High-frequency repeats need lighter feedback so motion does not outlast the beat

What it is

The first swipe-to-dismiss can use a short motion to say “it went”. Twenty in a row, each playing the same clip, and the clip is longer than the hand: the beat is held by feedback. Feedback on high-frequency repeats has to slim down to the operational beat: when the same action happens again inside a short interval, visual confirmation collapses to an instant state change, and does not walk through an entrance every time. Whether intervals line up is a prior question. Here the person is already firing in a string; the issue is whether feedback itself is slower than the hand.

Why it happens

Feedback has two durations. A state flip (ticked, row gone, count minus one) can finish in a frame, on the hand’s beat. Narrative motion (slide out, scale then settle, play a success clip) has its own few hundred milliseconds, meant for “this one has to be understood”. In a string, people do not need every instance understood. They need the beat not to be padded. If motion duration exceeds the firing interval, a queue builds: the hand is on the fifth item, the screen is still playing the first item’s exit. People either wait for the clip, or keep tapping on a target that is moving.

Slimming is not turning confirmation off. Confirmation can drop a gear: full on the first, state-only on the rest; or an instant change on a count instead of an exit per row. Once the hand pauses past a seam, the next one can take the full motion again — the seam means the string broke, and narrative is useful again.

Where it stops holding

Irreversible, high-cost actions (permanent delete, pay, send-to-everyone) should not drop confirmation to a one-frame flip even when someone is hammering; rhythm yields to consequence. If accessibility depends on motion to understand what happened, slimming must be backfilled with status text or speech, not only a one-frame colour change. First use, a demo, teaching need the full motion to build a model, then slim; slimming from the first beat never builds the model. Tracking a drag is already continuous feedback; it should not stack a per-beat success clip on top.

Applying it

• Give actions that can fire several times in a second two feedback gears: full confirmation on the first beat or after a seam, state flip and count only during a string.
• Motion duration must be shorter than the minimum repeat interval. If it would queue, switch to instant state, or merge into one summary when the string ends.
• Targets must not walk out from under the finger during a string. If they must move, gather them after the string ends.
• How to check: dismiss a dozen items in a row. The screen should not still be playing the previous exit while the hand has already moved down; count or state should track the hand. Pause two seconds and dismiss one; full motion may return. A 400 ms slide-out on every item, as a counterexample, should visibly hold the string.

Related

• Same group: I4.09.1 Consecutive actions need a steady beat; speeding up and slowing down breaks flow · I4.09.3 After delay breaks the beat, people slow their own rate to match unreliable response · I4.09.4 A reliable rhythm lets skilled users drop into automaticity without watching each confirmation
• Nearby: I1.05 Latency jitter · F7.02 Motion duration
• Search terms: repeat feedback · animation duration · rapid serial interaction