a real slip rather than a random wrong answer: one flips the direction, one flips the sign of h, and one flips the sign of k. None of them is a mistake in arithmetic; all three are mistakes in reading.

The single test that catches all three is the one from Part 2: substitute the vertex's x into your answer and see whether the vertex's y comes back. Only one of the four survives it.

The plausibility check is faster still. The given point is at height negative one and the vertex is at negative five, so the curve rises away from the vertex, which no negative value of a can do.

Finish by testing the point you were actually given as well. At x equal to two the winning function returns four minus five, which is negative one, so both of the facts in the question are reproduced and the answer is confirmed rather than merely chosen.

60. Check: two crossings and one point

Check

Now the intercept-form version. Get the signs in the brackets right first, because everything after that is one division.

Check your understanding

The x-intercepts of a parabola are −2 and 5, and it passes through (6, 2). Which function is it?

  • A. y = (1/4)(x + 2)(x − 5) (correct)
  • B. y = (1/4)(x − 2)(x + 5)
  • C. y = 4(x + 2)(x − 5)
  • D. y = 2(x + 2)(x − 5)

Answer: A

Why: A crossing at negative 2 gives the bracket x plus 2, and a crossing at 5 gives x minus 5. Substituting the point gives 2 equal to a times 8 times 1, so a is one quarter. Testing the answer at x equal to 6 returns 2, and at x equal to negative 2 and at 5 it returns zero.

Why B tempts people
The signs in the brackets are reversed, so this crosses at 2 and at negative 5, which is not what was given.
Why C tempts people
This is the reciprocal of the right coefficient: 2 equals 8a gives a equal to one eighth of 2, not four times it. At x equal to 6 it returns 32.
Why D tempts people
This uses only part of the arithmetic: at x equal to 6 the brackets give 8, and twice 8 is 16, not the height of 2 that was given.

The crossings check themselves in this form, so the only substitution worth doing is the third point. That is one of the quiet advantages of intercept form.

The bracket signs are where nearly all the marks go. A bracket reading x plus 2 crosses at negative 2, and saying that out loud each time you write one is enough to fix it.

Two of the wrong options here share the correct brackets and differ only in the coefficient in front, which is the reminder that getting the brackets right is necessary but not sufficient. The point has to be substituted and the resulting equation solved, not guessed at.

If you are ever unsure between two candidate coefficients, put the given point into each. Only one will return the given height, and the arithmetic takes ten seconds. That habit is faster than re-deriving the answer and it catches a sign slip as well as a size slip.

It is also worth predicting the sign of a before you compute it, as Guided Practice 3 did. The given point is above the axis and lies to the right of both crossings, so the curve must open upward and a has to be positive. Every option offered here is positive, so that prediction does not choose between them; it only tells you that none of them can be dismissed on sight, and the arithmetic has to be done.

61. Check: three ordinary points

Check

The system route. Set it up, remember that a point with x equal to zero would have given c free, and notice that none of these three does.

Check your understanding

Write the quadratic in standard form through (−1, 9), (1, 1) and (3, 17).

  • A. y = 3x² − 4x + 2 (correct)
  • B. y = 3x² + 4x + 2
  • C. y = 2x² − 4x + 3
  • D. y = 3x² − 4x − 2

Answer: A

Why: The three points give a minus b plus c equal to 9, a plus b plus c equal to 1 and 9a plus 3b plus c equal to 17. Subtracting the first two gives 2b equal to negative 8, so b is negative 4; the rest of the elimination gives a equal to 3 and c equal to 2. Testing: at x equal to negative 1 the value is 3 plus 4 plus 2, which is 9.

Why B tempts people
The sign of the middle coefficient is wrong. At x equal to negative 1 this gives 3 minus 4 plus 2, which is 1, not the 9 that was given.
Why C tempts people
The first and last coefficients have been swapped. At x equal to 3 this gives 18 minus 12 plus 3, which is 9, not 17.
Why D tempts people
The constant has the wrong sign. At x equal to 1 this gives 3 minus 4 minus 2, which is negative 3, not 1.

Every wrong option fails on at least one of the three points, which is exactly why the method ends with testing all three rather than one.

The fastest route through this particular system is the one Guided Practice 6 used: the equations from x equal to negative 1 and x equal to 1 differ only in the sign of b, so subtracting them hands you b in one line.

