3.7 Determinants and Cramer's Rule

The determinant of a 2 by 2 and a 3 by 3 matrix, using a determinant to find the area of a triangle, the coefficient matrix, Cramer's rule for solving a system, and what a zero determinant tells you.

Subject: Algebra 2 · 69 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Evaluate Determinants and Apply Cramer's Rule

Title

Algebra 2 · Lesson 3.7

One number, squeezed out of a square matrix, that measures an area and decides whether a system has a single answer

2. What this lesson gives you

Objectives

Lessons 3.5 and 3.6 treated a matrix as a container: you added them, scaled them, multiplied them. This lesson does the opposite. It squeezes a square matrix down to one single number and then asks what that number knows.

It turns out to know a surprising amount. It measures an area, it finds the size of a triangle from nothing but the coordinates of its corners, and it tells you in advance whether a system of equations has one answer, none, or infinitely many.

Keep the last one in view from the start. It is the objective that turns Cramer's rule from a formula you copy into something you understand, because it explains what the formula is protecting you from.

One habit runs through the whole lesson: compute the determinant of the coefficient matrix before you compute anything else. It costs two products and it tells you whether the rest of the work is even possible.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — Before, Now, Why and the Key Vocabulary

3. What you already know about the three cases

Warm-up

Lesson 3.1 sorted every system of two linear equations into three cases, and you decided which case you were in by comparing the two lines. Bring that back now, because this lesson is about to attach a number to the same decision.

\[ \begin{array}{ll} \text{A:} & 2x + 3y = 7 \quad \text{and} \quad 4x - y = 1 \\ \text{B:} & 2x + 3y = 7 \quad \text{and} \quad 4x + 6y = 5 \end{array} \]

Discussion prompt

The two systems differ only in the second equation. Decide which one has exactly one solution, and say what you compared to reach your answer.

Notice how much work that comparison took: rearrange both equations, read off two slopes, compare them, then check the intercepts. Four steps before you know whether solving is worth starting.

By the end of this lesson you will make the same decision with two multiplications and one subtraction, using only the four coefficients. That single number is the determinant, and system B is the case where it comes out zero.

Lessons 3.5 and 3.6 are also about to be useful. You know how to read a matrix by rows and columns and how to pull one row or one column out of it, and every recipe in this lesson is built out of exactly that.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.1, p. 156 — the three cases for a system of two linear equations

4. The 2 by 2 determinant

Section

Part 1

5. One number attached to every square matrix

Concept

Every square matrix, and only a square matrix, has a determinant: a single real number worked out from its entries. It has two notations, and the book uses both, so learn to read both.

determinant — A real number associated with every square matrix. The determinant of a matrix A is written either det A or with vertical bars in place of the square brackets.

\[ \det\begin{bmatrix} 5 & 4 \\ 3 & 1 \end{bmatrix} = \begin{vmatrix} 5 & 4 \\ 3 & 1 \end{vmatrix} = -7 \]

Figure (svg): The matrix with rows 5, 4 and 3, 1 drawn in square brackets on the left, an arrow to the right, and the same entries drawn between vertical bars with the value negative 7 beside them.

The brackets and the bars hold the same entries and mean different things. One is an array; the other is a number you can compare, divide by, or set equal to zero.

The change of brackets is doing real work. Square brackets mean the array itself, which you can add to another array; vertical bars mean the number, which you can divide by or compare with zero. Writing one when you mean the other is the commonest notation slip in this lesson.

Why only square matrices? Because the recipe you are about to meet runs along diagonals, and a matrix that is wider than it is tall has no diagonal that reaches the far corner. A 2 by 3 matrix simply has no determinant to find.

One more thing worth saying now. This number is not a summary or an average of the entries; it is a specific combination of all of them, and changing any single entry changes it. That is what makes it useful as a test.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — the definition of a determinant

6. Reading the 2 by 2 formula piece by piece

Notation

Here is the whole rule for a 2 by 2 matrix. It is two products and a subtraction, and every part of it is worth a moment before you start using it at speed.

Annotate

On: \( \begin{vmatrix} a & b \\ c & d \end{vmatrix} = ad - cb \)

  • These say that what follows is a number, not a matrix. The bars are not absolute value signs either, which matters: a determinant is allowed to come out negative.
  • The product along the main diagonal, from the top left corner to the bottom right. This one is written first and stays positive.
  • The product along the other diagonal, from the bottom left corner to the top right. This one is subtracted, and the order is not negotiable.
  • The book writes ad minus cb rather than ad minus bc, which is the same product written in the order you meet the letters when you trace the diagonal upwards. Tracing the picture rather than memorising letters is what keeps the order right.

The order of the two products is the only thing in this formula that can go wrong, and it goes wrong often. Swapping them does not give a slightly different answer; it gives exactly the opposite sign, every time.

It helps to say the rule aloud as a sentence about position rather than about letters: top left times bottom right, minus bottom left times top right. Letters change from problem to problem, positions do not.

Notice also that all four entries appear, and each appears exactly once. If a product of yours uses an entry twice, or misses one out, you have traced a diagonal wrongly.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — The Determinant of a Matrix

7. Two diagonals, one subtraction

Picture it

The formula is really a picture, and the picture is easier to carry around than the letters. Look at the two arrows before you look at anything else.

Figure (svg): The letters a, b, c, d drawn as a 2 by 2 determinant between vertical bars, with a blue arrow from a down to d and a red arrow from c up to b, and a legend naming the main diagonal product and the other diagonal product.

Two products, one subtraction, and the order is fixed. Take the diagonals in the other order and every answer comes out with the wrong sign.

The blue arrow runs down to the right through the main diagonal, and its product is added. The red arrow runs up to the right through the other diagonal, and its product is taken away. That is the entire 2 by 2 rule.

Drawing those two arrows on your own paper, in two colours, costs about three seconds and eliminates the sign error completely. It is worth the three seconds; every teacher who has marked this topic will tell you so.

Hold on to the direction of the arrows as well. When the 3 by 3 recipe arrives in Part 3 it is the same idea with three arrows each way, and the downward ones will still be the ones that get added.

8. Example 1(a): the determinant of the matrix with rows 5, 4 and 3, 1

Worked example

The book's first example, in full. Two products, one subtraction, and a deliberate check at the end that uses a property of determinants rather than just repeating the arithmetic.

\[ \text{Evaluate } \begin{vmatrix} 5 & 4 \\ 3 & 1 \end{vmatrix} \]

Multiply along the main diagonal

Why: top left times bottom right, so the 5 and the 1. This product keeps its sign.

\[ 5(1) = 5 \]

Multiply along the other diagonal

Why: bottom left times top right, so the 3 and the 4. Compute it second, because it is the one that will be subtracted.

\[ 3(4) = 12 \]

Keep the two products apart on the page rather than combining them in your head. The whole difficulty of this topic is bookkeeping, not arithmetic.

Subtract the second product from the first

Why: the definition fixes the order: the main diagonal product comes first and the other is taken away from it.

\[ 5 - 12 = -7 \]

Figure (svg): The determinant with rows 5, 4 and 3, 1, with a blue arrow through the 5 and the 1 labelled 5 times 1 is 5, and a red arrow through the 3 and the 4 labelled 3 times 4 is 12, and the subtraction 5 minus 12 is negative 7 written beneath.

The two products are 5 and 12, and the picture fixes which of them is taken away from which. Nothing here depends on the entries being small or positive.

The answer is negative, and that is not a mistake to hunt for. A determinant is a signed number, and Part 2 will show you exactly what the sign is recording.

Check by swapping the two rows

Why: swapping two rows of a matrix reverses the sign of its determinant, so the swapped version must come out as positive 7. It does, which confirms both products at once.

\[ \begin{vmatrix} 3 & 1 \\ 5 & 4 \end{vmatrix} = 3(4) - 5(1) = 12 - 5 = 7 \]

That is a better check than recomputing the same two products, because a check that repeats your original work will happily confirm your original mistake.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — Example 1(a)

9. Exercise 9: a determinant with one negative entry

Fill the middle

Your turn, on a matrix from the exercise set. One entry is negative, which changes nothing about the method and quite a lot about the care needed.

Fill in the blanks

\begin-57 9 & 4 \\ 3 & -5 \end___ = 9(-5) - 3(4) = -45 - 12 = ___

Why: The main diagonal product is 9 times negative 5, which is negative 45. The other diagonal product is 3 times 4, which is 12, and it gets subtracted. Taking 12 away from negative 45 moves further down the number line, not back towards zero, so the answer is negative 57.

Watch the second subtraction. Both terms are working in the same direction here: the first product is already negative, and then you take a positive 12 away from it, so the answer is further from zero than either product.

A quick way to predict the sign before computing: if one diagonal product is negative and the other is positive, the subtraction cannot cancel anything, so expect a large answer. If both have the same sign, expect a small one.

Hold on to this determinant. The matrix with rows 9, 4 and 3, negative 5 is the coefficient matrix of the system in Example 3, and negative 57 is the number that whole solution will be divided by.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 207 — Exercise 9

10. Guided Practice 1: two negatives in one product

Worked example

This checkpoint from page 204 puts the negative entry on the other diagonal, which is where the double negative can bite. Work out each product on its own line before subtracting.

\[ \text{Evaluate } \begin{vmatrix} 3 & -2 \\ 6 & 1 \end{vmatrix} \]

Multiply along the main diagonal

Why: 3 times 1, with no signs to worry about.

\[ 3(1) = 3 \]

Multiply along the other diagonal

Why: 6 times negative 2. A positive times a negative is negative, so this product is negative 12.

\[ 6(-2) = -12 \]

Now the step that catches people. You must subtract that product, and the product itself is negative, so you are subtracting a negative number. Write it in brackets first and simplify afterwards.

Subtract, keeping the brackets

Why: taking away negative 12 is the same as adding 12, which is why the answer grows instead of shrinking.

\[ 3 - (-12) = 3 + 12 = 15 \]

If you got negative 9, you subtracted 12 instead of subtracting negative 12. That is by far the most common wrong answer here, and writing the brackets is what prevents it.

Check against the area picture coming in Part 2

Why: the rows (3, negative 2) and (6, 1) point into different quarters of the plane, so they genuinely span a shape, and a determinant of 15 says that shape has area 15. A zero here would have meant the two rows lay on one line, and they plainly do not.

\[ \begin{vmatrix} 3 & -2 \\ 6 & 1 \end{vmatrix} = 15 \ne 0 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — Guided Practice 1

11. Which of these determinants are zero?

Discrimination

Six matrices, and the only question is whether the determinant is zero or not. Try to answer each one by looking at the rows rather than by computing, then compute to confirm.

Sort into buckets

Sort each matrix by whether its determinant is zero.

