What the Grader Sees When Gaussian Elimination Misses a Pivot
When you skip swapping rows and miss a pivot in Gaussian elimination, one of your variables gets hidden or lost, so your reduced matrix doesn't reveal the real solution. If your second row never gets a leading 1, x2 disappears from the answer, as it does in the example below.
Suppose you are handed the system of equations:
2x₁ + 4x₂ = 8
x₁ + 2x₂ = 5
You write your work for Gaussian elimination as follows:
Student Solution (Incorrect, as seen by the grader):
- Write the augmented matrix:
[ 2 4 | 8 ]
[ 1 2 | 5 ] - Leave the rows as is (no swap):
- Row 1: 2x₁ + 4x₂ = 8
- Row 2: x₁ + 2x₂ = 5
- Eliminate x₁ in Row 2 by subtracting half of Row 1:
- Row 2 - (1/2)Row 1: (x₁ + 2x₂) - (1/2)(2x₁ + 4x₂) = 5 - (1/2)(8)
- (x₁ - x₁) + (2x₂ - 2x₂) = 5 - 4
- 0 = 1
At this point, you conclude the system has no solution, since 0 = 1 is impossible. But that's not what the grader sees — and not what the math says.
The x₂ That Disappears: Where the Mark Is Lost
A first-year linear algebra course teaches that Gaussian elimination relies on creating leading 1s (pivots) by swapping rows if necessary, as required by the Gaussian elimination algorithm. If you skip this, your elimination can turn a perfectly solvable system into an inconsistent one, or erase a variable entirely (the 'hidden variable' trap). In this example, the grader marks you down because you never swapped Row 1 and Row 2 to put the row with the smaller x₁ coefficient on top, which is required here since Row 2 has a simpler leading coefficient.
Correct Approach (with pivot):
- Swap Row 1 and Row 2 first:
- New Row 1: x₁ + 2x₂ = 5
- New Row 2: 2x₁ + 4x₂ = 8
- Eliminate x₁ from Row 2:
- Row 2 - 2(Row 1): [2x₁ + 4x₂] - 2[x₁ + 2x₂] = 8 - 2·5
- (2x₁ - 2x₁) + (4x₂ - 4x₂) = 8 - 10
- 0 = -2
- But this seems wrong — let's check.
Wait: both methods seem to suggest a contradiction. But let’s check the original equations:
2x₁ + 4x₂ = 8 x₁ + 2x₂ = 5
Multiply the second by 2: 2x₁ + 4x₂ = 10
Compare to the first: 2x₁ + 4x₂ = 8
Those are NOT the same equation — in fact, the system really is inconsistent. The student's conclusion is correct, but the grader still deducts marks if you don't show the required row swap and pivoting, since the process must always allow for the possibility that a nonzero pivot emerges from a swapped row. If you never attempt the swap, you miss the step that would clearly reveal the contradiction.
But let's see a system where missing the pivot truly hides the solution.
When the System Does Have a Solution and the Missed Pivot Destroys It
Change the original system slightly:
2x₁ + 4x₂ = 8 x₁ + 2x₂ = 2
Now:
- Multiply the second by 2: 2x₁ + 4x₂ = 4
- Compare to the first: 2x₁ + 4x₂ = 8
- Now these are parallel lines with no intersection (again, inconsistent). So let’s set up a well-behaved system:
x₁ + 2x₂ = 5 2x₁ + 4x₂ = 8
Let’s see the correct elimination:
- Row 2 - 2(Row 1): (2x₁ + 4x₂) - 2(x₁ + 2x₂) = 8 - 2·5 (2x₁ - 2x₁) + (4x₂ - 4x₂) = 8 - 10 0 = -2
Still inconsistent! Let’s try another concrete system:
x₁ + x₂ = 3 2x₁ + 3x₂ = 7
Correct elimination:
-
Row 2 - 2(Row 1): (2x₁ + 3x₂) - 2(x₁ + x₂) = 7 - 2·3 (2x₁ - 2x₁) + (3x₂ - 2x₂) = 7 - 6 x₂ = 1
-
Back to Row 1: x₁ + 1 = 3 → x₁ = 2
Right answer: x₁ = 2, x₂ = 1
What if you did not get the pivot (forgot to scale to get a 1 in the first entry)?
Suppose you wrote:
- No row swap needed, but you skip dividing by the coefficient:
- Row 1: x₁ + x₂ = 3
- Row 2: 2x₁ + 3x₂ = 7
- Subtract 2×Row 1 from Row 2:
- (2x₁ + 3x₂) - 2(x₁ + x₂) = 7 - 2·3
- (2x₁ - 2x₁) + (3x₂ - 2x₂) = 7 - 6
- x₂ = 1
- Substitute back: x₁ + 1 = 3 → x₁ = 2
The answer comes out right — so missing the scaling did not break the answer here. But in more complicated systems, especially with zeros or identical multiples, failing to swap or scale can leave a zero row that hides information about a variable.
The grader's rubric, based on the definition of a row-echelon form (every leading entry in a row is to the right of the leading entry of the row above it, and every leading entry is 1), demands visible pivots — not just a final answer.
Boundary Case: When Zero Pivots Are Unavoidable
If you have a row of all zeros (e.g., after elimination, 0x₁ + 0x₂ = 0), this does not mean you did something wrong. It means the system has infinitely many solutions (the equations are dependent). For instance:
x₁ + 2x₂ = 6 2x₁ + 4x₂ = 12
After Row 2 - 2×Row 1: 0x₁ + 0x₂ = 0
Here, a grader gives full marks for reducing the matrix correctly, even if one row is all zeros. The mistake is only when a nonzero right-hand side appears (0 = 5), or when you fail to swap rows to reveal a pivot that should be there.
Your Next Attempt: Check for Pivots Before You Eliminate
If an x-variable seems to vanish, or if your lower rows lose all their variables without showing a clear contradiction or dependency, pause. Ask: could you have swapped rows to bring in a nonzero entry? Is there a step where dividing by a coefficient could have produced a leading 1?
Try this system on your own:
x₁ + 3x₂ = 7 2x₁ + 5x₂ = 12
Carry Gaussian elimination with no row swaps. What are your pivots? Now try swapping if needed. The correct answer is x₂ = 2, x₁ = 1.
One missed pivot can cost the whole solution — but catching it is a skill you can practice. If you spot when a leading 1 is missing, you can rescue the marks before submitting.
Summary
Suppose you are handed the system of equations:
