UFC Immediate Rematch Win Rate: Do Winners Repeat?

Quick answer

Yes. In a strict sample of 17 modern-era UFC immediate title rematches, the winner of the preceding fight went 11-5-1. Excluding the draw, that is a 68.8% repeat rate across 16 decided bouts. The sample is small and highly selected, so it is a useful baseline, not a rule. Broader data points the same way: a UFCStats.com analyst found first winners repeated in 63.2% of 76 decided title-fight rematches through June 2024.

Data snapshot: 17 immediate UFC title rematches from November 2000 through UFC 320 in October 2025; prior winner 11 wins, 5 losses, 1 draw. Broader cross-check: 83 title-fight rematches through June 28, 2024; 76 decided, with 48 repeat winners.

What does the UFC immediate rematch win rate show?

ESPN listed 16 modern-era cases before UFC 320. Its definition was strict: the title changed hands, and the same fighters met in their next bout. The deposed champions went 4-11-1. Read from the other side, the prior winner went 11-4-1.

Alex Pereira then beat Magomed Ankalaev at UFC 320. It was the next fight for both men after Ankalaev took Pereira's light heavyweight belt at UFC 313. UFC Stats records Pereira's first-round stoppage at 1:20.

Adding that verified result to ESPN's table changes the record to 11-5-1 for prior winners. This is arithmetic from two named sources, not a new AgentMMA outcome study.

If you keep the draw in the denominator, repeat wins account for 64.7% of all 17 cases. That is also correct arithmetic, but it answers a different question. For “who won?”, the 16 decided bouts are the cleaner denominator.

One result moves this sample by more than six percentage points. Keep that in mind before treating 68.8% as a stable law.

What counts as an immediate UFC title rematch?

“Immediate” does not mean the UFC booked the fight quickly. It means neither fighter had another bout before meeting again.

That rule can include the third fight in a rivalry. If Fighter A wins the second meeting and both fighters go straight into a third, the relevant prior winner is Fighter A. It is not always the winner of the rivalry's first fight.

A normal title rematch is wider. The fighters may face other opponents, move divisions, or wait years before meeting again. An ordinary UFC rematch is wider still because no belt needs to be involved.

Those groups cannot share one headline rate. They contain different fighters and different matchmaking filters. A strict immediate title runback is the smallest and most selected group.

Do broader UFC rematch studies find the same pattern?

Yes, though the size of the lean changes with the definition.

FightAlpha matched UFC fighters by ID and audited second-or-later meetings through May 2026. It excluded draws, no contests, and missing winners. Across 192 later meetings, the previous winner repeated 115 times, or 59.9%.

The title split sits close to the strict immediate figure. The non-title rate is much nearer a coin flip. Time also matters in the raw groups: repeat rates fall as the gap grows.

An independent count supports the title pattern. Legal Sports Report published an audit from Michael Carroll, a UFCStats.com research analyst. He found 83 UFC title-fight rematches through June 28, 2024. Seven produced no winner.

The first winner repeated in 48 of the 76 decided bouts, or 63.2%. At light heavyweight, the count was 7 of 11, or 63.6%. Eleven bouts cannot establish a dependable division effect.

An older FightMetric count also pointed the same way. MMA Fighting reported 49 championship rematches in 2017: 26 repeat wins, 21 reversals, and two draws. The history grew, but the basic lean survived.

Why might the prior winner keep an edge?

These counts show association. They do not prove that winning once causes an advantage next time.

The preceding fight reveals information. A hard wrestling matchup can remain hard. A speed gap can remain visible. The loser can prepare a better answer, but the winner can adjust too.

Quick runbacks also leave less time for either athlete to change. That fits FightAlpha's raw time-gap split. It does not prove time caused the difference because the bins do not control for age, odds, injuries, or matchmaking.

Selection is the larger problem. Former champions do not receive immediate rematches at random. Long reigns, close decisions, disputed endings, star power, and thin contender queues can all shape who gets another chance.

The 17 cases therefore describe champions whom the UFC chose to run back. They do not describe an average pair of UFC fighters.

What do Alex Pereira's rematches teach us?

Pereira has landed on both sides of the pattern. That makes him a better example than a tidy slogan.