Determinant is zero
[[2, 3], [4, 6]]; [[1, 2], [2, 4]]; [[-2, -3], [4, 6]]
Determinant is not zero
[[5, 4], [3, 1]]; [[9, 4], [3, -5]]; [[3, 8], [2, 5]]
zero
In each of these the second row is a fixed multiple of the first: twice it, twice it again, and negative two times it. Multiplying a row by a number multiplies both diagonal products by that same number, so the difference between them stays zero.
no
Here no multiple of one row gives the other, and the two diagonal products come out different: negative 7, negative 57 and negative 1 respectively.

The pattern is worth naming. A determinant is zero exactly when one row is a multiple of the other, and you can usually see that faster than you can multiply.

Look back at the warm-up. System B had coefficient rows 2, 3 and 4, 6, which is the first matrix on this slide, and that system was the one with no solution. That is not a coincidence, and Part 6 is where it gets explained properly.

Try the same test on the columns while you are here. In the first matrix the second column is one and a half times the first, and the determinant is zero again. Rows or columns, proportional is proportional.

12. What happens if you double a row?

Prediction

Commit to an answer before you compute anything. Take any 2 by 2 matrix, double every entry of its first row only, and ask what that does to the determinant.

\[ \begin{vmatrix} 5 & 4 \\ 3 & 1 \end{vmatrix} = -7 \qquad \text{so what is} \quad \begin{vmatrix} 10 & 8 \\ 3 & 1 \end{vmatrix}\;? \]

Predict first

You double every entry in the first row of a 2 by 2 matrix. What happens to its determinant?

  • It stays the same
  • It doubles
  • It quadruples
  • It becomes zero

Correct: It doubles

Why: Each diagonal product uses exactly one entry from the first row, so doubling that row doubles both products, and doubling both parts of a subtraction doubles the difference. Here 10 times 1 minus 3 times 8 is 10 minus 24, which is negative 14, exactly twice negative 7. It would quadruple only if you doubled both rows.

The reason is worth more than the answer. Each of the two products takes exactly one entry from the first row, so each product doubles, and doubling both parts of a subtraction doubles the result.

That also tells you what doubling both rows would do. Each product would then pick up two factors of 2, so the determinant would be multiplied by 4. Exercise 39 asks you to justify exactly that.

Once Part 2 gives you the area reading, this stops being an algebra fact and becomes obvious. Stretching one side of a parallelogram to twice its length doubles the area, which is all this prediction is saying.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 39(b)

13. What the determinant measures

Section

Part 2

14. The two rows are two arrows, and the determinant is their area

Concept

So far the determinant is a recipe. Here is what it is actually measuring, and once you have seen it you will never again wonder why zero is the value that matters.

Read each row of the matrix as an arrow drawn from the origin: the row 4, 0 is an arrow 4 units east, and the row 0, 3 is an arrow 3 units north. Two arrows from one point span a parallelogram.

\[ \begin{vmatrix} 4 & 0 \\ 0 & 3 \end{vmatrix} = 4(3) - 0(0) = 12 \]

Figure (svg): Two arrows from the origin, one 4 units along the horizontal axis and one 3 units up the vertical axis, with the rectangle they span shaded and labelled area 12.

Read the two rows of the matrix as two arrows from the origin. The determinant is the area of the shape they span, which here is an ordinary rectangle 4 wide and 3 tall.

This first case is deliberately easy: the two arrows are at right angles, so the parallelogram is a plain rectangle, 4 wide and 3 tall, with area 12. The determinant is 12 as well.

That is the whole claim, and it holds for every 2 by 2 matrix: the determinant is the area of the parallelogram spanned by its two rows. Not a metaphor and not an approximation, but the same number.

You can run the test on the columns instead and get the same answer, because a matrix and its sideways flip have the same determinant. For now, rows are the easier picture to hold.

OpenStax Algebra and Trigonometry 2e, §11.8 Solving Systems with Cramer's Rule §11.8 — determinants as areas, alongside Cramer's rule

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — the area of a triangle from a determinant

15. Slide the top edge across: the area does not care

Intuition

A rectangle is a special case, so test the claim on a shape that is genuinely slanted. Keep the first arrow at 4, 0 and tip the second one over so it reaches 2, 3 instead of 0, 3.

\[ \begin{vmatrix} 4 & 0 \\ 2 & 3 \end{vmatrix} = 4(3) - 2(0) = 12 \]

Figure (svg): The same base arrow 4 units long, with the second arrow now going 2 across and 3 up, so the shape is a slanted parallelogram of the same height, its area still 12, with the original rectangle drawn dashed behind it.

Sliding the top edge sideways keeps the base at 4 and the height at 3, so the area does not budge. That is why the determinant of the sheared matrix is 12 as well.

The determinant has not moved. Neither has the area, and the picture says why: the base is still 4 units long and the top edge is still 3 units above it, so base times height is still 12.

This is worth sitting with, because it is the first sign that the determinant is measuring something geometric rather than just crunching the entries. The entries changed, the shape changed, and the number did not.

It also explains an algebraic fact you will meet later. Adding a multiple of one row to another row shears the parallelogram without changing its area, which is exactly why row operations leave a determinant alone.

Try the extreme case in your head. Keep tipping the second arrow over until it points along the first one; the height drops to nothing and the area, and the determinant, go to zero. That case is two slides away.

16. Example 1(a) as a sliver of area, and what the minus sign means

Picture it

Back to the book's first example, now as a picture. The rows 5, 4 and 3, 1 are two arrows pointing in fairly similar directions, so the parallelogram they span is thin.

Figure (svg): Two arrows from the origin, one to the point 5 comma 4 and one to the point 3 comma 1, spanning a thin sliver of a parallelogram whose fourth corner is 8 comma 5, labelled area 7.

The determinant came out as negative 7 and the sliver has area 7. The minus sign records the turn from the first row to the second, not a shortage of area.

The determinant of that matrix was negative 7, and the sliver in the picture has area 7. Areas are never negative, so the minus sign must be recording something else, and it is: the order of the two rows.

Trace from the first row to the second. Going from the arrow at 5, 4 round to the arrow at 3, 1 turns clockwise, and a clockwise pair gives a negative determinant. Swapping the rows turns the pair anticlockwise and flips the sign, which is exactly the check you used on the worked example.

So the determinant carries two pieces of information in one number: how much area, and which way round. The area part is what the triangle formula in Part 4 wants, which is why that formula has a plus or minus in front of it.

Thin slivers are the interesting case. The closer the two arrows come to pointing the same way, the smaller the area and the smaller the determinant, and a system built from nearly proportional rows is a system whose answer is hard to pin down.

17. Slide one entry until the determinant hits zero

Tweak it

Here is the determinant of a matrix with one entry left free. The bottom right entry runs along the horizontal axis, and the curve is the determinant it produces.

Parameter explorer

The horizontal axis is the bottom right entry and the curve is the determinant. With the top left entry set to 2, read off the value of the bottom right entry that makes the determinant zero, then look at what the two rows are doing there.

\[ \begin{vmatrix} {a} & 3 \\ 4 & d \end{vmatrix} = {a}d - 12 \]

  • a — from 1 to 5: top left entry

The line crosses zero in exactly one place, and that place is where the second row becomes a multiple of the first. Everywhere else the determinant is some other number, positive on one side and negative on the other.

Change the top left entry and watch the crossing point move. The determinant is a straight line in this entry, so there is always exactly one value that kills it, which is worth remembering: a matrix is almost never singular by accident.

The crossing is also the moment the parallelogram collapses. On one side of it the two arrows turn one way, on the other side they turn the other, and at the crossing they lie on top of each other with nothing in between.

18. A zero determinant means the two rows lie on one line

Concept

Now the case the whole lesson is built around. Take the rows 2, 3 and 4, 6, which is the matrix the slider just found, and draw them.

\[ \begin{vmatrix} 2 & 3 \\ 4 & 6 \end{vmatrix} = 2(6) - 4(3) = 12 - 12 = 0 \]

Figure (svg): Two arrows from the origin, one to 2 comma 3 and a longer one to 4 comma 6, lying exactly on top of one another along a dashed line, with no parallelogram between them and the area marked zero.

The second row is twice the first, so the two arrows point the same way and span nothing. A determinant of zero is this picture: a parallelogram flattened onto a line.

The second arrow is exactly twice the first, so it points in the same direction and lands further along the same line. There is no parallelogram at all, only a segment, and a segment has no area.

So the three descriptions are one description. The determinant is zero, the rows are proportional, and the parallelogram has collapsed onto a line: say any one of them and you have said all three.

Keep this picture for Part 6. When those two rows are the coefficients of a system, the collapse is what stops the system having a single answer, and it is also what makes Cramer's rule divide by zero.

It is worth noticing how fragile the collapse is. Move either arrow by the smallest amount off the line and an area reappears. Zero determinants are exact conditions, not approximate ones, which is why you test for them rather than eyeballing them.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the condition det A is not zero in Cramer's rule

19. Why should a collapsed parallelogram mean no single answer?

Socratic

Two apparently unrelated facts have now turned up side by side: a zero determinant collapses a parallelogram, and a zero determinant is the case where a system misbehaves. Argue the connection out before the lesson hands it to you.

\[ \begin{array}{l} 2x + 3y = 7 \\ 4x + 6y = 5 \end{array} \qquad \text{coefficient rows } (2, 3) \text{ and } (4, 6) \]

Discussion prompt

The coefficient rows of this system are the two arrows that collapsed onto one line. Explain in your own words why proportional coefficient rows stop a system having exactly one solution, using either the two lines from Lesson 3.1 or the two arrows from this part.

The reason this matters is that it makes one number do the work of the whole Lesson 3.1 comparison. Rearranging two equations and comparing two slopes becomes two multiplications and a subtraction.

What the determinant will not tell you is which of the two bad cases you are in. Zero says no single answer, and elimination is still what separates no solution from infinitely many.

Hold on to the phrase that connects the parts: proportional rows, zero area, no unique solution. Every remaining part of this lesson leans on it at least once.

20. The 3 by 3 determinant

Section

Part 3

21. Repeat the first two columns, then take six products

Concept

A 3 by 3 matrix has a determinant too, and the book's method for it is a direct extension of the diagonal picture. There is one piece of setup first, and skipping it is what makes the rest look impossible.

  1. Step 1. Copy the first two columns again, immediately to the right of the determinant.
  2. Step 2. Add the three products running down to the right, subtract the three running up to the right.

\[ \begin{vmatrix} a & b & c \\ d & e & f \\ g & h & i \end{vmatrix} = (aei + bfg + cdh) - (gec + hfa + idb) \]

Figure (svg): A 3 by 3 determinant of the letters a to i between vertical bars, with the first two columns copied again to the right of the bars in a faint colour, three blue arrows running down to the right through the array and three red arrows running up to the right.

