Verified tape
18+ / 21+ by jurisdiction. Gambling help: 1-800-GAMBLER. Losing cards are normal and expected. Analysis only. No wagers are placed.
Likelier winner, but the price is too short to be a value bet.
Informational only. No wagers placed on your behalf. Past accuracy does not guarantee future results.
Past fights the rates below are counted on. Not either fighter's record.
Men · no-vig band ±5 · band-5 · 155 lb · 2010–2026
24 markets · 21 with model · 8 with market price
2.0% of model mass is method “other” (kept out of every market id).
Decision markets: model over-predicted decisions on the one graded card (5.1 expected vs 2 actual, n=14). Not a clean recommendation.
Decision markets: model over-predicted decisions on the one graded card (5.1 expected vs 2 actual, n=14). Not a clean recommendation.
Either fighter, either finish method. Not the same as a named fighter winning a given round by KO/TKO or submission.
This section carries three bases. On method and distance cards, MODEL and MARKET are conditional shares within each de-vig group. On round cells, MODEL is the unconditional fighter × method × round share of all fight outcomes, while COHORT is a share of that fighter's historical wins (same family as the method-card Cohort — e.g. TKO in R1 / that fighter's wins), not a share of all outcomes, and not comparable to MODEL. Round cells and “fight ends in this round” have no complementary book — MARKET reads No fair-price group, and the posted line is shown separately. EV/$100 uses the unconditional model probability against the raw American line.
Informational — recommendations balance confidence and EV, never EV alone.