What the flight benchmark tells us

3.8%

Mean absolute carry percentage error against normalized TrackMan outputs. This is a flight-model comparison, not an independent outdoor measurement of the whole fitting experience.

Validation sample	573 primary shots · 28 players
Model	Base flight model · measured launch and report weather
Exact percentage error	3.7580% (displayed as 3.8%)
Mean absolute carry error	6.08 m · 6.6 yd
95th-percentile absolute error	15.06 m · 16.5 yd
Snapshot	September 4, 2026 · America/Chicago

How to interpret it

Players are separated between training and validation. The final holdout has not been evaluated. The percentage is the average absolute percentage difference per shot, not “96.2% accurate” and not a guarantee for an individual shot or club.

The fitting UI starts from a swing profile or imported input and uses equipment-response assumptions. That is different from supplying exact measured launch conditions to this benchmark. An experimental lower-error calibration remains disabled.

What is not validated

Equipment recommendations, ranking weights, or causal gains from changing a shaft, head, or ball.
Real-turf bounce and roll, outdoor wind variability, or every club/player cohort.
The accuracy of an estimated swing profile or the certainty of a one-shot equipment comparison.

All fitting gains are simulated estimates. Confirm promising setups in an independent launch-monitor fitting.

Reproducible snapshot

Run: 2026-09-05T02-12-20-403Z-83f18219
Group: validation/primary/measuredBase
Source hash: 83f18219a59dffa0e9a3cdccc51f5ee9cb40e60ac8a94009a2a5ea8a68c904b8

Only aggregate statistics are shown here; no player records are published. Refresh this snapshot after changing the flight model.