Cross-sport calibration scoreboard. Brier scores compare model
consensus_prob to the picked-side actual outcome; market Brier uses
square_prob (no-vig anchor). Lower Brier = better. Lift = model − market;
negative = model beats market. Status uses the tennis tripwire pattern:
calibrated if |lift| < 0.005,
watch if < 0.020,
action otherwise.
Cross-sport calibration scoreboard: N picks, model Brier, market Brier, lift, and calibration status per sport.