A sanitized example of what a coach receives: Driver Cards with effect, interval, evidence state, decision license, caveats, and next-data guidance.
In this sample, longer sleep before testing is associated with higher CMJ height after accounting for load and soreness. The coach can use it, but the interval and caveats travel with the recommendation.
This sample athlete does not have enough repeatable threshold tests for a personal estimate. A comparable-athlete prior can guide observation, explicitly flagged as borrowed prior evidence.
The pattern looks directionally plausible, but the robust interval still overlaps no effect. The correct output is not a weaker recommendation; it is a no-claim state.
There are not enough repeatable outcome events to estimate this relationship. AthDash declines rather than filling the gap with an effect direction.