Fast, short movements are dominated by an open-loop component

What it is

A racket swing, a quick tap, a brief flick-style swipe — movements this short are called ballistic movements, borrowing the image of a shell that's no longer steerable once fired: once launched, the movement flies out mostly on parameters fixed before it started. These movements are open-loop-dominated not because they're unimportant, but because they're simply too fast — too fast for feedback to have any chance of acting before the movement is already over.

Why it happens

Visual or proprioceptive feedback has a minimum-duration processing chain to get from being sensed to actually changing a movement in progress — that chain typically takes on the order of a hundred-plus milliseconds to run once. When a movement's total duration, start to finish, is already shorter than that chain, feedback — no matter how accurate — can't arrive in time to correct course before the movement ends, so the movement has no option but to run entirely on the direction, force, and timing set before launch. The shorter and faster the movement, the narrower the window left for feedback to act, and the larger the open-loop share becomes; once total duration clearly exceeds the feedback loop's own latency, there's enough room for in-flight correction to enter, and open-loop's dominance recedes.

Studying it

Researchers typically vary the allowed duration of a target movement systematically and watch for a turning point in accuracy: below some duration threshold that matches the feedback loop's own latency, artificially cutting feedback doesn't further degrade accuracy — meaning movements at or below that duration are already decoupled from feedback and open-loop-dominated. Once duration exceeds that threshold, cutting feedback clearly hurts accuracy, showing that feedback correction has entered the picture. Another common approach directly measures the feedback loop's own minimum latency — for instance, the shortest interval between visual information being presented and it being able to trigger a change in muscle response — and uses that number as a benchmark for whether a given class of movement could possibly make use of feedback at all.

Where it stops holding

This conclusion depends on the relative size of two quantities — movement duration and feedback-loop latency — and the exact cutoff shifts with the feedback channel (vision, touch, and proprioception each have their own processing speed), the body part involved, and individual differences; there's no single millisecond figure that applies across every context. A short overall movement duration also doesn't mean every segment of it is open-loop — many fast movements still pick up a small amount of coarse proprioceptive (rather than visual) correction right at the very end, though this correction is far less precise than the visual closed-loop correction seen in slower movements.

Applying it

• For operations that are inherently brief — a quick tap, a flick-style swipe — put the effort to improve accuracy into what happens before the movement launches: target size, starting position, grip stability — rather than counting on the user to correct course mid-execution based on interface feedback, since that feedback simply can't arrive in time to be used.
• If accuracy on some operation genuinely needs improving, consider deliberately lengthening its required execution time (requiring a longer hold or a longer swipe distance), pushing the movement out of the purely open-loop regime into one where closed-loop correction can operate — rather than simply enlarging the target or adding more cues.
• How to check: measure accuracy on the target operation under two conditions — real-time feedback provided, and feedback fully occluded. If the two don't differ significantly, the operation already sits in the open-loop-dominated regime, and further investment in real-time feedback design has limited payoff.

Related

• Same group: A8.06.1 An open-loop action does not draw on feedback once it has launched · A8.06.3 Open-loop error comes from programming precision, not execution correction · A8.06.4 Without any feedback, control has to stay open-loop, and accuracy has a fixed ceiling
• Nearby: A8.08 Two-Phase Structure of Target Acquisition · A8.01 The Composition of Reaction Time
• Search terms: ballistic movement · open-loop control · feedback processing time