RHY.14:5.1 - A smaller average hides an unmatched attack

The intended passage has five attacks at 0, 0.5, 1, 1.5 and 2 seconds, with no additional attacks. The local task accepts a deviation up to 0.10 seconds for event matching. It requires all five attacks.

Recording A contains attacks at 0.02, 0.52, 1.20, 1.52 and 2.02. Inspection confirms that these are performed attacks rather than detector echoes. The matching rule pairs four events, each 0.02 seconds late; the intended location at 1 has no match within its window, and the attack at 1.20 lies outside all admitted windows. The recording still contains five attacks; the two unmatched entries describe this rule’s failed correspondence, not proof that two separate physical errors occurred.

Recording B has five matched deviations: −0.04, 0.03, −0.02, 0.04 and −0.01 seconds. Its mean absolute deviation is 0.028 seconds. A’s matched-event mean is smaller, 0.02, but A does not satisfy the required event sequence under the stated rule.

The useful comparison retains B for the complete passage and directs A’s next repair to the misplaced attack. It does not select A by silently discarding its unmatched events or mispair its late attack with the next beat. If the task instead allowed that displacement as an ornament, the matching and success conditions would have to change.

An ambiguous correspondence. In a different two-attack fragment, the expected times are 0 and 0.12 seconds and the allowed deviation is 0.10 seconds. The recording contains one confirmed attack at 0.06 seconds. It can match either expected attack, but cannot satisfy both. The complete-fragment requirement already fails; the missing attack’s identity remains unresolved. A decision to repeat the whole fragment needs no further observation. If the correction depends on which action was omitted, inspect an identifiable sound or action in the existing recording. If it supplies no distinction, retain that uncertainty rather than assigning the nearest event by convention.