Look for that shortcut whenever two given points are symmetric about the vertical axis. It turns a three-unknown elimination into a one-line result followed by an easy pair, and it is the reason textbook questions so often include a point at negative one and a point at one.

Adding the same two equations is worth a moment too. It kills b instead and leaves two a plus two c equal to ten, so a plus c is five, which combined with the third equation finishes the job just as quickly from the other direction.

Whichever route you take, finish by testing all three points rather than the one you solved with last. Two of these wrong options satisfy one of the three given points, which is precisely why one substitution is not a check.

McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models §4.10, p. 313 — Exercise 35

62. Say the choice rule in your own words

Explain it to yourself

One decision has carried this whole lesson, and putting it into your own sentences is what makes it stick. Write your answer out before reading the one provided.

Discussion prompt

In three or four sentences, explain how you decide which form to use, why a vertex or a pair of crossings saves you so much work, and why three ordinary points do not. Then say what all three routes have in common at the end.

If your explanation leaned on remembering which form goes with which clue, try writing it again in terms of counting unknowns. Counting is a reason; remembering is not, and reasons survive questions worded in ways you have not seen.

A good test of your own understanding is whether you could invent a question for each route. If you can write down a vertex problem, an intercept problem and a three-point problem and solve your own, the decision rule is yours.

Try the same explanation on the modelling case as well. Data is the situation where no exact route exists at all, and saying why is a good check that you understood the difference between fitting through points and fitting near them.

63. Exit ticket

Exit ticket

One question that uses the whole lesson: choose the form, carry it out, and finish by testing every fact you were given.

\[ \text{crossings at } -6 \text{ and } 3, \quad \text{through } (0, -9) \]

Discussion prompt

Name the form you would use and why, write the function, convert it to standard form, and then say what the lowest point of this parabola is.

The conversion step is worth doing even when nobody asks for it, because standard form immediately confirms the y-intercept: the constant term is negative nine, which is the point you were given.

If you found the vertex by using negative b over two a on the standard form instead, you should get the same answer: negative three halves divided by one, which is negative one and a half. Two routes to one number is a check in itself.

Finally, notice how little algebra this took once the form was chosen. That is the lesson's claim in miniature, and it is why the decision comes before the pen touches the page.

64. What you can do now

Recap

Three exact routes and one approximate one, chosen by a single question: what were you actually given? Everything else in this lesson is Chapter 4 and Lesson 3.4 arithmetic you already had.

what you are givenform to writewhat is left to find
the vertex and one more pointy = a(x − h)² + ka, from one equation
both x-intercepts and one more pointy = a(x − p)(x − q)a, from one equation
three ordinary pointsy = ax² + bx + ca, b and c, from a system of three
a table of measurementsa best-fitting modela, b and c, by quadratic regression

\[ y = -\tfrac12 (x + 1)(x - 4) = -\tfrac12 x^2 + \tfrac32 x + 2 \]

The line above is the shape of every answer in the first three rows: a form chosen for the clue you were given, and the same function rewritten in standard form when that is what was asked for.

The fourth row is the one that is different in kind. No curve passes through real measurements, so regression finds the curve that comes closest to all of them at once, and every answer from it carries the range of the data as a condition.

Two habits are worth carrying out of this lesson. Decide the form before writing anything, and test the finished function on every fact you were given rather than the one you used last.

Chapter 5 takes this further. Once you can build a polynomial from data and features, the same questions get asked about cubics and higher degrees, and the counting argument from this lesson is what tells you how many points each one needs.

OpenStax Algebra and Trigonometry 2e, §5.1 Quadratic Functions §5.1 — the same three forms, and modelling with quadratics

Sources

  1. McDougal Littell Algebra 2 (Texas Edition), Ch. 4 Quadratic Functions and Factoring — Lesson 4.10 Write Quadratic Functions and Models — Larson, Boswell, Kanold & Stiff, McDougal Littell / Houghton Mifflin, 2007, pp. 309-313
  2. OpenStax Algebra and Trigonometry 2e, §5.1 Quadratic Functions
  3. OpenStax Algebra and Trigonometry 2e, §4.3 Fitting Linear Models to Data

Want this taught 1-on-1? Alexander tutors Algebra 2 — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108