Row reduction and RREF intuition

Row reduction is a bookkeeping method for solving a system of linear equations. You write the system as an augmented matrix, then use three safe moves until the matrix is in reduced row echelon form, or RREF. In that form the answer can be read off directly: one solution, no solution, or infinitely many.

The idea is the same elimination you met in school algebra. The matrix just drops the letters so you can see the pattern. Zoom out on the linear algebra pillar, follow the linear algebra first course, and check each example below on the RREF calculator.

The three safe moves

1. Swap two rows.
2. Multiply a row by a number that is not zero.
3. Add a multiple of one row to another row.

Each move can be undone by another move of the same kind. That is the whole reason they are safe: the new system has exactly the same solutions as the old one. Multiplying a row by zero is not allowed, because it throws an equation away and cannot be undone.

What RREF looks like

A matrix is in reduced row echelon form when every nonzero row starts with a 1 (a pivot), each pivot sits to the right of the pivot above it, any rows of zeros are at the bottom, and each pivot is the only nonzero entry in its column. A useful fact: the RREF of a matrix is unique. Two people can use different moves and still land on the same final matrix.

Worked example: one solution

Solve $x+2y=5$ and $3x+4y=6$. The augmented matrix is $\left[\begin{array}{cc|c}1&2&5\\3&4&6\end{array}\right]$.

1. Replace row 2 with row 2 minus 3 times row 1: $\left[\begin{array}{cc|c}1&2&5\\0&-2&-9\end{array}\right]$.
2. Divide row 2 by $-2$: $\left[\begin{array}{cc|c}1&2&5\\0&1&4.5\end{array}\right]$.
3. Replace row 1 with row 1 minus 2 times row 2: $\left[\begin{array}{cc|c}1&0&-4\\0&1&4.5\end{array}\right]$.

Read it back as equations: $x=-4$ and $y=4.5$. Check in the originals: $-4+9=5$ and $-12+18=6$. Both hold.

Worked example: no solution

Take $x+y=2$ and $2x+2y=5$. Subtracting 2 times row 1 from row 2 gives the row $\left[\begin{array}{cc|c}0&0&1\end{array}\right]$, which says $0=1$. That contradiction means the system is inconsistent. Geometrically, the two equations are parallel lines that never meet.

Worked example: infinitely many solutions

Take $x+y=2$ and $2x+2y=4$. The same move turns row 2 into all zeros, which says $0=0$ and adds nothing. Only one pivot remains, in the $x$ column, so $y$ is a free variable. Write $y=t$ and $x=2-t$ for any number $t$. Both equations describe the same line.

Pivots, rank and free variables

The number of pivots is the rank of the matrix. For a consistent system with $n$ unknowns, the number of free variables is $n$ minus the rank. A square coefficient matrix whose RREF is the identity has full rank, so the system has exactly one solution for every right hand side. That same matrix is invertible and has a nonzero determinant, which links row reduction to the 3 by 3 matrix calculator and to eigenvalues in plain language.

Common mistakes

• Changing two rows at once using each other. Update one row per step, using the current version of the others.
• Forgetting the right hand column. Every move applies to the whole row, including the augmented part.
• Stopping at echelon form and calling it reduced. Echelon form still has entries above the pivots; RREF clears them.
• Reading a zero row as no solution. A row of all zeros is harmless. Only $0=$ a nonzero number signals no solution.

Try it with practice tools

Citations and further reading

• Gilbert Strang, Introduction to Linear Algebra, elimination chapters. math.mit.edu
• MIT OpenCourseWare, 18.06 Linear Algebra. ocw.mit.edu
• Paul’s Online Math Notes, augmented matrices. tutorial.math.lamar.edu
• Khan Academy, linear algebra. khanacademy.org

Last reviewed: 1 November 2026.