Skip to content
UFC ARCHIVE ANALYSIS / TWO QUESTIONS

How the fight changed.

One archive, two studies: what three decades changed inside the cage, and what changes when the same two fighters meet again.

Fights
8,806
Fighter-rounds
41,488
Coverage
1994–2026
01
THE LONG VIEW / 987 EARLY · 1319 RECENT / DIVISION-STANDARDIZED

How has the way UFC fights are fought changed over time?

The fight moved outward.

Compared with 2006–2010, modern fights carry one-third more significant-strike attempts, land more accurately, and place 87% of those attempts at distance. Ground and clinch striking receded; the finish rate fell only four points.

More striking volume
+33.5%
6.31 → 8.43 attempts per fighter minute
Now thrown at distance
87.2%
Up 16.6 percentage points
More accurate
+7.2 pts
41.4% → 48.6% landed
Fewer submission attempts
-66.7%
1.02 → 0.34 per 15 fighter minutes
ANNUAL ARCHIVE / ALL DIVISIONS

Follow the change year by year.

Distance88.0% of significant attempts
95.377.259.040.822.71994: 29.7 % of significant attempts; 29 fights1995: 52.4 % of significant attempts; 34 fights1996: 37.0 % of significant attempts; 36 fights1997: 36.8 % of significant attempts; 38 fights1998: 55.0 % of significant attempts; 21 fights1999: 51.1 % of significant attempts; 44 fights2000: 44.1 % of significant attempts; 43 fights2001: 54.7 % of significant attempts; 40 fights2002: 53.7 % of significant attempts; 53 fights2003: 51.5 % of significant attempts; 41 fights2004: 51.4 % of significant attempts; 39 fights2005: 60.4 % of significant attempts; 80 fights2006: 68.0 % of significant attempts; 158 fights2007: 65.5 % of significant attempts; 171 fights2008: 68.0 % of significant attempts; 201 fights2009: 72.7 % of significant attempts; 215 fights2010: 76.4 % of significant attempts; 253 fights2011: 79.1 % of significant attempts; 300 fights2012: 82.8 % of significant attempts; 341 fights2013: 79.9 % of significant attempts; 386 fights2014: 80.1 % of significant attempts; 503 fights2015: 80.1 % of significant attempts; 473 fights2016: 82.3 % of significant attempts; 493 fights2017: 83.8 % of significant attempts; 457 fights2018: 84.0 % of significant attempts; 474 fights2019: 85.9 % of significant attempts; 516 fights2020: 86.4 % of significant attempts; 456 fights2021: 88.1 % of significant attempts; 509 fights2022: 87.3 % of significant attempts; 511 fights2023: 86.4 % of significant attempts; 520 fights2024: 88.3 % of significant attempts; 516 fights2025: 88.0 % of significant attempts; 517 fights1994200020052010201520202025
2006–2010 comparison window 2021–2025 comparison window2026 excluded — incomplete seasonHover points for annual values
Read the annual values
YearFightsDistance
19942929.7 % of significant attempts
19953452.4 % of significant attempts
19963637.0 % of significant attempts
19973836.8 % of significant attempts
19982155.0 % of significant attempts
19994451.1 % of significant attempts
20004344.1 % of significant attempts
20014054.7 % of significant attempts
20025353.7 % of significant attempts
20034151.5 % of significant attempts
20043951.4 % of significant attempts
20058060.4 % of significant attempts
200615868.0 % of significant attempts
200717165.5 % of significant attempts
200820168.0 % of significant attempts
200921572.7 % of significant attempts
201025376.4 % of significant attempts
201130079.1 % of significant attempts
201234182.8 % of significant attempts
201338679.9 % of significant attempts
201450380.1 % of significant attempts
201547380.1 % of significant attempts
201649382.3 % of significant attempts
201745783.8 % of significant attempts
201847484.0 % of significant attempts
201951685.9 % of significant attempts
202045686.4 % of significant attempts
202150988.1 % of significant attempts
202251187.3 % of significant attempts
202352086.4 % of significant attempts
202451688.3 % of significant attempts
202551788.0 % of significant attempts
THE REST OF THE PATTERN

More output did not create more chaos.

The modern fight became cleaner and more spatial. More strikes are attempted, but exchanges migrated away from the clinch and ground while knockdown rates remained essentially flat.

On the ground% of significant attempts
15.8THEN
6.6NOW
-9.2
From the clinch% of significant attempts
13.6THEN
6.2NOW
-7.4
Control share% of fight time
53.5THEN
36.7NOW
-16.8
Body targeting% of significant attempts
10.8THEN
14.2NOW
+3.5
Takedown pressureper 15 fighter minutes
4.42THEN
3.72NOW
-0.7
Finished before the cards% of resolved fights
60.7THEN
56.4NOW
-4.3
NOT A DIVISION-MIX ILLUSION

The direction holds across all five divisions.