Each arrow passes through exactly one entry from every row and every column. The copies exist only so the three upward diagonals have somewhere to run.

The copied columns are scaffolding. They are not part of the matrix, they do not change its size, and they exist only so that all six diagonals have three entries to pass through.

Count what each product uses: one entry from every row and one from every column, never two from the same. That is the rule behind the pattern, and it is the fastest way to catch a mistraced diagonal.

Six products is a lot of bookkeeping for one number, so give yourself room on the page. Write the three downward products on one line and the three upward ones underneath, then subtract the totals in a third line.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — Determinant of a 3 by 3 Matrix

22. Example 1(b): the six products with their values

Picture it

Here is the book's second example set up the same way, with real numbers in the cells. Read the arrows first, then read the products.

Figure (svg): Example 1b drawn as a 3 by 3 determinant with its first two columns copied to the right, the three downward products negative 4, 0 and 48 listed in blue adding to 44, and the three upward products negative 9, 0 and 8 listed in red adding to negative 1, with the final subtraction giving 45.

Only two of the six products needed real work: the 48 and the negative 9. Spotting the zero entry first tells you which products are going to vanish.

The three downward products come to negative 4, 0 and 48, adding to 44. The three upward products come to negative 9, 0 and 8, adding to negative 1. The determinant is the first total minus the second.

Notice how much the single zero entry in the middle row does for you. It lies on one downward diagonal and one upward diagonal, so two of the six products are zero before you start multiplying anything.

That is a habit worth forming: scan for zeros before you compute. A row or a column with two zeros in it can reduce this whole calculation to a single product.

23. Example 1(b): a 3 by 3 determinant from start to finish

Worked example

The same example written out as working, one product per line. It is longer than a 2 by 2 but not harder, and the length is the whole reason to be systematic.

\[ \text{Evaluate } \begin{vmatrix} 2 & -1 & -3 \\ 4 & 1 & 0 \\ 3 & -4 & -2 \end{vmatrix} \]

Repeat the first two columns to the right

Why: the array now has five columns, and every one of the six diagonals has three entries to run through.

\[ \begin{array}{ccc|cc} 2 & -1 & -3 & 2 & -1 \\ 4 & 1 & 0 & 4 & 1 \\ 3 & -4 & -2 & 3 & -4 \end{array} \]

Multiply along the three downward diagonals

Why: starting from each entry of the top row and stepping down and right: 2 times 1 times negative 2, then negative 1 times 0 times 3, then negative 3 times 4 times negative 4.

\[ -4, \quad 0, \quad 48 \qquad \text{sum } 44 \]

The middle product vanished because the top row's negative 1 happens to share a diagonal with the zero. Free products like that are the reward for setting the array out properly.

Multiply along the three upward diagonals

Why: starting from each entry of the bottom row and stepping up and right: 3 times 1 times negative 3, then negative 4 times 0 times 2, then negative 2 times 4 times negative 1.

\[ -9, \quad 0, \quad 8 \qquad \text{sum } -1 \]

Subtract the upward total from the downward total

Why: and mind the double negative: taking away negative 1 adds 1.

\[ 44 - (-1) = 44 + 1 = 45 \]

The book's own line reads 44 minus negative 1 equals 45, so if you wrote 43 you subtracted instead of adding. That single step is where most of the lost marks in this exercise set live.

Check that every product used each row and each column once

Why: six products, three added and three subtracted, and each one takes exactly one entry from every row and every column. The zero entry appears in exactly two products, which is what a single zero should do, so no diagonal was traced twice.

\[ \begin{vmatrix} 2 & -1 & -3 \\ 4 & 1 & 0 \\ 3 & -4 & -2 \end{vmatrix} = 45 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — Example 1(b)

24. Put the 3 by 3 recipe in order

Ranking

The steps of this method are easy to name and easy to do in the wrong order, which produces a number that is not the determinant of anything. Fix the order in your head now.

Put in order

Put the steps of the 3 by 3 determinant in the order you carry them out.

  1. Copy the first two columns to the right of the determinant
  2. Multiply along each of the three downward diagonals and add the three products
  3. Multiply along each of the three upward diagonals and add the three products
  4. Subtract the upward total from the downward total

Why: The copying has to come first, because without it three of the diagonals run off the edge of the array and cannot be traced at all. The two totals can honestly be computed in either order, but the subtraction cannot happen until both exist, and it must take the upward total away from the downward one rather than the other way round.

The only step whose position is truly forced is the copying, and it is also the step people skip when they are in a hurry. Without it, half the diagonals do not exist.

The subtraction at the end is forced in a different way: the two totals are not interchangeable. Downward minus upward gives the determinant, and upward minus downward gives its negative.

One thing this list does not include is copying the last two columns instead of the first two. That produces six products, all of them wrong, and it is a real mistake rather than an invented one.

25. Guided Practice 3: a 3 by 3 with a zero in the corner

Worked example

This checkpoint has larger entries and a zero in the bottom left, which is exactly the sort of matrix where a careless downward product goes missing. Work through all six.

\[ \text{Evaluate } \begin{vmatrix} 10 & -2 & 3 \\ 2 & -12 & 4 \\ 0 & -7 & -2 \end{vmatrix} \]

Repeat the first two columns

Why: so that the six diagonals can be traced across a five-column array.

\[ \begin{array}{ccc|cc} 10 & -2 & 3 & 10 & -2 \\ 2 & -12 & 4 & 2 & -12 \\ 0 & -7 & -2 & 0 & -7 \end{array} \]

Multiply along the three downward diagonals

Why: 10 times negative 12 times negative 2, then negative 2 times 4 times 0, then 3 times 2 times negative 7.

\[ 240, \quad 0, \quad -42 \qquad \text{sum } 198 \]

The first product is positive because two negatives multiply to a positive, and then the 10 keeps it positive. Getting negative 240 there is the single most likely slip on this matrix.

Multiply along the three upward diagonals

Why: 0 times negative 12 times 3, then negative 7 times 4 times 10, then negative 2 times 2 times negative 2.

\[ 0, \quad -280, \quad 8 \qquad \text{sum } -272 \]

Subtract the upward total from the downward total

Why: 198 take away negative 272, so the two amounts add.

\[ 198 - (-272) = 198 + 272 = 470 \]

An answer in the hundreds should not worry you. Determinants grow quickly with the size of the entries, and nothing in this lesson requires them to be small.

Check the two products that involved the zero

Why: the zero in the bottom left corner sits on one downward diagonal and one upward diagonal, so exactly two of the six products should be zero, and they are. If three had vanished, a diagonal would have been traced through that corner twice.

\[ \begin{vmatrix} 10 & -2 & 3 \\ 2 & -12 & 4 \\ 0 & -7 & -2 \end{vmatrix} = 470 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — Guided Practice 3

26. Guided Practice 2: finish the subtraction

Fill the middle

The six products of this checkpoint have already been computed for you. What is left is the part where the signs decide everything, so take it slowly.

\[ \begin{vmatrix} 4 & -1 & 2 \\ -3 & -2 & -1 \\ 0 & 5 & 1 \end{vmatrix} \]

Fill in the blanks

(-8 + 0 - 30) - (0 - 20 + 3) = -38 - (-17) = -21

Why: The three upward products are 0, negative 20 and 3, which add to negative 17. Subtracting negative 17 from negative 38 means adding 17, so the determinant is negative 21. Subtracting 17 instead would have given negative 55, which is the usual wrong answer here.

The bracket around the second total is not decoration. It holds the three upward products together so that the minus sign in front applies to their sum, and not just to the first of them.

Notice the direction the answer moves. Negative 38 with 17 added lands at negative 21, which is closer to zero than you started. Whenever a determinant comes out closer to zero than the first total, the upward products were negative overall.

A sanity check on the sign: the second row of this matrix is not a multiple of the first, so the determinant had no business being zero, and it is not.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — Guided Practice 2

27. Find the error: the two totals were added

Error analysis

This is the book's own error-analysis exercise, and the mistake in it is the one this whole part is designed to prevent. Find it before you read the notes.

\[ \begin{vmatrix} 2 & 0 & -1 \\ 4 & 1 & 6 \\ -3 & 2 & 5 \end{vmatrix} \]

Annotate

On: \( (10 + 0 - 8) + (3 + 24 + 0) = 2 + 27 = 29 \)

  • These are the three downward products, and they are right: 2 times 1 times 5 is 10, then 0 times 6 times negative 3 is 0, then negative 1 times 4 times 2 is negative 8. They total 2.
  • These are the three upward products, and they are right as well: negative 3 times 1 times negative 1 is 3, then 2 times 6 times 2 is 24, then 5 times 0 times 4 is 0. They total 27.
  • Here is the error. The upward total must be subtracted, not added. Every product was computed correctly and the answer is still wrong, which is what makes this mistake so easy to make and so hard to spot.
  • Two minus twenty-seven is negative twenty-five, so the determinant is negative 25, not 29. Notice that the wrong answer is positive and the right one is negative: the sign alone would have given the error away.

\[ (10 + 0 - 8) - (3 + 24 + 0) = 2 - 27 = -25 \]

Why is this slip so common? Because the six products are the hard part and the subtraction is the easy part, so attention has usually run out by the time it arrives.

One defence is to write the minus sign in front of the second bracket before you compute anything that goes inside it. Then the hard work happens inside a structure that is already correct.

Another is to glance at the sign of your answer. A determinant whose upward products dominate must come out negative, and this one clearly did.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 207 — Exercise 19

28. Trap: the subtraction the wrong way round

Trap

The trap

Evaluating the determinant with rows 5, 4 and 3, 1, here is a very common first line:

\[ 3(4) - 5(1) \]

\[ = 12 - 5 = 7 \]

Both products are right. The arithmetic is right. The answer is the exact opposite of the truth.

The fix

The main diagonal product comes first. Top left times bottom right, then take away bottom left times top right.

\[ 5(1) - 3(4) \]

\[ = 5 - 12 = -7 \]

A wrong sign is not a small error here. In Part 4 it turns an area into a negative number, and in Part 5 it flips the sign of both answers a system has.

The cure is the picture rather than the letters: draw the two arrows, start at the top left corner, and the order looks after itself.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 203 — the order of the products in the Key Concept box

29. The 2 by 2 and the 3 by 3 side by side

Comparison

Before leaving determinants as a calculation, set the two recipes against each other. The shape of the work is the same and the amount of it is not.

