Orders tests so failures surface earliest — runs tests covering changed code first, historically flaky/failing tests early, and slow low-value tests last. Use when the suite is too slow to run in full on every change, when CI feedback takes too long, or when deciding what to run in a smoke-test tier.