Muscle Memory

Muscle memory is not stored in the muscles at all, it is a popular label for procedural motor learning happening in the brain, primarily in the motor cortex, basal ganglia, and cerebellum. Fitts and Posner's (1967) three-stage model describes how this happens: a cognitive stage, where the learner consciously works out what to do and makes large, variable errors; an associative stage, where the movement smooths out and errors shrink as the correct pattern gets reinforced; and an autonomous stage, where the movement runs with minimal conscious attention, freeing cognitive resources for higher-level decisions happening at the same time. A player's crosshair placement or ability-cast timing moving from the associative to the autonomous stage is exactly what lets them think about strategy mid-fight instead of the mechanics of the fight itself.

Reference: Fitts & Posner, 1967. The NeuroRank implementation holds the canonical form and scales interference via task-irrelevant stimulus density.

1. Executing a practiced recoil-control pattern without consciously tracking the crosshair correction on each shot.
2. Hitting a combo's input sequence from years of repetition while attention is spent reading the opponent instead.
3. Landing a skillshot's lead and timing automatically after enough reps, rather than calculating travel time each cast.

NeuroRank measures the output of automaticity rather than the learning process itself: the Aim and Tracking modules' consistency scores reflect how stable a player's motor execution is within a single session, which is the signature of a movement that has already reached, or is close to, the autonomous stage described by Fitts and Posner.