RREF Step-by-Step Tutorial

Quick Review: The Four RREF Conditions

A matrix is in Reduced Row Echelon Form (RREF) when all four conditions hold simultaneously:

1. All-zero rows at the bottom. Any row of all zeros sits below all non-zero rows.
2. Pivots equal 1. The first non-zero entry in each non-zero row is exactly 1.
3. Staircase pattern. Each pivot is strictly to the right of the pivot in the row above.
4. Zeros above AND below each pivot. This is what separates RREF from REF — all entries in a pivot column except the pivot itself are zero.

Every matrix has a unique RREF — no matter which sequence of row operations you use, you reach the same end state. (This is the uniqueness theorem for RREF, discussed in Strang's Introduction to Linear Algebra, §1.5.) New to the subject? The linear algebra basics guide covers vectors, matrices, and the motivation for row reduction from scratch.

Example 1: 2×2 Matrix

Find the RREF of A = [[4, 7], [2, 6]].

Start:

[ 4  7 ]
[ 2  6 ]

Step 1. Pivot column is column 1. The top entry is 4 (non-zero), so no row swap needed. Scale R₁ → (1/4)R₁ to make pivot = 1:

[ 1  7/4 ]
[ 2  6   ]

Step 2. Eliminate below pivot: R₂ → R₂ − 2R₁:

[ 1  7/4 ]
[ 0  5/2 ]

Step 3. Move to next pivot: column 2, row 2. Scale R₂ → (2/5)R₂:

[ 1  7/4 ]
[ 0  1   ]

Step 4. Eliminate ABOVE this pivot: R₁ → R₁ − (7/4)R₂:

[ 1  0 ]
[ 0  1 ]

RREF is the 2×2 identity matrix. Verification: A has full rank 2 and is invertible — the matrix inverse calculator can confirm A⁻¹ exists and compute it step by step.

Example 2: 3×3 Matrix with Fractions

Find RREF of B = [[0, 2, −1], [4, 0, 3], [−2, 1, 0]].

Start:

[  0  2  -1 ]
[  4  0   3 ]
[ -2  1   0 ]

Step 1. Column 1: the top entry is 0. Swap R₁ ↔ R₂ (row with non-zero first entry):

[  4  0   3 ]   ← R₂ moved up
[  0  2  -1 ]   ← R₁ moved down
[ -2  1   0 ]

Step 2. Scale R₁ → (1/4)R₁:

[  1   0   3/4 ]
[  0   2  -1   ]
[ -2   1   0   ]

Step 3. Eliminate column 1 below: R₃ → R₃ + 2R₁:

[  1  0   3/4 ]
[  0  2  -1   ]
[  0  1   3/2 ]

Step 4. Pivot column 2, row 2. Scale R₂ → (1/2)R₂:

[  1  0   3/4 ]
[  0  1  -1/2 ]
[  0  1   3/2 ]

Step 5. Eliminate below: R₃ → R₃ − R₂:

[  1  0   3/4 ]
[  0  1  -1/2 ]
[  0  0   2   ]

Step 6. Pivot column 3, row 3. Scale R₃ → (1/2)R₃:

[  1  0   3/4 ]
[  0  1  -1/2 ]
[  0  0   1   ]

Step 7. Eliminate above: R₁ → R₁ − (3/4)R₃, then R₂ → R₂ + (1/2)R₃:

[  1  0  0 ]
[  0  1  0 ]
[  0  0  1 ]

RREF is the 3×3 identity — B is invertible with rank 3.

Example 3: Augmented Matrix — Infinite Solutions

Solve the system: x + 2y − z = 1, 2x + 4y − 2z = 2, 3x + 6y − 3z = 3.

Augmented matrix:

[ 1  2  -1 | 1 ]
[ 2  4  -2 | 2 ]
[ 3  6  -3 | 3 ]

Step 1. R₂ → R₂ − 2R₁:

[ 1  2  -1 | 1 ]
[ 0  0   0 | 0 ]
[ 3  6  -3 | 3 ]

Step 2. R₃ → R₃ − 3R₁:

[ 1  2  -1 | 1 ]
[ 0  0   0 | 0 ]
[ 0  0   0 | 0 ]

This is already in RREF. Interpretation: only one pivot (column 1). Columns 2 and 3 have no pivots — y and z are free variables. The solution is:

x = 1 - 2s + t
y = s    (free parameter s)
z = t    (free parameter t)

This system has infinitely many solutions — a 2-dimensional family (plane in ℝ³). All three original equations are multiples of each other: row 2 = 2×row 1, row 3 = 3×row 1. The matrix has rank 1, nullity 2.

Example 4: No Solution (Inconsistent)

Augmented matrix: [[1, 2, 3 | 4], [2, 4, 6 | 9]].

[ 1  2  3 | 4 ]
[ 2  4  6 | 9 ]

R₂ → R₂ − 2R₁:

[ 1  2  3 | 4 ]
[ 0  0  0 | 1 ]

Row 2 reads "0 = 1", which is impossible. No solution exists. The coefficient rows are linearly dependent (row 2 = 2×row 1), but the right-hand sides are not proportionally related (9 ≠ 2·4). Geometrically: two parallel planes that don't intersect.