He stopped Israel Adesanya at 2:01 of round five at UFC 281. Their immediate title rematch came five months later. Adesanya reversed the result with a knockout at 4:21 of round two at UFC 287.

Pereira later stopped Jiří Procházka at 4:08 of round two at UFC 295. Both men took another fight before their UFC 303 rematch, so this was not immediate under ESPN's rule. Pereira repeated the result 13 seconds into round two.

Then came Ankalaev. He beat Pereira by unanimous decision at UFC 313 in March 2025. Both went straight into UFC 320 seven months later, where Pereira reclaimed the belt after 1:20.

One fighter supplied a strict reversal, a later repeat, and another strict reversal. “Good in rematches” is not a fixed trait. Each opponent creates a new problem.

Does a 68.8% repeat rate equal a 68.8% prediction?

No. It is a group average from 16 decided, selected title bouts. It is not a ready-made forecast for the next rematch.

Suppose one champion closes as a heavy favorite after a clear first win. Another closes near even money after a disputed decision. Giving both 68.8% would discard useful information in the matchup and price.

The rate also mixes divisions, eras, methods, ages, and preparation windows. A five-round decision followed six months later differs from a stoppage revisited after injuries or a rushed camp.

For betting education, convert current odds into implied probability and remove the bookmaker margin. Compare that estimate with a fresh forecast. Record the price available at the time. A historical rematch rate alone does not show value or promise profit.

What should a UFC rematch prediction include?

Start with the previous bout, then test which parts can survive.

• Did the winner control several phases or land one decisive sequence?
• Was the result a clear decision, a close score, a stoppage, or an injury?
• How long is the gap, and did either fighter compete since?
• Have age, health, weight class, camp, or preparation changed?
• What adjustment can the former loser make without opening a new weakness?
• Do current odds already price the strongest lesson from the first bout?

An AI model should treat the prior meeting as evidence, not an automatic pick. It can pair that result with pre-fight age, reach, stance, form, layoff, division, and market probability.

Testing must follow time. Train on information available before a date, then test on later fights. Otherwise, future records can leak backward and inflate apparent accuracy.

Probability calibration matters too. Yet immediate title rematches are so rare that any dedicated model will carry wide uncertainty. Seventeen cases cannot support many fine-grained rules.

How solid is this data?

The strict sample is transparent and easy to define. ESPN supplied a 16-case modern-era table. The official UFC 320 result supplies the seventeenth case. The broader counts from FightAlpha, Carroll, and FightMetric point in the same direction.

The limits are large. Seventeen immediate title rematches make a small sample. One ended in a draw. Several rivalries contribute more than one bout, so the observations are not fully independent.

Matchmaking creates selection bias. The raw record does not adjust for closing odds, age, injury, layoff, short notice, division, or how the preceding fight ended. It also starts in November 2000 and excludes ordinary UFC rematches.

The broader sources use different cutoffs and definitions. FightAlpha counts every later UFC meeting. Carroll counts title-fight rematches. ESPN requires the next bout for both fighters after a title change. Their percentages are context, not numbers to pool.

The defensible conclusion is narrow: in this modern-era immediate title sample, the prior winner won 11 of 16 decided rematches. That is useful evidence. It is not proof of a causal edge.

FAQ

How often does the first winner win an immediate UFC rematch?

The winner of the preceding fight won 11 of 16 decided modern-era UFC immediate title rematches through UFC 320, a 68.8% rate. One additional rematch ended in a draw.

How often do UFC champions win their belt back immediately?

Former champions reversed the result in 5 of 16 decided immediate title rematches in this sample, or 31.3%. The full 17-case record also includes one draw.

What qualifies as an immediate rematch in the UFC?

Under the strict definition used here, the title changes hands and the same two fighters meet in their next bout. If either fighter competes against someone else first, the later meeting is not immediate.

Are UFC title rematches different from non-title rematches?

FightAlpha found previous winners repeated in 66.7% of 87 title rematches, compared with 54.3% of 105 non-title rematches. Selection and matchup quality differ, so this does not prove the belt causes the gap.

Should the first fight decide a UFC rematch prediction?

No. Use it as one strong input. Recheck the method, competitive phases, time gap, age, health, camp, weight class, recent form, and current implied probability.