Not every test deserves to live forever. Some tests stop pulling their weight and become liabilities. Knowing when to delete a test is a skill.
Multiple tests assert the same thing with slightly different wording. If deleting one doesn't reduce confidence, it's dead weight.
The test breaks constantly due to harmless refactors (renames, formatting changes) without actually catching real bugs.
Tests that assert trivial facts (e.g., assertTrue(true)) or just hit code for coverage don't provide value.
The functionality under test has been removed, or the test is checking a deprecated path no one relies on.
A test that takes 20 minutes to run and rarely finds a bug might be better off archived or reimagined.
👉 If a test doesn't increase confidence, it's technical debt. Delete it.