The suite, and the pass it will not confirm

Suite lab: L4.2's four text assertions catch 29 of 34 failures and clear 5 of them by mistake; L4.3's reference grader catches 30 and abstains on 20. Composed, the pair catches all 34 — and 25 rows are caught twice, so the two catch counts sum to 59 verdicts for 34 failures (173.5%). Under the rule that a pass needs every rung to pass, the suite decides 34 rows, catches 34 and clears 0, reading 100.0% while answering only 68.0% — the constant-fail figure exactly. Under the rule that a pass needs no rung to accuse, it clears all 16 passes and still catches all 34, licensed by 0 failures escaping both rungs. Adding the reference rung withdraws 0 of the 16 accusations the obvious variants raise, and the cheaper order saves 57.0 ms without changing a catch.

This Evals Engineering lesson is part of a paid plan. Sign in with a plan that includes this course to read it.