โ METHODOLOGY
No politics.
Just data.
Traditional MMA rankings are voted on by journalists and promotion insiders โ they reflect politics, recency bias, and name recognition. FightTruth rankings are built entirely from fight outcomes, and the same model powers our fight picks. Here's how.
S01Two rating systemsOVERVIEW
FightTruth uses two different rating models because no single model is best for every question. The right tool depends on whether you're asking โwho is the best fighter right now?โ or โwho is the greatest of all time?โ
FT SCORE
6.38
Active rankings
Makhachev ยท #1 Active P4P
ELO
2109
All-time rankings
Jon Jones ยท #1 All-Time
S02FT Score โ Active RankingsBRADLEY-TERRY MODEL
The FT Score answers the question: โIf these two fighters fought today, who would win?โIt's designed for comparing fighters who are currently active and competing.
โ
Who you beat matters more than how many you beat
A win over a top-5 fighter moves your score dramatically. A win over an unranked fighter barely moves it. The model learns each fighter's strength from the full web of results โ if A beats B and B beats C, then A beating C is expected and counts less.
โ
Recent fights count more
A fight from last month counts more than a fight from three years ago. The model gives more weight to recent performances so that a fighter's current form drives their ranking, not a win they had six years ago.
โฒ
Finishes are worth slightly more
A KO, TKO, or submission is a more decisive signal than a split decision. Finishes carry a 30% bonus weight โ not because we value violence, but because a dominant finish is stronger evidence of superiority than a close points win.
โ
Championship fights count more
A title fight in a major organization (UFC, ONE, Bellator, PFL) carries a 30% weight bonus. Winning 5 regional belts isn't the same as winning one UFC title fight โ the model knows the difference.
โ
Scores stabilize after retirement
When a fighter stops competing, their FT Score locks in relative to their last fight. It won't drift down to zero as time passes โ a retired champion keeps the score they earned.
WHERE IT'S USED
Active RankingsFighter ProfileMatch Simulator
S03Elo Rating โ All-Time RankingsCHRONOLOGICAL ELO MODEL
The Elo rating answers a different question: โWho were the greatest fighters in history?โIt's borrowed from chess and adapted for MMA to make fair comparisons across eras.
โถ
Fights are scored as they happened
The model walks through every fight in history, oldest to newest. When a fight happens, each fighter's rating at that moment determines how much the result is worth. Beat a highly-rated fighter and your rating jumps. Lose to an unrated newcomer and it drops hard.
โ
Opponent strength at the time of the fight
This is the key difference from FT Score. BJ Penn's career-long rating has been dragged down by losses late in his career. But when Georges St-Pierre beat him in 2009, Penn was ranked top 3 pound-for-pound โ and GSP gets credit for beating that version of Penn, not the 2016 version.
โฒ
Finishes are worth slightly more
Same logic as FT Score โ a KO or submission carries a small bonus because it's a more convincing result.
โ
Championship fights count more
A title fight in a major organization (UFC, ONE, Bellator, PFL) carries a 30% weight bonus โ applied at the time of the fight. Josรฉ Aldo's 10 consecutive UFC title defences each get this bonus against the opponent's rating at that moment in history.
โ
Each fight matters less as your career grows
Early in a career, results shift ratings significantly because there's little data. By fight 100, the model is quite confident in a fighter's level, so each additional result barely moves the needle. This prevents journeymen with 200 regional fights from outranking elite fighters with 20 high-level bouts.
โ EXAMPLE ยท GEORGES ST-PIERRE
FT SCORE (active model)
4.34
Low โ because his opponents' career-wide scores are pulled down by their post-GSP losses
ELO RATING (all-time)
2045
#3 all-time โ because he beat prime Hendricks, prime BJ Penn, prime Hughes
WHERE IT'S USED
All-Time RankingsFighter Profile
S04Active vs All-TimeRANKING VIEWS
Every ranking page has two views โ you can switch between them at any time.
โ ACTIVE
Current Landscape
Only fighters who have competed in the last 2 years. Sorted by FT Score. Best for answering: โWho should fight for the title right now?โ
โ ALL TIME
Greatest Ever
All fighters including retired legends. Sorted by Elo rating. Best for answering: โWas GSP better than Khabib?โ
S05Fight Picks & Betting EdgePREDICTIVE MODEL
The same rating data that powers rankings feeds directly into a prediction model โ generating win probabilities for upcoming fights and comparing them against what the bookmakers are implying. When the model disagrees with the market, that's the edge.
โ
Win Probability
Each fighter's FT Score is fed into a model that outputs a win probability. This is our independent estimate of who wins โ derived purely from fight outcomes, not betting market sentiment or public perception.
โฒ
Model Edge
We compare the model's win probability against the implied probability baked into each bookmaker's odds. If the model gives a fighter 60% but the book implies 48%, that's a +12% edge โ a meaningful signal that the market may be mispriced.
โ
Kelly Sizing
Edge alone doesn't tell you how much to bet. The Kelly Criterion calculates the mathematically optimal fraction of your bankroll to wager, given the edge and the odds. FightTruth displays the Kelly-sized recommendation alongside each pick.
โ
Best Odds Across 7 Bookmakers
We compare moneyline odds from BetOnline, MyBookie, Ladbrokes, FanDuel, BetMGM, Stake, and Polymarket โ refreshed every 6 hours. Fight Picks always shows you the best available line, and Best Odds lets you compare all 7 side by side.
WHERE IT'S USED
Fight PicksBest OddsMatch Simulator
S06What we don't useINTENTIONAL OMISSIONS
โ
Promotion
A win in the UFC, Bellator, ONE, or a regional show is treated the same. Promotion prestige is not a factor.
โ
Title belts
Holding a belt doesn't boost your score. Winning the fight that earned it did.
โ
Journalist votes
No human bias, no name recognition, no recency bias beyond what the data shows.
โ
Weight cutting
We rank fighters in their listed weight class. Extreme cuts are not rewarded or penalized.