Comparison matrix

2 by 23 by 3
columns copied firstnonethe first two
products to compute26
products added13
entries in each product23
what the size of the answer measuresarea of a parallelogramvolume of a box

The last row is a genuine promise, not a slogan. Three rows of three numbers are three arrows in space, and the 3 by 3 determinant is the volume of the slanted box they span, with a sign for orientation just as before.

That means the zero story carries over exactly. A 3 by 3 determinant is zero when its three arrows lie in one flat plane, which is the three-variable version of two arrows lying on one line.

Count the work in the second row and you can see why bigger determinants are not done by hand. Four by four needs twenty-four products, and the diagonal trick stops working altogether at that size.

30. Area from a determinant

Section

Part 4

31. A triangle is half of the parallelogram

Concept

The area reading from Part 2 is about to pay for itself. Suppose a triangle has one corner at the origin and the other two at 5, 2 and 3, 6. Those two corners are exactly the two arrows of a matrix.

\[ \begin{vmatrix} 5 & 2 \\ 3 & 6 \end{vmatrix} = 5(6) - 3(2) = 30 - 6 = 24 \]

Figure (svg): A triangle with corners at the origin, 5 comma 2 and 3 comma 6, shaded darker, with a second copy of it turned round to complete a parallelogram whose fourth corner is 8 comma 8, so the triangle is visibly half of it.

The triangle and its turned copy tile the parallelogram exactly. So the triangle holds half of the determinant, which is where the one half in the area formula comes from.

The determinant 24 is the area of the whole parallelogram. The triangle is one of the two identical halves that tile it, so the triangle has area 12. That factor of one half is the only new thing in the area formula.

The book's version works for any triangle, not just one with a corner at the origin, and it hides that half in front of a 3 by 3 determinant. The column of ones is what lets the triangle sit anywhere in the plane.

\[ \text{Area} = \pm\tfrac12 \begin{vmatrix} x_1 & y_1 & 1 \\ x_2 & y_2 & 1 \\ x_3 & y_3 & 1 \end{vmatrix} \]

Check the two versions against each other on this triangle: put 0, 0 in the third row of the big determinant and the six products collapse to the two you already computed, giving 24 again, and half of it is 12.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — Area of a Triangle

32. Reading the area formula piece by piece

Notation

Four things in this formula need explaining, and the plus or minus is the one that gets ignored. Take them in order.

Annotate

On: \( \text{Area} = \pm\tfrac12 \begin{vmatrix} x_1 & y_1 & 1 \\ x_2 & y_2 & 1 \\ x_3 & y_3 & 1 \end{vmatrix} \)

  • One row per vertex, in the order you happen to list them. The order affects the sign of the determinant and nothing else, which is precisely what the plus or minus is there to absorb.
  • This is what lets the triangle sit anywhere rather than having a corner at the origin. It is not a coefficient and not a count; it is the column that makes the determinant depend on the differences between the vertices instead of on their raw positions.
  • The determinant gives the parallelogram, and a triangle is half of it, exactly as in the picture on the previous slide.
  • Choose whichever sign makes the answer positive. An area cannot be negative, and the determinant's sign only records the order you listed the vertices in, so you are allowed to discard it.

The plus or minus is not the book being vague. It is telling you that the sign of the determinant carries no information about the triangle, only about the order of your list, so you may throw it away.

That also means there is no wrong order to list the vertices in. Two different orders give determinants that are negatives of each other, and both give the same area.

One small practical point: keep the column of ones as a column of ones. Every entry in it is 1, whatever the coordinates are, and forgetting one of them changes the whole answer.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — the plus or minus in the area formula

33. Example 2: the area of the sea lion triangle

Worked example

Off the California coast, a triangle of ocean between Bodega Bay, the Farallon Islands and Ano Nuevo Island is home to huge populations of sea lions and seals. The map gives coordinates in miles.

Nobody is going to measure this region with a ruler. Three coordinate readings and one determinant are all the information the calculation needs.

\[ (-1, 41), \quad (38, -43), \quad (0, 0) \]

Write the three vertices as rows, with a column of ones

Why: the order does not matter, so take them as the book lists them.

\[ \text{Area} = \pm\tfrac12 \begin{vmatrix} -1 & 41 & 1 \\ 38 & -43 & 1 \\ 0 & 0 & 1 \end{vmatrix} \]

Multiply along the three downward diagonals

Why: negative 1 times negative 43 times 1, then 41 times 1 times 0, then 1 times 38 times 0. The row of zeros kills two of the three.

\[ 43, \quad 0, \quad 0 \qquad \text{sum } 43 \]

Multiply along the three upward diagonals

Why: 0 times negative 43 times 1, then 0 times 1 times negative 1, then 1 times 41 times 38.

\[ 0, \quad 0, \quad 1558 \qquad \text{sum } 1558 \]

Having a vertex at the origin is worth a lot of arithmetic. Two of its three coordinates are zero, so four of the six products disappear and only two real multiplications survive.

Subtract, then halve, choosing the sign that makes the area positive

Why: 43 minus 1558 is negative 1515, and half of 1515 is 757.5. The minus sign records the order the vertices were listed in, so discard it.

\[ \text{Area} = \pm\tfrac12(43 - 1558) = \tfrac12(1515) = 757.5 \]

Figure (svg): A coordinate map in miles with the Farallon Islands at the origin, Bodega Bay at negative 1 comma 41 and Ano Nuevo Island at 38 comma negative 43, with the triangle joining them shaded.

The three readings are coordinates in miles, so the determinant they build is measured in square miles. The shape is long and thin, which is why estimating its area by eye is hopeless.

Check against a rough triangle you can compute in your head

Why: the region stretches about 84 miles from north to south and reaches about 38 miles out to sea, so a triangle of base 84 and height 38 would hold about 1600 square miles; the real region is thinner than that, and 758 is comfortably less. The book reports about 758 square miles.

\[ \tfrac12(84)(38) = 1596 \quad \text{versus} \quad 757.5 \]

Rough bounds like that are the only realistic check on a figure this size, and they catch the errors that matter: a lost factor of one half, or a determinant that came out ten times too big.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — Example 2

34. Trap: reporting a negative area

Trap

The trap

Finishing the sea lion calculation, here is a tempting last line:

\[ \tfrac12(43 - 1558) = -757.5 \]

The arithmetic is perfect. The determinant really is negative 1515, and half of it really is negative 757.5.

But the question asked for an area, and no region of ocean has a negative size.

The fix

The formula carries a plus or minus in front precisely so that you can choose the sign that makes the answer positive.

\[ \text{Area} = \tfrac12(1515) = 757.5 \]

The minus sign was never about the triangle. Listing the three vertices in the other order gives positive 1515, and the same 757.5 square miles.

So treat the sign as a note about your list, not a feature of the region, and write the area as a positive number every time.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — the meaning of the plus or minus sign

35. Estimate before you compute

Estimation

Here is the triangle from the next checkpoint, with the smallest rectangle that will hold it drawn in dashes. Before touching a determinant, commit to a range for its area.

Figure (svg): The triangle with corners 5 comma 11, 9 comma 2 and 1 comma 3 drawn on axes, inside a dashed rectangle from 1 comma 2 to 9 comma 11 that measures 8 by 9.

The smallest rectangle round a triangle always holds more than the triangle does, and usually between half and all of it. That makes it a cheap check on any area you compute.

Predict first

The triangle has vertices at 5 comma 11, 9 comma 2 and 1 comma 3, and the dashed box around it measures 8 by 9. Roughly how large is the triangle's area?

  • under 20 square units
  • between 20 and 50 square units
  • between 50 and 80 square units
  • more than 80 square units

Correct: between 20 and 50 square units

Why: The box holds 72 square units and the triangle fills a good deal of it but certainly not all, so anything above 72 is impossible and anything very small is unlikely. A triangle drawn inside a box like this usually covers something near half of it, which points at the middle of that range, and the exact answer turns out to be 34.

The bounding box is the cheapest estimate there is. The triangle cannot be bigger than the box, and a triangle with a vertex on each of three sides of the box is usually somewhere near half of it.

Estimating first changes what an answer means. A determinant that hands you 340 or 3.4 is now obviously wrong, whereas without the estimate both look like plausible outputs of a calculation.

This habit costs about ten seconds and it is the only defence against a lost factor of one half, which is the error this formula produces more than any other.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 23, the same triangle in a box

36. Guided Practice 4: the area of triangle A, B, C

Fill the middle

Now compute the triangle you just estimated. The three vertices are A at 5 comma 11, B at 9 comma 2 and C at 1 comma 3, and the six products have been done for you.

\[ \begin{vmatrix} 5 & 11 & 1 \\ 9 & 2 & 1 \\ 1 & 3 & 1 \end{vmatrix} = (10 + 11 + 27) - (2 + 15 + 99) \]

Fill in the blanks

\text116 = \pm\tfrac12(48 - 34) = \tfrac12(68) = ___

Why: The upward products 2, 15 and 99 add to 116, so the determinant is 48 minus 116, which is negative 68. Half of 68 is 34 square units, and the sign is discarded because an area is positive. That lands inside the range you predicted, and well under the 72 square units of the bounding box.

Compare the answer with your estimate. The box was 72 square units, the triangle is 34, and the ratio is a little under half, which is exactly what the picture suggested.

It is worth noticing what you never had to do here: no base, no height, no perpendicular from a vertex to the opposite side. Six coordinates went in and an area came out.

That is the real gain of this formula. Finding a height means finding a perpendicular distance, which is a much harder calculation than any of the multiplications you just performed.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 204 — Guided Practice 4

37. Exercise 41: the triangular garden

Worked example

You are turning a triangular patch of your yard into a garden. Its corners are at 0, 0 and 5, 2 and 3, 6, measured in feet, and you want to know how much ground you are digging.

You have met this triangle before: it is the one from the half-a-parallelogram picture at the start of this part. That means you already know the answer and can watch the formula produce it.

\[ (0, 0), \quad (5, 2), \quad (3, 6) \]

Write the three vertices as rows with a column of ones

Why: the corner at the origin goes in the first row, where its zeros will do the most good.

\[ \text{Area} = \pm\tfrac12 \begin{vmatrix} 0 & 0 & 1 \\ 5 & 2 & 1 \\ 3 & 6 & 1 \end{vmatrix} \]

Multiply along the three downward diagonals

Why: 0 times 2 times 1, then 0 times 1 times 3, then 1 times 5 times 6. Two of the three vanish on the zeros.

\[ 0, \quad 0, \quad 30 \qquad \text{sum } 30 \]

Multiply along the three upward diagonals

Why: 3 times 2 times 1, then 6 times 1 times 0, then 1 times 5 times 0.

\[ 6, \quad 0, \quad 0 \qquad \text{sum } 6 \]

Subtract and halve

Why: 30 take away 6 is 24, and half of 24 is 12 square feet.

\[ \text{Area} = \tfrac12(30 - 6) = \tfrac12(24) = 12 \]

Twelve square feet of garden, from three pairs of coordinates and no measuring tape. If you had paced out a base and a height instead, the height would have been the hard part.