Example 5: 3×4 Augmented Matrix — One Free Variable

Solve the system: x₁ + 2x₂ − x₃ = 3, 2x₁ + 5x₂ + x₃ = 9, x₁ + 3x₂ + 2x₃ = 6. This gives a 3×4 augmented matrix [A|b] with 3 equations and 3 unknowns — but the matrix has rank 2, yielding one free variable.

Start:

[ 1  2  -1 | 3 ]
[ 2  5   1 | 9 ]
[ 1  3   2 | 6 ]

Step 1. Eliminate column 1 below: R₂ → R₂ − 2R₁, R₃ → R₃ − R₁:

[ 1  2  -1 | 3 ]
[ 0  1   3 | 3 ]
[ 0  1   3 | 3 ]

Step 2. Note R₂ = R₃ exactly. Eliminate column 2 in R₃: R₃ → R₃ − R₂:

[ 1  2  -1 | 3 ]
[ 0  1   3 | 3 ]
[ 0  0   0 | 0 ]

Row 3 is all zeros — the three equations are not independent. R₁ and R₂ already give us RREF in the first two pivot columns.

Step 3. Back-eliminate column 2 in R₁: R₁ → R₁ − 2R₂:

[ 1  0  -7 | -3 ]
[ 0  1   3 |  3 ]
[ 0  0   0 |  0 ]

This is RREF. Pivot columns: 1 and 2 (x₁ and x₂ are basic variables). Column 3 has no pivot — x₃ is a free variable. Set x₃ = t (any real number). Then:

x₁ = −3 + 7t
x₂ =  3 − 3t
x₃ = t   (free parameter)

This is a line of solutions in ℝ³ (parametric form). At t = 0: (−3, 3, 0). At t = 1: (4, 0, 1). Verify each satisfies all three original equations. Rank = 2, nullity = 1 (one free variable). The RREF calculator identifies free variables automatically when Augmented mode is enabled.

Common Mistakes in Row Reduction

These five errors account for the majority of wrong answers when computing RREF by hand. Each is fixable once you know to look for it.

Mistake 1: Arithmetic errors in fraction multiplication. When computing R₃ → R₃ − (5/3)R₁, you must multiply (5/3) by every entry in R₁ and subtract. Missing even one entry cascades: every subsequent step inherits the error. Fix: multiply the entire row explicitly before subtracting, and double-check each entry.

Mistake 2: Stopping at REF instead of RREF. After forward elimination (zeroing below each pivot), you have REF. RREF requires a backward pass that zeros above each pivot too. If your textbook teaches both Gaussian elimination (→ REF) and Gauss-Jordan (→ RREF), confirm which one the problem is asking for. When in doubt, RREF is the standard for most homework and exam problems.

Mistake 3: Applying the row operation to the wrong row. "R₂ → R₂ − 3R₁" means replace row 2 with (row 2 minus 3 times row 1). A common error is accidentally modifying R₁ instead. Keep R₁ untouched — only R₂ changes.

Mistake 4: Scaling the wrong row or by the wrong factor. To make the pivot in row k equal 1, scale row k by the reciprocal of the pivot. If the pivot is 3/4, scale by 4/3 — not by 3/4. This is the most common single-step error for students first encountering fraction pivots.

Mistake 5: Confusing free variables with the zero-row variable. In a system with a zero row [0 0 0 | 0], beginners sometimes think the "third variable" is 0. It is not — the zero row simply means one equation was redundant. The third variable is free (it can be anything), not zero. Only [0 0 0 | c] with c ≠ 0 means no solution.

Using a TI-84 to Compute RREF: Pros, Cons, and When to Use Each

The TI-84 graphing calculator has a built-in rref( function (found under MATRIX → MATH → rref). It computes RREF in one step. Here is an honest comparison of the TI-84 vs. computing by hand (or using this online calculator):

Recommendation: Use the TI-84 on time-pressured exams where the calculator is allowed and you only need the final RREF. Use this calculator (or work by hand) when you need to understand the steps, check individual row operations, or study for an exam where calculators are not allowed. For learning and homework verification, the step-by-step output of this calculator is far more valuable than the single-answer output of rref( on a TI-84.

The Gauss-Jordan vs. Gaussian elimination guide explains the underlying algorithm that both the TI-84 and this calculator use.

How to Use the RREF Calculator to Check Your Work

After computing RREF by hand:

1. Enter your original matrix into the RREF calculator.
2. Toggle "Augmented" on if your matrix has a right-hand side column.
3. Click "Calculate RREF" and compare your result to the calculator's RREF.
4. If they differ, expand the steps panel and compare step-by-step to find where your row operation went wrong.

The calculator uses exact rational arithmetic, so even complex fraction arithmetic (like 7/4 subtracted from 3/2) is handled without rounding. Have more questions about the calculator's features? See the FAQ page.