Fight IQ Analytics — Powered by UCFR™
Fight IQ Analytics GOAT Model

UFC GOAT Leaderboard

How exceptional was each fighter's UFC career? Scored chronologically from Cito bout history and pre-fight UCFR ratings — every component is shown. A transparent model, not an objective verdict.

Methodology & normalisation
  • Peak: Best rolling window of N consecutive UFC bouts (default 5): 60% mean own pre-fight UCFR, 25% mean result quality (win = opponent pre-fight UCFR + 10, loss = opponent UCFR − 40, draw = 60% of opponent UCFR; unknown opponent = 45), 15% mean pre-fight Fight IQ. Mapped linearly 40→0, 80→100.
  • Resume: Sum of win points by opponent pre-fight UCFR tier (elite ≥ elite threshold 4 pts, strong 2, solid 1, below 0.4, unrated 0.25; title-fight win ×1.5). Score = 100·(1 − e^(−points/K)) — diminishing returns so raw volume cannot dominate.
  • Championship: Undisputed title wins 12, interim title wins 5, successful defenses 7, longest defense streak 2 each, years holding a belt 4 each, extra divisions won 12 each; multiplied by (0.7 + 0.3·title-fight win %). Score = 100·(1 − e^(−points/K)). Each title fight counted once as exactly one of: championship win, interim win, defense or loss.
  • Dominance: Per bout with round statistics: sig-strike differential per round (−10→15), knockdown differential (−0.5→1), takedown differential per round (−1→2), control-time share (20%→80%), accuracy differential (−15%→+20%), strike defense (35%→70%). Weighted .30/.15/.15/.15/.10/.15, averaged over bouts, shrunk toward 50 by n/(n+2). With era adjustment on, differentials are divided by the running standard deviation of the same era (and division when ≥30 prior samples) from strictly earlier dates.
  • Longevity: Meaningful activity only: UFC fights (0→25), competitive seasons = distinct years with a fight (0→12), strong-tier opponents faced (0→12), elite opponents faced (0→6), years holding a title (0→5), fights entered at strong-tier UCFR (0→15). Years on the roster without fights add nothing.
  • Consistency: Needs ≥4 bouts. Pre-fight UCFR standard deviation (12→0, 2→100) 35%, recency-damped win rate (last 3 bouts count half, so late-career losses are not over-punished) 35%, frequency of UCFR drops > 5 points (40%→0, 0%→100) 20%, win rate vs comparable opponents (±5 UCFR) 10%.
  • Finishing: Finish rate of wins shrunk by wins/(wins+3) toward 35% (10%→0, 80%→100) 40%, total finishes (0→12) 30%, finishes over strong-tier opponents (0→6) 30%. Decision wins are not penalised beyond not counting as finishes.
  • Fight IQ: The fighter's pre-fight Fight IQ skill score entering the most recent bout in the window — never a later value. Missing → Insufficient Evidence, weight redistributed and Evidence Strength reduced.
  • GOAT score: Weighted mean of available components (weights total 100%). Peak and Resume are required; a missing Dominance or Fight IQ has its weight redistributed and is flagged. Fewer than the minimum UFC fights → Insufficient Evidence.
  • Evidence Strength: Separate from GOAT: UFC fights (0→15) 25%, share of bouts with round statistics 25%, share of opponents with a pre-fight UCFR 20%, own pre-fight UCFR coverage 10%, Fight IQ available 10%, title-fight data completeness 10%.
  • No look-ahead: each snapshot after a fight uses only that fight and earlier ones; opponents are valued at their pre-fight UCFR, never a later career rating. "n/a" means Insufficient Evidence — its weight is redistributed and Evidence Strength is lowered.

Fight IQ Analytics provides statistical and analytical information for research and entertainment. Model outputs are not guarantees of future results.