What the workflow does
The test uses a structured 75-putt sweep design. Three Meta captures failed, leaving 72 matched pairs. The report separates per-sweep validation from an all-putts view and evaluates bias, absolute error, RMSE, scaling, R², limits of agreement, and per-stroke error.
Why it matters
The output makes it possible to distinguish strong channels from calibration issues instead of treating the device as a single pass/fail result.
Design approach
Metrics are grouped, labeled, and annotated so technical readers can inspect the evidence while nontechnical readers can still see the conclusion quickly.