Chance first, price second
FightAlpha starts by estimating each fighter's chance from historical performance, matchup traits, recent form, and available public data. The model then compares that estimate with the market price. A fighter can be likely to win and still be a bad bet if the price is too short. If you want to translate a line before reading the edge, use the FightAlpha odds calculator.
PASS is part of the method
Many fights are not worth forcing. If the available data is thin, the edge is too small, or the odds do not justify the risk, the fight should be marked PASS instead of turned into artificial confidence.
Model League
Candidate models are tracked silently before they can become public. Promotion requires forward-tested evidence across accuracy, ROI, calibration, odds buckets, favorite and underdog splits, PASS discipline, and stake behavior. The current public model remains in place until a clear contender earns promotion.
Data sources and freshness
FightAlpha uses public fight records, fighter profile data, matchup statistics, odds snapshots, and manually reviewed card information where needed. Pages are updated when cards change, odds are refreshed, or results are settled.
Responsible use
The model is built to support decision-making, not to encourage chasing losses or oversized bets. Stake sizing is conservative and every card includes reminders that outcomes are uncertain.