Each long-running division became faster, more accurate, and more distance-oriented. Finish rates varied; the underlying style shift did not.

DivisionPaceAccuracyDistanceFinish rate
Heavyweight+1.46+5.1 pts+20.6 pts-15.0 pts
Light Heavyweight+1.41+9.5 pts+13.1 pts-2.1 pts
Middleweight+2.48+5.8 pts+19.7 pts-5.5 pts
Welterweight+2.37+5.1 pts+19.0 pts-4.0 pts
Lightweight+2.30+10.2 pts+11.8 pts+0.6 pts
How this comparison works+

Five complete seasons are compared with five recent complete seasons. Both windows are standardized to the same pooled division distribution across heavyweight, light heavyweight, middleweight, welterweight, and lightweight. Uncertainty uses 2,000 event-cluster bootstrap samples.

  • The early tournament era is shown descriptively but excluded from the headline era comparison because rules, bout duration, and sample sizes differ sharply.
  • The archive includes 2026 events through August 15, but 2026 is excluded from annual trends and era comparisons because the season is incomplete.
  • Recorded control time is missing before part of 1999; missing values are never interpreted as zero.
  • Changes describe the UFC archive and do not identify a single causal mechanism such as training, rules, matchmaking, or judging.

Every displayed interval is a 95% event-cluster bootstrap interval. The annual chart is descriptive and includes all divisions; the headline comparison fixes the division distribution across eras.

02
200 REPEATED MATCHUPS / FIRST REMATCH ONLY

Do UFC rematches produce fundamentally different fights than first meetings?

Different—but not wilder.

Rematches do not reliably become faster, longer, or more likely to finish. The repeatable change is precision: significant-strike accuracy rises by about five points while the broader tactical shape changes less than it normally does against a new opponent.

A different winner
40.8%
71 of 174 matchups with a winner both times
Higher accuracy
+4.9 pts
43.9% → 48.8% landed
Change in pace
+0.01
No meaningful shift in significant-strike volume
Change in finish rate
+0.5 pts
The route to the result barely changes
THE REPEATABLE SIGNAL2 IN 3

rematches were more accurate.

Accuracy increased in 66.7% of the 141 tactically comparable pairs. The average gain remained 4.3 points when both meetings had the same scheduled length.

Significant-strike accuracy% landed
FIRST43.9
REMATCH48.8
N=141 paired matchups · 95% CI for change: 2.8 to 7
Significant-strike attemptsper fighter minute
FIRST8.79
REMATCH8.8
N=141 paired matchups · 95% CI for change: -0.69 to 0.71
Recorded control time% of fight time
FIRST34.4
REMATCH34.2
N=141 paired matchups · 95% CI for change: -4.8 to 4.3
Inside-distance result% of resolved fights
FIRST52.4
REMATCH53
N=185 paired matchups · 95% CI for change: -8.6 to 9.7
THE FAMILIARITY TEST

Same opponent, smaller tactical rewrite.

Across 138 eligible repeated matchups, a style-change score combines pace, accuracy, target mix, position, takedowns, submissions, and control. Each rematch is compared with a same-fighter transition across the same number of career fights, in the same divisions, against different opponents.

SAME OPPONENT0.78REMATCH STYLE DISTANCE
DIFFERENT OPPONENT0.92MATCHED CONTROL DISTANCE
-14.2%less tactical change in a rematch95% CI -0.2 to -0.05 standardized units
HOW THE RESULT ARRIVED

The ending was almost a coin flip between repeating and changing.

REMATCH: DECISIONREMATCH: INSIDE DISTANCEFIRST: DECISION
50same route
38changed route
FIRST: INSIDE DISTANCE
37changed route
60same route

Fifteen pairs include a no contest or another unresolved meeting and are excluded from this matrix. The aggregate inside-distance rate moved from 52.4% to 53%.

How the rematch study works+

The primary contrast pairs each matchup's first UFC meeting with its first UFC rematch. Outcome and duration use every eligible pair. Tactical estimates require both fights to last at least three minutes, at least 20 combined significant attempts in each fight, and recorded control data. Pair-cluster bootstrap intervals use 2,000 samples.

Within-fighter tactical change across the first two meetings, compared with the closest same-fighter, same-career-gap transition against different opponents in the same divisions.

  • A first UFC meeting may not be the fighters' first meeting in another promotion.
  • Rematches are selected by matchmaking and are disproportionately competitive, consequential, or commercially valuable; estimates are descriptive, not causal.
  • Third and fourth meetings are counted in archive coverage but excluded from the primary first-versus-second comparison.
Sensitivity checkLooser threshold: -13.7% (170 pairs; CI -0.2 to -0.06)Primary threshold: -14.2% (138 pairs; CI -0.2 to -0.05)Stricter threshold: -20.2% (106 pairs; CI -0.26 to -0.12)