About this model
A FRAM model of Boyd's OODA loop, built function by function from sensory encoding to motor output and run in milliseconds. A ball comes over the net. If its spin matches the internal model, prediction error stays low and the fast path acts in about 200 ms. If it surprises, the beta hold pauses the motor command while the model is revised, which costs 300 to 500 ms. Run many trials and the surprises are learned: expertise shortens the loop.
The band traces are schematic envelopes. The readiness trace sums the model's activity in an accumulator (Schurger and colleagues, 2012), drawn negative-up as Libet recorded it.
Structure: the FRAM model "OODA model ds 2" (31 functions), as drawn in FMV. Timings follow "Timing, Prediction, and Volition" (Slater). Thresholds, error rates and the learning rule are illustrative assumptions, not measurements.