Check against the 2 by 2 determinant of the two edges

Why: with a corner at the origin the other two corners are the rows of a 2 by 2 matrix, whose determinant is the parallelogram: 5 times 6 minus 3 times 2 is 24, and the triangle is half of that. The two routes agree, which also confirms the column of ones is doing what the previous slide claimed.

\[ \begin{vmatrix} 5 & 2 \\ 3 & 6 \end{vmatrix} = 24, \qquad \tfrac12(24) = 12 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 41

38. The Bermuda Triangle, from three coordinates

Real world

Exercise 40 takes the same formula to a region nobody could measure directly. The Bermuda Triangle is the stretch of Atlantic between Miami, Bermuda and Puerto Rico, and the map gives coordinates in miles.

Figure (svg): A coordinate map in miles with Miami at the origin, Bermuda at 938 comma 454 and Puerto Rico at 900 comma negative 518, joined into a shaded triangle.

Three coordinates, one determinant, and an area no tape measure could reach. The numbers are large but the recipe is the one you used on a triangle 8 units wide.

Discussion prompt

Miami is at 0, 0, Bermuda at 938, 454 and Puerto Rico at 900, negative 518. Set up the area determinant, use the vertex at the origin to kill four of the six products, and find the area to the nearest thousand square miles.

The scale of the numbers is the only thing that has changed. The recipe is the one you used on a garden 5 feet across, and it does not care whether the units are feet or miles.

This is what makes the determinant method worth learning rather than just the base-times-height one. Survey data arrives as coordinates, and this formula eats coordinates directly.

It is also why the sign has to be discarded rather than fixed. A surveyor walking the corners in the other order would get negative 894,484, and the plot of ocean would be no smaller.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 40

39. Three points on one line

Prediction

Test the area formula on a case where there is nothing to measure. The points 1 comma 2, then 3 comma 6, then 5 comma 10 all satisfy the same rule: the second coordinate is twice the first.

\[ (1, 2), \quad (3, 6), \quad (5, 10) \qquad \text{all on the line } y = 2x \]

Predict first

What does the area determinant give for three points that lie on one straight line?

  • Zero
  • A negative number
  • Half the distance between the two outer points
  • It cannot be computed

Correct: Zero

Why: Three points on one line do not enclose anything, so the area has to be zero, and the determinant obliges: the downward products give 10 plus 10 plus 18, which is 38, and the upward products give 30 plus 6 plus 2, which is also 38. The difference is zero, so half of it is zero. The calculation is perfectly possible; it simply reports a flat triangle.

So a zero here means the three points are collinear, which is the same message the zero determinant of Part 2 carried: the shape has collapsed and there is no area left.

That makes the formula useful in a direction the book does not stress. Given three points, one determinant tells you whether they lie on a line, without finding any slopes at all.

Surveyors care about exactly this. Three readings that should outline a plot but produce an area of zero are three readings taken along one line, which means one of them is wrong.

Notice how the two collapses match. Two arrows on one line span no area; three points on one line enclose no area. Same picture, one dimension apart.

40. Cramer's rule

Section

Part 5

41. Each variable is one determinant divided by another

Concept

Determinants can solve systems. The method is named after the Swiss mathematician Gabriel Cramer, and it starts by stripping a system down to the matrix of its coefficients.

\[ \begin{array}{ll} ax + by = e \\ cx + dy = f \end{array} \qquad A = \begin{bmatrix} a & b \\ c & d \end{bmatrix} \]

coefficient matrix — The matrix of the coefficients of the variables in a system, in the order the variables appear, with one row per equation and the constants left out.

The constants on the right of the equals signs are deliberately not in that matrix. They have a separate job: each one takes its turn as a replacement column.

\[ x = \dfrac{\begin{vmatrix} e & b \\ f & d \end{vmatrix}}{\det A} \qquad\qquad y = \dfrac{\begin{vmatrix} a & e \\ c & f \end{vmatrix}}{\det A} \]

Read the two numerators carefully and the pattern is unmistakable. To find x, the column of x coefficients is thrown out and the constants are put in its place; to find y, the column of y coefficients goes instead.

The denominator never changes. You compute the determinant of the coefficient matrix once, and it divides both numerators, which is why it is the first thing you should work out and the thing that decides whether to continue.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — Cramer's Rule for a 2 by 2 System

42. The same matrix, one column swapped

Picture it

Here is the rule drawn out on the system from Example 3, which has coefficients 9 and 4 in the first equation and 3 and negative 5 in the second, with constants negative 6 and negative 21.

Figure (svg): Three determinants side by side: the coefficient matrix with rows 9, 4 and 3, negative 5 giving negative 57; the same matrix with its first column replaced by the constants negative 6 and negative 21, that column highlighted; and the matrix with its second column replaced instead.

Every determinant here is built from the same system. Only one column moves, and which column moves decides which variable you get.

Three determinants, and all three are built from the same eight numbers. The leftmost is the coefficient matrix and the divisor; the other two are it with one column swapped for the constants.

The rule for which column is the only thing to remember: the column belonging to the variable you want is the column that leaves. For x, that is the first column, and the constants move in.

Notice that the column you are not replacing stays exactly where it was, in the same order. Cramer's rule never rearranges a matrix; it substitutes one column and leaves everything else alone.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the note on replacing the coefficients with the column of constants

43. Reading the rule for x piece by piece

Notation

One formula, four things worth naming. The condition at the end is part of the rule, not a footnote to it.

Annotate

On: \( x = \dfrac{\begin{vmatrix} e & b \\ f & d \end{vmatrix}}{\det A}, \quad \det A \ne 0 \)

  • The constants from the right-hand sides of the two equations, in equation order. They have taken the place of the x coefficients, because x is the variable being found.
  • The y coefficients, untouched. Only one column ever moves, and it is the column of the variable you are solving for.
  • The determinant of the coefficient matrix, computed once and used as the divisor for every variable in the system.
  • The rule only applies when det A is not zero. That is not caution about arithmetic: a zero determinant means the system has no single solution for the rule to find, so there is nothing to compute.

Writing the condition down every time is a habit worth having. It puts the determinant of the coefficient matrix first in your working, which is where it belongs.

The formula for y has the identical shape with the other column replaced, so there is really only one thing to learn here and then one choice to make.

Cramer's rule is at its best on exactly this size of problem: a 2 by 2 system where you want both answers, with no rearranging and no substituting. Its cost grows quickly with size, which is why Lesson 3.2's methods do not go away.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the condition det A is not zero

44. Example 3: solving a 2 by 2 system by Cramer's rule

Worked example

The book's worked system, in full. Notice that the very first move is the determinant of the coefficient matrix, before any thought is given to x or y.

\[ \begin{array}{l} 9x + 4y = -6 \\ 3x - 5y = -21 \end{array} \]

Evaluate the determinant of the coefficient matrix

Why: 9 times negative 5 is negative 45, and 3 times 4 is 12, which is subtracted.

\[ \det A = \begin{vmatrix} 9 & 4 \\ 3 & -5 \end{vmatrix} = -45 - 12 = -57 \]

That number is not zero, so the system has exactly one solution and the rule may be applied. Had it been zero, the correct next step would have been to stop and reach for elimination.

Replace the first column with the constants and divide

Why: the x coefficients leave, the constants negative 6 and negative 21 move in, and the result is divided by negative 57.

\[ x = \dfrac{\begin{vmatrix} -6 & 4 \\ -21 & -5 \end{vmatrix}}{-57} = \dfrac{30 - (-84)}{-57} = \dfrac{114}{-57} = -2 \]

Watch the double negative in that numerator. The main diagonal gives 30, the other diagonal gives negative 84, and subtracting negative 84 adds it, which is where the 114 comes from.

Replace the second column with the constants and divide

Why: this time the y coefficients leave and the same divisor is used again.

\[ y = \dfrac{\begin{vmatrix} 9 & -6 \\ 3 & -21 \end{vmatrix}}{-57} = \dfrac{-189 - (-18)}{-57} = \dfrac{-171}{-57} = 3 \]

Figure (svg): The lines nine x plus four y equals negative six and three x minus five y equals negative twenty-one drawn on axes, crossing at the point negative 2 comma 3.

Cramer's rule never draws this picture, but it is the picture being computed: the one point that lies on both lines. The determinant negative 57 is the guarantee that there is exactly one.

The point marked on the graph is the answer: negative 2 and 3. Cramer's rule never drew this picture, but the single crossing it shows is what a nonzero determinant guarantees.

Check the pair in both original equations

Why: substituting negative 2 and 3 into the first equation gives negative 18 plus 12, which is negative 6, and into the second gives negative 6 minus 15, which is negative 21. Both check, so the solution is right.

\[ 9(-2) + 4(3) = -6 \qquad 3(-2) - 5(3) = -21 \]

Check in both equations, never just one. A single arithmetic slip in a numerator will usually still satisfy one of the two equations by luck, and it will almost never satisfy both.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — Example 3, including the book's own check

45. Guided Practice 6: finish the rule

Fill the middle

Another system from page 207, with the coefficient determinant already computed. Fill in the numerator and the value of x.

\[ \begin{array}{l} 4x + 7y = 2 \\ -3x - 2y = -8 \end{array} \qquad \det A = \begin{vmatrix} 4 & 7 \\ -3 & -2 \end{vmatrix} = -8 - (-21) = 13 \]

Fill in the blanks

x = \dfrac52 2 & 7 \\ -8 & -2 \end4}___ = \dfrac______ = \dfrac___}___ = ___

Why: The main diagonal of the numerator gives 2 times negative 2, which is negative 4, and the other diagonal gives negative 8 times 7, which is negative 56. Subtracting negative 56 adds it, so the numerator is 52, and 52 divided by 13 is 4. The matching value of y works out as negative 2, and the pair 4 and negative 2 satisfies both equations.

Both diagonal products were negative here, and the subtraction turned the second one into an addition. That is the single most error-prone line in the whole method, so write the brackets.

Finish the job for yourself: replace the second column instead, and you should get negative 26 over 13, which is negative 2. Then check the pair 4 and negative 2 in both equations.

