Sarrus Rule Calculator guide
Why the Rule of Sarrus Only Works for 3×3 Matrices
A 3×3 determinant has exactly 6 terms, and the widened Sarrus grid produces exactly 6 diagonals — so the picture captures the whole determinant. A 4×4 determinant has 24 terms, but the same construction yields only 8 diagonals, missing two thirds of the answer. The coincidence that makes Sarrus work exists only at 3×3.
It comes down to counting
If you have ever wondered whether you could just draw more diagonals and make Sarrus work on a bigger matrix, this guide is the answer. The reason you cannot comes down to two numbers not matching, and once you see them side by side it is quite clear.
How many terms a determinant has
The determinant of an n×n matrix is a sum of n factorial terms. Each term is a product of n entries, picked so that no two of them share a row or a column.
For a 2×2 matrix that is 2 terms — ad and bc, which you already know. For a 3×3 it is 6 terms. For a 4×4 it is 24, and for a 5×5 it is 120.
That growth is brutal, and it is the reason nobody computes large determinants this way.
How many diagonals the trick gives you
Now count the other side. Widening an n×n matrix by repeating its first n−1 columns gives you n diagonals running down-right and n running up-right — 2n in total.
At n = 3 that is 6 diagonals for 6 terms. An exact match, which is precisely why the method works.
At n = 4 it is 8 diagonals for 24 terms. Sixteen of the terms have no diagonal to sit on, and no amount of redrawing the grid will create room for them.
The numbers side by side
| Size | Terms in the determinant | Diagonals available | Does Sarrus work? |
|---|---|---|---|
| 2×2 | 2 | 4 | Not needed — ad − bc is already minimal |
| 3×3 | 6 | 6 | Yes — exact match |
| 4×4 | 24 | 8 | No — 16 terms missing |
| 5×5 | 120 | 10 | No |
What to use on bigger matrices
For 4×4 and up, use cofactor expansion along whichever row or column has the most zeros in it. Each zero removes an entire sub-determinant from your workload, so a well-chosen row can cut the work in half.
The other option, and the one computers actually use, is to apply row operations until the matrix is triangular — everything below the diagonal is zero — and then just multiply the entries along the diagonal.
That approach matters more than it might sound. Its cost grows roughly like n³ rather than n factorial, which for a large matrix is the difference between a fraction of a second and a length of time longer than the universe has existed.
