Runnable proof — not screenshots. Every card is a machine-readable rule; run it on your own screen against the same historical data and reproduce the same trace. Metrics are cost- and behaviour-focused; the jury selection and community favourites are separate lanes.
Loading…