Notice that the divisor 13 was computed once and reused. On a 2 by 2 system Cramer's rule asks for three determinants in total, never four.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 207 — Guided Practice 6

46. Guided Practice 5: a system whose determinant is positive

Worked example

One more of the book's checkpoints, this time with a positive determinant, so there is one fewer sign to keep track of. The routine is unchanged.

\[ \begin{array}{l} 3x - 4y = -15 \\ 2x + 5y = 13 \end{array} \]

Evaluate the coefficient determinant first

Why: 3 times 5 is 15, and 2 times negative 4 is negative 8, which is subtracted, so the two amounts add.

\[ \det A = \begin{vmatrix} 3 & -4 \\ 2 & 5 \end{vmatrix} = 15 - (-8) = 23 \]

Twenty-three is not zero, so there is a unique solution to find. An odd-looking divisor like this is a hint that the answers will come out as whole numbers only if the numerators are multiples of it, which in a textbook exercise they always are.

Replace the first column for x

Why: the constants negative 15 and 13 take the place of the x coefficients.

\[ x = \dfrac{\begin{vmatrix} -15 & -4 \\ 13 & 5 \end{vmatrix}}{23} = \dfrac{-75 - (-52)}{23} = \dfrac{-23}{23} = -1 \]

Replace the second column for y

Why: the same divisor, the other column swapped.

\[ y = \dfrac{\begin{vmatrix} 3 & -15 \\ 2 & 13 \end{vmatrix}}{23} = \dfrac{39 - (-30)}{23} = \dfrac{69}{23} = 3 \]

Both numerators turned out to be exact multiples of 23, which is the textbook telling you that you have not slipped. Ugly fractions in a problem like this are usually a symptom rather than an answer.

Check the pair in both equations

Why: in the first, 3 times negative 1 minus 4 times 3 is negative 3 minus 12, which is negative 15; in the second, 2 times negative 1 plus 5 times 3 is negative 2 plus 15, which is 13. Both hold, so the solution is negative 1 and 3.

\[ 3(-1) - 4(3) = -15 \qquad 2(-1) + 5(3) = 13 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 207 — Guided Practice 5

47. Find the error: the wrong column was replaced

Error analysis

Here is a student's attempt at Example 3's system. Every determinant on the page has been evaluated correctly, and the answer is still wrong. Work out what happened before reading the notes.

\[ \begin{array}{l} 9x + 4y = -6 \\ 3x - 5y = -21 \end{array} \qquad \det A = -57 \]

Annotate

On: \( x = \dfrac{\begin{vmatrix} 9 & -6 \\ 3 & -21 \end{vmatrix}}{-57} = \dfrac{-171}{-57} = 3 \)

  • This value is right. The matrix with rows 9, negative 6 and 3, negative 21 really does have determinant negative 171, and dividing by negative 57 really does give 3.
  • Here is the error. The constants have gone into the second column, which is the y column, so this ratio is the value of y, not x.
  • The student will report x equal to 3 and y equal to negative 2, which is the correct solution with the two numbers swapped. Both equations then fail: 9 times 3 plus 4 times negative 2 is 19, not negative 6.
  • For x, the constants replace the first column, giving 114 over negative 57, which is negative 2. The solution is negative 2 and 3, in that order, which is what the check confirms.

\[ x = \dfrac{\begin{vmatrix} -6 & 4 \\ -21 & -5 \end{vmatrix}}{-57} = \dfrac{114}{-57} = -2 \]

This error is dangerous precisely because it produces sensible-looking numbers. Nothing in the arithmetic complains, and the answers are even the right two numbers, just attached to the wrong letters.

The check is what saves you, and only a check in both equations will do it. Swapped answers usually fail both, loudly.

A wording that prevents the mistake: the variable I want is the column that leaves. Say it out loud before you write the numerator and the columns stop swapping themselves.

48. Match each determinant to its job

Matching

Four determinants from one system, each with a different role. Pair them up by what they are for rather than by how they look.

Match the pairs

  • a. the determinant of the coefficient matrix
  • b. the coefficient matrix with the first column replaced by the constants
  • c. the coefficient matrix with the second column replaced by the constants
  • d. a coefficient determinant that comes out as zero
  • w. the divisor for every variable in the system
  • x. the numerator for the first variable
  • y. the numerator for the second variable
  • z. the signal to stop and use elimination instead

Why: The coefficient determinant is computed once and divides everything. Replacing a column with the constants builds the numerator for the variable whose column left, so the first column gives the first variable and the second gives the second. A zero coefficient determinant is not a numerator or a divisor at all; it is the news that the system has no unique solution and that the rule does not apply.

The pairing to be careful with is the middle two, because the two matrices differ by only one column and the difference is the entire meaning.

The last pairing is the one worth memorising. Zero is not a value you divide by carefully; it is a value that ends the attempt.

Notice that three determinants answer a 2 by 2 system completely. That is the efficiency Cramer's rule is offering: no rearranging, no substitution, no elimination, just three small calculations and two divisions.

49. Exercise 42: the rock concert ticket sales

Worked example

A concert drew 6700 people. Floor seats cost 40 dollars and every other seat cost 25, and the ticket sales came to 185,500 dollars. How many of each kind of seat were sold?

Nothing in that paragraph is a matrix yet. The first job is always to turn the sentences into two equations, and only then to look at the coefficients.

Write a system from the two facts

Why: let x be the number of floor seats and y the number of other seats. One equation counts people, the other counts money.

\[ \begin{array}{l} x + y = 6700 \\ 40x + 25y = 185\,500 \end{array} \]

Evaluate the coefficient determinant

Why: 1 times 25 is 25, and 40 times 1 is 40, which is subtracted.

\[ \det A = \begin{vmatrix} 1 & 1 \\ 40 & 25 \end{vmatrix} = 25 - 40 = -15 \]

Negative 15 is not zero, so there is exactly one seating split that fits both facts. That is worth knowing before you start dividing large numbers.

Replace the first column for the floor seats

Why: the constants 6700 and 185,500 move into the x column.

\[ x = \dfrac{\begin{vmatrix} 6700 & 1 \\ 185\,500 & 25 \end{vmatrix}}{-15} = \dfrac{167\,500 - 185\,500}{-15} = \dfrac{-18\,000}{-15} = 1200 \]

Replace the second column for the other seats

Why: same divisor, other column.

\[ y = \dfrac{\begin{vmatrix} 1 & 6700 \\ 40 & 185\,500 \end{vmatrix}}{-15} = \dfrac{185\,500 - 268\,000}{-15} = \dfrac{-82\,500}{-15} = 5500 \]

Figure (svg): Two lines on axes counting floor seats against other seats: one for a total attendance of 6700 people and one for a total income of 185,500 dollars, crossing at 1200 floor seats and 5500 other seats.

One line counts heads and the other counts money. The crossing is the only seating split that satisfies both, and a determinant of negative 15 promises it is unique.

Check both facts

Why: 1200 and 5500 add to 6700 people, and 40 times 1200 plus 25 times 5500 is 48,000 plus 137,500, which is 185,500 dollars. Both conditions hold, so the split is right.

\[ 1200 + 5500 = 6700 \qquad 40(1200) + 25(5500) = 185\,500 \]

The exercise also asks you to solve this by substitution and by elimination and to say which you prefer. With both answers wanted and no awkward rearranging, three small determinants is hard to beat.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 42

50. Cramer's rule against the methods from Lesson 3.2

Comparison

Cramer's rule does not replace what you already know. Set the three methods side by side and the question becomes which one suits the system in front of you.

Comparison matrix

methodwhat you compute firstat its best whenat its worst when
graphingtwo linesyou want the picturethe answer is not a whole number
substitutionone variablea coefficient is already 1every coefficient is awkward
eliminationa matching pair of coefficientsone variable cancels easilyyou want to know in advance whether it works
Cramer's ruledet Ayou want both answers and a unique-solution testthe system is large

The row that makes the case for this lesson is the last one. Cramer's rule is the only method that tells you whether a unique solution exists before it tries to find one.

The worst-case column matters too. A 3 by 3 system needs four 3 by 3 determinants, which is twenty-four products, and beyond that the arithmetic grows faster than elimination does.

Elimination keeps one clear advantage: when the determinant is zero, elimination is what tells you which of the two bad cases you are in. The two methods work together rather than competing.

OpenStax Algebra and Trigonometry 2e, §11.8 Solving Systems with Cramer's Rule §11.8 — the same comparison of methods

51. Example 4: three atomic weights from three compounds

Worked example

A chemistry problem the book solves with Cramer's rule. Glucose weighs 180, carbon dioxide 44 and hydrogen peroxide 34, and their formulas tell you how many atoms of each element they contain.

Turn each formula into an equation

Why: let C, H and O be the three atomic weights. Glucose has six carbons, twelve hydrogens and six oxygens, carbon dioxide has one carbon and two oxygens, and hydrogen peroxide has two hydrogens and two oxygens.

\[ \begin{array}{l} 6C + 12H + 6O = 180 \\ C + 2O = 44 \\ 2H + 2O = 34 \end{array} \]

Notice the missing terms. Carbon dioxide has no hydrogen, so its hydrogen coefficient is zero, and that zero must go into the matrix. A coefficient of zero is information, not an absence.

Evaluate the coefficient determinant

Why: six products, three down and three up, with the zeros doing most of the work: the downward products are 0, 0 and 12, and the upward products are 0, 24 and 24.

\[ \det A = \begin{vmatrix} 6 & 12 & 6 \\ 1 & 0 & 2 \\ 0 & 2 & 2 \end{vmatrix} = 12 - 48 = -36 \]

Not zero, so the three compounds between them pin the three weights down to exactly one set of values. Three equations in three unknowns do not always manage that, and the determinant is what confirms it.

Replace each column with the weights in turn

Why: the same rule as for a 2 by 2, one column wider: the column of the variable you want leaves and the constants 180, 44 and 34 move in.

\[ C = \dfrac{-432}{-36}, \qquad H = \dfrac{-36}{-36}, \qquad O = \dfrac{-576}{-36} \]

Figure (svg): Four 3 by 3 determinants in a row: the coefficient matrix of the chemistry system giving negative 36, then the same matrix with the atomic weights 180, 44 and 34 replacing the first, second and third columns in turn, giving negative 432, negative 36 and negative 576.

The same pattern as the 2 by 2 rule, one column wider. The coefficient determinant is computed once and then divides all three numerators.

