Pairwise testing
Compare all combinations with Pairwise, then apply exclusions to build your test set.
Generate combinations →- No account or installation
- Copy or download results
- Documents processed in your browser
File sizes and generation counts have limits. Check the limits shown in each tool.
Input and result example
With no exclusions, the three two-value factors below can be covered pairwise by these four rows. This is one valid arrangement; the tool may produce a different order or set.
| Browser | OS | Role |
|---|---|---|
| Chrome | Windows | User |
| Chrome | macOS | Admin |
| Edge | Windows | Admin |
| Edge | macOS | User |
All combinations: 8. Example pairwise set: 4. For exclusions, see the example below.
An eight-combination example
Browser: Chrome, Edge OS: Windows, macOS Role: User, Admin
Two values for each of three factors produce 2 × 2 × 2 = 8 combinations. Suppose your test scope excludes Edge on macOS—not because the browser cannot run there, but because your project does not support that combination.
OS=macOS & Browser=Edge
What exclusion changes
The rule removes both roles for that OS/browser pair, leaving six feasible combinations. Pairwise must cover every pair that occurs in at least one feasible complete combination. It must not simply delete rows from an already-reduced set, which could lose coverage.
Choosing a method
- Use all combinations when the set is small or every supported configuration must be checked.
- Use Pairwise when two-factor interactions are an acceptable coverage target.
- Use targeted extra cases for high-risk interactions involving three or more factors.
Review before export
Check the total, feasible and generated counts. Investigate any value that exclusions make unreachable. Fewer tests are not automatically better. The current algorithm guarantees feasible pair coverage within its limits, not the smallest possible test suite.