What It Studies
The gap between training and competition. A tennis player moves smoothly and automatically in practice — skill operating at 90%. Under match pressure, the same player hesitates, tightens, starts checking their technique. Performance drops to 70%. The skill is still there. It just cannot be accessed.
This is not nerves. Something structural changes when the stakes become real.
Why It Matters
Every approach to athletic performance improvement assumes the athlete’s problem is skill acquisition — learn the shot, drill it, deploy it. But the performance gap is not a skill gap. Players already have the skill. They lose access to it under pressure.
If pressure is the variable that controls the transition between having a skill and being able to use it, then the intervention point is not more drilling. It is understanding what pressure does to the cognitive system that coordinates execution.
What We Hypothesized
Pressure is a form of internal data — generated by anxiety, fear, uncertainty. The player must simultaneously process external data: opponent movement, ball spin, trajectory, court positioning. All of this competes for a constrained processing space — a single memory slot, analogous to RAM.
When internal noise fills the slot, there is no room left for the signals execution depends on. The player chokes. When the slot is clear — as in flow state — the system runs at full bandwidth: automatic execution, anticipatory positioning, filtered perception. The ball appears to move in slow motion. The body feels lighter. Every shot is as planned.
Three layers emerge from this framework:
Layer 1 — Self-Discovery. Identify the player’s strongest natural tactical sequences — their “trump cards.”
Layer 2 — Opponent Tendencies. Map the patterns where the opponent consistently struggles.
Layer 3 — Cue Design. Build simplified triggers that activate multi-step patterns without overloading the memory slot. A good cue carries maximum tactical information with minimum cognitive load — the way “hit a rainbow” teaches a child to serve without explaining swing path, contact point, or toss height.
What We Found
The engineering investigation began with video: 45 commits over 9 days building rally detection, shot annotation, and movement segmentation from 2D broadcast footage. The sensor was wrong. 2D video lacks the granularity to detect the substrate-level shifts that precede performance collapse.
But the core hypothesis held — execution breakdown does have detectable precursors. The question was not whether the signal exists, but which sensor can reach it. Voice turned out to be the answer.
The pivot was in the sensor, not in the question. The cognitive framework — memory slot, cue design, pattern activation — remains active research.
Open Questions
- Why does a player exit automatic mode even without external pressure? (e.g., choking when leading 5-0)
- Can a simple cue activate a whole multi-step tactical pattern without triggering overthinking?
- Do novices and experts respond to the same cues differently — and if so, why?
- Is there a trainable pathway to reliably enter flow state?