Divide, using the same determinant three times

Why: each numerator is a 3 by 3 determinant of its own, and all three are divided by negative 36.

\[ C = 12, \qquad H = 1, \qquad O = 16 \]

Check all three weights in the original equations

Why: six carbons at 12, twelve hydrogens at 1 and six oxygens at 16 give 72 plus 12 plus 96, which is 180; one carbon and two oxygens give 12 plus 32, which is 44; two hydrogens and two oxygens give 2 plus 32, which is 34. All three hold, and the answers match the periodic table.

\[ 6(12) + 12(1) + 6(16) = 180 \qquad 12 + 2(16) = 44 \qquad 2(1) + 2(16) = 34 \]

Getting carbon at 12, hydrogen at 1 and oxygen at 16 out of three molecular weights is a genuinely satisfying check, because the answer was verifiable from outside the problem all along.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 206 — Example 4

52. Before you solve: set up the 3 by 3 rule

Step zero

Here is Guided Practice 7, a three-variable system. Do not solve it yet. The setting up is where 3 by 3 Cramer goes wrong, so practise only that.

\[ \begin{array}{l} 3x - 4y + 2z = 18 \\ 4x + y - 5z = -13 \\ 2x - 3y + z = 11 \end{array} \]

Discussion prompt

Write down the coefficient matrix, then write the determinant whose value is the numerator for z, and say which determinant you would compute first and why.

The two things that go wrong at this stage are both bookkeeping. Coefficients must go in variable order, and a variable missing from an equation contributes a zero rather than nothing at all.

The order of the constants matters just as much. They keep their equation order when they move into the column, so the constant from the first equation goes in the first row.

Once the four determinants are set up correctly, the rest is the six-product recipe four times over and three divisions. Long, but not difficult.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 206 — Cramer's Rule for a 3 by 3 System

53. What a zero determinant means

Section

Part 6

54. Zero says no unique solution, and it says nothing more

Concept

Everything in this part follows from one line of the Key Concept box: the rule applies when the determinant of the coefficient matrix is not zero. Here is what the zero case actually is.

\[ \begin{array}{l} 2x + 3y = 7 \\ 4x + 6y = 5 \end{array} \qquad \det A = \begin{vmatrix} 2 & 3 \\ 4 & 6 \end{vmatrix} = 12 - 12 = 0 \]

Figure (svg): The lines two x plus three y equals seven and four x plus six y equals five drawn on axes, running parallel and never meeting.

Both lines fall two units for every three across, so they can only be the same line or no meeting at all. The coefficient determinant is zero in either case.

Two things have happened at once, and they are the same thing seen twice. The coefficient rows are proportional, so the arrows of Part 2 lie on one line; and the two equations describe lines of the same slope, so they cannot cross exactly once.

That is why Cramer's rule has to refuse. There is no single pair of numbers for the formula to produce, and the arithmetic reports this by asking you to divide by zero, which is not a difficulty to work around but a message.

Be careful about what zero does not tell you. It says there is no unique solution; it does not say whether there are none or infinitely many. Those two cases have the same determinant and different answers.

So a zero determinant ends one method and starts another. Elimination or substitution on the original equations is what separates the parallel case from the coincident one.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the condition on Cramer's rule

55. A zero determinant with no solution at all

Worked example

Take the system from the warm-up and push it all the way through. The determinant stops Cramer's rule, and elimination then says which of the two bad cases you are in.

\[ \begin{array}{l} 2x + 3y = 7 \\ 4x + 6y = 5 \end{array} \]

Evaluate the coefficient determinant first

Why: 2 times 6 is 12, and 4 times 3 is 12, which is subtracted, so the two products cancel exactly.

\[ \det A = 12 - 12 = 0 \]

Stop here as far as Cramer's rule is concerned. Both numerators would have to be divided by zero, so neither x nor y can be produced this way, however carefully you compute.

Multiply the first equation by 2 to compare the two

Why: the coefficients then match the second equation exactly, which is what proportional rows guarantee.

\[ 2(2x + 3y) = 2(7) \;\Longrightarrow\; 4x + 6y = 14 \]

Compare the two constants

Why: the same left-hand side is being set equal to two different numbers, 14 and 5, which nothing can satisfy.

\[ 4x + 6y = 14 \quad \text{and} \quad 4x + 6y = 5 \]

So the system is inconsistent: it has no solution. In the picture, the lines are parallel and separate, which is exactly the case Lesson 3.1 called inconsistent.

Check by substituting any point of the first line

Why: the point 2, 1 lies on the first line because 4 plus 3 is 7, and in the second equation it gives 8 plus 6, which is 14 rather than 5. Every point of the first line misses the second in the same way, by the same 9, which is what parallel means.

\[ 2(2) + 3(1) = 7 \qquad 4(2) + 6(1) = 14 \ne 5 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the case det A equals zero

56. A zero determinant with infinitely many solutions

Worked example

Now a system with the same zero determinant and a completely different answer. This is the pair of cases the determinant cannot tell apart.

\[ \begin{array}{l} 4x - 3y = 8 \\ 8x - 6y = 16 \end{array} \]

Evaluate the coefficient determinant

Why: 4 times negative 6 is negative 24, and 8 times negative 3 is negative 24, which is subtracted.

\[ \det A = -24 - (-24) = 0 \]

Zero again, so Cramer's rule is out again. Notice how little the determinant has told you: exactly the same value as the previous system, whose answer was completely different.

Multiply the first equation by 2

Why: and compare it with the second.

\[ 2(4x - 3y) = 2(8) \;\Longrightarrow\; 8x - 6y = 16 \]

Compare the two equations

Why: this time the constants agree as well, so the second equation is the first one written twice as large and adds no new information.

\[ 8x - 6y = 16 \quad \text{is the first equation doubled} \]

So this system is consistent and dependent: the two lines are the same line, and every point on it solves both equations. There are infinitely many solutions, not none.

Check two different points on that line

Why: the point 2, 0 gives 8 in the first equation and 16 in the second, and the point 5, 4 gives 20 minus 12, which is 8, and 40 minus 24, which is 16. Two different solutions is already proof that the solution is not unique.

\[ 4(2) - 3(0) = 8 \qquad 4(5) - 3(4) = 8 \]

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.1, p. 158 — consistent and dependent systems

57. Trap: dividing by a zero determinant

Trap

The trap

On the system with rows 2, 3 and 4, 6, a student presses on:

\[ \det A = 0 \]

\[ x = \dfrac{27}{0} = 0 \]

The reasoning is that dividing something by nothing gives nothing, so x must be zero.

The fix

Dividing by zero does not give zero. It gives nothing at all: the expression has no value.

\[ \dfrac{27}{0} \quad \text{is undefined} \]

And the answer is not hiding somewhere else either. There is no unique solution to find, so no arithmetic performed on these four coefficients can produce one.

The correct response is to stop and change method. Eliminate, and the system will tell you whether it has no solutions or infinitely many.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the condition det A is not zero

58. Does Cramer's rule apply?

Sorting

Six systems. For each one, compute the determinant of the coefficient matrix in your head if you can, and sort by whether the rule can be used at all.

Sort into buckets

Sort each system by whether Cramer's rule can solve it.

The rule applies
9x + 4y = −6 and 3x − 5y = −21; 2x − y = −2 and x + 2y = 14; 5x + y = −40 and 2x − 5y = 11; x + y = 6700 and 40x + 25y = 185,500
The determinant is zero, so it does not
2x + 3y = 7 and 4x + 6y = 5; 4x − 3y = 8 and 8x − 6y = 16
yes
These four have coefficient determinants of negative 57, 5, negative 27 and negative 15. None is zero, so each system has exactly one solution and the rule produces it.
no
In both of these the second equation's coefficients are exactly twice the first equation's, so the determinant is zero. One of them turns out to have no solution and the other infinitely many, which is why the determinant alone cannot finish the job.

The quick test is the one from Part 1: look for one row of coefficients being a multiple of the other. Both of the rejected systems were doubled, and you can see that without multiplying anything.

Notice that the two rejected systems look alike and behave differently. Same test, same verdict from the determinant, two different final answers, which is exactly the limit of what this one number can do.

The habit being drilled here is putting the determinant first. Sorting these six took a few seconds and would have saved a lot of pointless dividing.

59. What happens just before the determinant reaches zero

Edge cases

Zero is a wall, but it is worth looking at the approach to it. Keep the first equation fixed and slide one coefficient of the second towards the value that makes the determinant vanish.

\[ \begin{array}{l} 2x + 3y = 7 \\ 4x + ky = 5 \end{array} \qquad \det A = 2k - 12 \]

kdet Asolution
5−2x = −10, y = 9
5.9−0.2x = −131.5, y = 90
5.99−0.02x = −1346.5, y = 900
60no unique solution

Discussion prompt

Describe what happens to the solution as the determinant gets closer to zero, and explain what that behaviour looks like in terms of the two lines.

This is why a determinant close to zero is a warning even though it is not zero. The answer exists, but it is enormous and extremely sensitive to the coefficients, so a small measurement error in the data moves it a long way.

It also explains the sliver in Part 2 in a new light. A thin parallelogram means nearly proportional rows, which means nearly parallel lines, which means a crossing point that is hard to pin down.

Nothing in this table is a new rule. It is the same division by the same determinant, watched as the divisor gets small, which is behaviour worth recognising anywhere in mathematics.

60. Three false claims about zero determinants

Elimination

Four statements about a 2 by 2 system whose coefficient determinant is zero. Exactly one of them is true, so knock out the other three and say what is wrong with each.

Eliminate the wrong options

A system of two linear equations has a coefficient determinant of zero. Which statement about it is true?

  • A. There is no unique solution, but this alone does not say whether there are none or infinitely many
  • B. The system has no solution at all, so its graphs must be parallel lines
  • C. The coefficient matrix must contain a row of zeros
  • D. Cramer's rule still works and gives zero for both variables

Survives elimination: A

Why: A zero determinant means the coefficient rows are proportional, so the two lines have the same slope and cannot meet exactly once. Whether they never meet or coincide completely depends on the constants, which the coefficient determinant never looked at, so elimination is still needed to decide. The other three claims each overstate what the determinant knows.

The claim about a row of zeros is worth dwelling on. Proportional rows are the condition, and a row of zeros is only the most extreme way of satisfying it.

The claim that the rule gives zero is the dangerous one, because it produces an answer that can be written down. Dividing by zero has no value, and a solution of zero and zero would be checkable in the original equations, which it fails.

Notice that the surviving statement is deliberately modest. It names what the determinant knows and stops, which is the right way to use a test that answers one question and not another.

61. Looking ahead to inverses

Prediction

Lesson 3.8 will introduce the inverse of a matrix, which is the matrix that undoes it, in the way that dividing by 5 undoes multiplying by 5. Not every matrix has one, and you can already guess the condition.

\[ A \begin{bmatrix} x \\ y \end{bmatrix} = \begin{bmatrix} e \\ f \end{bmatrix} \qquad \text{solved by} \qquad \begin{bmatrix} x \\ y \end{bmatrix} = A^{-1}\begin{bmatrix} e \\ f \end{bmatrix} \]

Predict first

A system can be written as a matrix times a column of variables. Which matrices do you expect to have an inverse?

