CRAP Score

CRAP stands for Cyclomatic Readability And Predictability. It measures how expensive a function is to test.

The Formula

\[ CRAP\left(CC, coverage\right) = CC^2 * \left(1 - \frac{coverage}{100} \right)^3 + CC \]

Where:

• \(CC\) - cyclomatic complexity (number of independent paths)
• \(coverage\) - test coverage as a percentage (0-100)

Interpretation

CRAP Range	Meaning
0 - 10	Well-tested, simple function
10 - 30	Moderate complexity, should be tested
30 - 50	High CRAP - refactoring or more tests needed
50+	Critical - likely hard to test, complex logic

How It Works

Cyclomatic Complexity

Computed by walking the AST and counting decision points:

• if, else if, for, case, &&, ||, ?

Each decision point adds 1 to the base complexity of 1.

Coverage

Obtained by running go test -coverprofile=... and reading the coverage profile. Coverage is expressed as a percentage of lines executed within each function.

Example

A function with CC=5 and 75% coverage:

\[ \begin{align} CRAP &= 5^2 * (1 - 0.75)^3 + 5 \\ &= 25 * 0.015625 + 5 \\ &= 0.39 + 5 \\ &= 5.39 \\ \end{align} \]

The same function with 0% coverage:

\[ \begin{align} CRAP &= 5^2 * (1 - 0)^3 + 5 \\ &= 25 * 1 + 5 \\ &= 30.0 \\ \end{align} \]

As coverage drops, the CRAP score rises - poorly tested complex functions score highest.

Reference

The original CRAP score was introduced by Adam Tornelius. go-crap uses the same formula as the cargo-crap Rust implementation.