  • Exactly those whose determinant is not zero
  • All square matrices
  • Only matrices whose entries are all positive
  • Only matrices with no zero entries

Correct: Exactly those whose determinant is not zero

Why: An inverse would solve the system in one multiplication, so it would hand you a single solution for every column of constants. A zero determinant means some systems with that coefficient matrix have no single solution at all, so no inverse can exist there. A nonzero determinant is exactly the condition, and Lesson 3.8 proves it by putting the determinant in the denominator of the inverse.

The reasoning matters more than the guess. If a matrix had an inverse, multiplying by it would produce one and only one solution for any constants, and Part 6 has just shown that a zero determinant makes that impossible.

So the two conditions you will meet in Lesson 3.8, has an inverse and has a nonzero determinant, are not two facts to learn. They are the same fact, and you have already done the work that shows why.

You will also see the determinant reappear in the formula for the inverse itself, sitting underneath everything as a divisor, in exactly the place it occupied in Cramer's rule.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 205 — the determinant as the divisor in Cramer's rule

62. Putting it together

Section

Part 7

63. The order of operations for any determinant question

Pattern

Every question in this lesson is answered in the same order, and the order matters because the first step decides whether the others are possible.

Figure (svg): A flow diagram: a system of equations leads to a box that computes the determinant of the coefficient matrix, which branches on whether that number is zero. Zero leads to a box saying no unique solution, use elimination to tell which case; not zero leads to a box saying swap one column for the constants and divide.

The determinant is the gate, not the last step. Computing it first is what stops you dividing by zero halfway through a solution.
  1. Compute the determinant of the coefficient matrix first. Two products for a 2 by 2, six for a 3 by 3.
  2. If it is zero, stop dividing. There is no unique solution; eliminate instead, to find whether there are none or infinitely many.
  3. If it is not zero, swap one column for the constants, one variable at a time, and divide each numerator by that same determinant.
  4. For an area, put the vertices in rows with a column of ones, take half the determinant, and choose the sign that makes the answer positive.
  5. Check in the original equations, both of them, or against a rough area you can estimate.

The first two lines are the ones worth building into a habit. Most of the wasted effort on this topic is spent computing numerators for a system that was never going to have a unique solution.

The area line is the same recipe wearing different clothes, and the plus or minus in it is the sign that the determinant carried all along, which an area is allowed to discard.

Nothing in this list needs the picture from Part 2, but everything in it makes more sense with the picture. Zero determinant, collapsed parallelogram, no unique solution: one idea in three languages.

64. Check: which determinant is the largest?

Check

Four matrices, and the question is which has the greatest determinant. Compute all four rather than guessing from the size of the entries, and watch the signs.

Check your understanding

Which of these matrices has the greatest determinant?

  • A. [[-4, 1], [6, 3]]
  • B. [[1, 6], [3, 8]]
  • C. [[5, -3], [7, -1]]
  • D. [[5, -2], [1, 5]] (correct)

Answer: D

Why: The four determinants are negative 18, negative 10, 16 and 27. The largest is 27, from the matrix with rows 5 and negative 2, then 1 and 5, where the main diagonal gives 25 and subtracting negative 2 adds a further 2.

Why A tempts people
The main diagonal gives negative 12 and the other diagonal gives 6, so this determinant is negative 18, the smallest of the four.
Why B tempts people
Eight minus eighteen is negative 10, so this one is negative as well.
Why C tempts people
Five times negative 1 is negative 5, and subtracting 7 times negative 3 adds 21, giving 16, which is positive but not the largest.

Two of the four come out negative, and a negative determinant is never the greatest when a positive one is available. Getting the sign right is most of this question.

The winner is the one where both diagonal products help: the main diagonal gives 25, and subtracting a negative product adds a further 2 instead of taking anything away.

Nothing about the size of the entries predicted this. The matrix with the largest single entry, an 8, has a negative determinant, because the two products nearly cancel.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 207 — Exercise 21

65. Check: which numerator finds y?

Check

The system is here and the coefficient determinant is already computed. The whole question is which matrix goes on top when you want y.

\[ \begin{array}{l} 5x + y = -40 \\ 2x - 5y = 11 \end{array} \qquad \det A = -25 - 2 = -27 \]

Check your understanding

For this system, which determinant divided by negative 27 gives y?

  • A. the determinant of [[-40, 1], [11, -5]]
  • B. the determinant of [[5, -40], [2, 11]] (correct)
  • C. the determinant of [[5, 1], [2, -5]]
  • D. the determinant of [[-40, 11], [1, -5]]

Answer: B

Why: To find y, the column of y coefficients leaves and the constants take its place, keeping their equation order. That puts negative 40 above 11 in the second column, giving 55 plus 80, which is 135, and 135 divided by negative 27 is negative 5. The matching value of x is negative 7, and the pair checks in both equations.

Why A tempts people
This has the constants in the first column, so it is the numerator for x, not y.
Why C tempts people
This is the coefficient matrix itself, which is the divisor rather than a numerator.
Why D tempts people
The constants have been written across a row instead of down a column, which mixes up the two equations.

The rule to say to yourself is that the variable you want is the column that leaves. For y, the y coefficients go, and nothing else in the matrix moves.

The last distractor is worth looking at twice. Writing the constants along a row instead of down a column is a real slip, and it silently mixes the two equations together.

Finish the problem for practice: the numerator for x is 200 minus 11, which is 189, and 189 divided by negative 27 is negative 7. Check the pair negative 7 and negative 5 in both equations.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 31

66. Check: what does a zero determinant tell you?

Check

Last one, and it is the idea this deck was built around. A system arrives with a coefficient determinant of zero, and you must say precisely what follows.

\[ \begin{array}{l} 3x - 6y = 9 \\ -x + 2y = 4 \end{array} \qquad \det A = 6 - 6 = 0 \]

Check your understanding

What can you conclude from the determinant of this system?

  • A. There is no unique solution; eliminate to find whether there are none or infinitely many (correct)
  • B. The solution is x equal to zero and y equal to zero
  • C. The two lines cross at exactly one point, which Cramer's rule will find
  • D. The system must have infinitely many solutions

Answer: A

Why: A zero determinant means the coefficient rows are proportional, so the lines have the same slope and cannot cross exactly once. Which of the two remaining cases holds depends on the constants: multiplying the second equation by negative 3 gives 3x minus 6y equal to negative 12, against 9 in the first, so this particular system has no solution at all. The determinant could not have told you that on its own.

Why B tempts people
Zero is the value of the determinant, not of the variables, and dividing by it produces no value at all.
Why C tempts people
A unique crossing point is exactly what a zero determinant rules out.
Why D tempts people
Infinitely many is one of the two possibilities, but the constants decide, and here they give no solution rather than infinitely many.

The determinant of this system is zero because the first row of coefficients is negative 3 times the second. That test is faster than computing the two products, and it is the same test you used in Part 1.

The reason the answer stops short of naming a case is the whole lesson in one line. The coefficient determinant never looked at the constants, so it cannot possibly know which side of that distinction the system falls.

Work the elimination through yourself to see it: multiplying the second equation by negative 3 gives 3x minus 6y equal to negative 12, while the first says 9. Two different values for the same expression, so nothing satisfies both.

67. Say the zero case in your own words

Explain it to yourself

One idea has carried this deck, and putting it into your own sentences is what makes it stay. Write your version before reading the one below.

Discussion prompt

In three or four sentences, explain what a determinant of zero means about the two rows of a matrix, about the parallelogram they span, and about the system whose coefficients they are. Then say why Cramer's rule has to fail there.

If your explanation used the word formula more than the word picture, try writing it again. The picture is what makes the condition memorable a month from now.

A good test of your own understanding: could you explain to somebody who has not met matrices why two proportional equations cannot pin down two unknowns? If yes, you have the idea.

And notice how far that one idea reaches. It explained the collapse in Part 2, the collinear points in Part 4, the refusal in Part 5, and it will explain which matrices have inverses in Lesson 3.8.

68. Exit ticket

Exit ticket

One system, three questions, and all three habits: determinant first, then the column swap, then a check in both equations.

\[ \begin{array}{l} 2x - y = -2 \\ x + 2y = 14 \end{array} \]

Discussion prompt

Compute the determinant of the coefficient matrix and say what it permits. Then use Cramer's rule to find x and y, and check the pair in both original equations.

Every subtraction in that solution involved taking away a negative number. Writing the brackets in first, before simplifying, is what keeps all three of them right.

Notice how the determinant earned its place at the front. It cost two products and it told you the rest of the work would succeed.

If you want one more, exercise 29 has a determinant of negative 4 and an answer of 11 and negative 6, which is a good test of whether large answers worry you. They should not: only a zero determinant is a problem.

McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule §3.7, p. 208 — Exercise 30

69. What you can do now

Recap

Two recipes, one meaning, and one test. The recipes are diagonal products; the meaning is area; the test is whether the answer is zero.

what you havewhat you dowhat it tells you
a 2 by 2 matrixmain diagonal product minus the otherthe signed area of the parallelogram its rows span
a 3 by 3 matrixrepeat two columns, add three products, subtract threethe signed volume of the box its rows span
three verticesrows with a column of ones, then half the determinantthe area of the triangle, sign discarded
a system with a nonzero determinantswap one column for the constants, dividethe one solution, both variables
a system with a zero determinantstop and eliminateno unique solution: none, or infinitely many

\[ \begin{vmatrix} a & b \\ c & d \end{vmatrix} = ad - cb \qquad x = \dfrac{\begin{vmatrix} e & b \\ f & d \end{vmatrix}}{\det A}, \quad \det A \ne 0 \]

Read the right-hand column of that table from top to bottom and you have the through-line of the lesson. Area, area, area, and then a test that is really the question of whether the area is zero.

The last row is the one to keep. A zero determinant is not a failure of your arithmetic and not a hole in the rule; it is the rule detecting that the system it was handed has no single answer.

Lesson 3.8 picks this up immediately with the inverse of a matrix, and the condition for one to exist turns out to be the condition you have been testing all lesson: a determinant that is not zero.

OpenStax Algebra and Trigonometry 2e, §11.8 Solving Systems with Cramer's Rule §11.8 — Cramer's rule, with the same 2 by 2 and 3 by 3 cases

Sources

  1. McDougal Littell Algebra 2 (Texas Edition), Ch. 3 Linear Systems and Matrices — Lesson 3.7 Evaluate Determinants and Apply Cramer's Rule — Larson, Boswell, Kanold & Stiff, McDougal Littell / Houghton Mifflin, 2007, pp. 203-207
  2. OpenStax Algebra and Trigonometry 2e, §11.8 Solving Systems with Cramer's Rule
  3. OpenStax College Algebra 2e, §7.8 Solving Systems with Cramer's Rule

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

Book on Wyzant · Text (657) 465-8108