Age · 8,641 decided fights · 1993–2026
The Undefeated Opponent
Every fighter eventually loses to the same opponent, and the tape is unusually clear about what it takes first — not the offence, which barely moves, but everything on the other side of the ledger.
What is the first thing to go when a fighter gets old?
Not the offence. Fighters past forty land no fewer significant strikes than fighters under twenty-five, and land them more accurately. What climbs by half is what comes back the other way — and that turns out to change what an ageing fighter should actually be training.
The arithmetic
Youth is worth a fixed number of points
Put two fighters in a cage and the younger one wins more often. That much is unsurprising. What the data adds is a price list.
Compare the two fighters inside each bout rather than across the roster, and the question answers itself cleanly — there is no era to control for, no weight class, no matchmaking to argue about. One is older. Did the other one win?
The edge scales almost linearly with the gap. A year or two is close to nothing. A decade is close to decisive.
Share of fights won by the younger fighter. Bars measured from 50% — the mark where age buys nothing.
What actually goes
The offence never leaves
Here is the part that ruins the obvious explanation. Fighters do not get old and stop landing. They get old and stop avoiding.
Under 25 · 1,542
They land 32.6 significant strikes a fight and absorb 29.4. The youngest fighters on the roster are the only group that gives less than it takes.
25 to 29 · 6,555
Output jumps to 36.3. This is the last age on the roster where the ledger still favours you.
30 to 34 · 6,625
37.4 landed against 37.9 absorbed — half a strike apart, which is another way of saying level. This is the band the lines meet in, not a moment you could put a date on.
40 and over · 265
Still landing 34.5 — as much as any fighter under 25 manages — and now absorbing 44.2. Accuracy is at its career best, 48.0%.
Read the two lines separately and they tell opposite stories. Output climbs through the twenties, settles around 37 significant strikes a fight, and simply stays there — a fighter past forty still lands 34.5, as many as a fighter under 25 manages. The two are two strikes apart, and on this sample that cannot be told from no gap at all: the interval runs from −2.3 to +6.3. What the tape supports is that output does not fall, not that it climbs. Accuracy doesn’t decay either. It improves, from 43.6% under 25 to 48.0% past forty, which is what three extra title camps of shot selection buys you.
The other line never stops climbing. Strikes absorbed run 29.4, 34.6, 37.9, 41.2, 44.2 — half as much again, end to end, and a fifteen-strike gap between the youngest and oldest bands that no amount of sampling error explains away. Takedown defence slides the same way, 63.7% at its peak in the late twenties down to 57.2% past forty. Nothing about the older fighter’s offence has failed. They are simply being hit half as much again while doing it.
What the data will not support is a precise crossing. In the 30–34 band the two lines sit a quarter of a strike apart, and the range consistent with that sample runs from one and a half strikes either way — wide enough to contain the crossing happening a band earlier or a band later. The divergence is the finding. The exact meeting point is not.
Getting old in a cage doesn’t look like losing your weapons. It looks like losing the answer to everyone else’s.
The check
The same man, ten years earlier
Everything above compares one age group to another, and there is an obvious objection to it: the men still fighting at forty are the men who could. The ones the sport broke are not in the forty-year-old column, because they are not in any column.
So here is the same question asked without them. Every fighter is measured against their own record at 27 to 32, which means a fighter who declined and retired at 34 still counts — his decline is inside his own line rather than missing from the sample. 15,365 fighter-fights, men’s divisions, bootstrapped over fighters.
The story holds, and it is not survivorship. Against his own prime, a fighter absorbs 0.48 more significant head strikes a minute at 33–34, 0.67 more at 35–36, 1.10 more at 37–38 and 1.88 more at 39 and over — against a prime rate of 2.43 a minute. That last figure is a 77% increase in how often he is hit, and every one of the four intervals clears zero.
The knockout column follows it exactly, as the argument says it should: losing by knockout runs +6.8, +11.7, +15.2, +25.8 percentage points above the same man’s prime, clearing zero at every step.
What does not go
The other thing everybody says about old fighters is that they fade — that the tank empties and the third round is where age shows. It does not. Measured as round-three output minus round-one output, the same fighter holds his pace at every age: +2.1 [−0.3, +4.4] at 33–34, +2.3 [−1.3, +5.2] at 35–36, −3.8 [−10.1, +1.4] at 37–38, +3.9 [−2.8, +11.0] past 39. Four bands, four intervals through zero, no trend.
A thirty-nine-year-old throws as much in the third round, relative to his first, as he did at twenty-eight. He is simply being hit twice as often while he does it.
Two limits, stated rather than buried. These figures are not adjusted for who the fighter was matched against, and opponent quality drifts with age — controlling for it shrinks the decline in win rate substantially, so read the shape here rather than the exact size. And a fighter’s age advances with the calendar, so an age effect and the sport’s own drift cannot be fully separated within one career.
One question this page deliberately does not answer: whether the chin itself weakens. The sharp test is knockdowns absorbed per hundred head strikes taken, which holds exposure constant. Measured that way the estimate rises with age but only two of four bands clear zero, and a model that also controls for opponent and era finds it flat. Two defensible methods disagree, so the honest answer is that we do not know. What is not in doubt is the exposure: whatever his chin is worth at forty, it is being tested twice as often.
The bill
Age charges the most where damage decides it
If the thing that goes is the ability to absorb, then age should cost most in the fights decided by absorbing. It does — and the gap between the three endings is wider than the effect it sits inside.
Share won by the younger fighter, by how the fight finished.
Hand an older fighter fifteen minutes and a scorecard and they lose 54.8% of the time. Hand them a fight that ends in damage and it climbs to 60.2%. Five and a half points is the difference between the two ways of losing to a younger opponent, and it points at the same place the tape does.
Somewhere in the early thirties
The crossing you can see but not locate
Every fight in the database, sorted by how old the fighter was on the night. The bars are how many. The line is how often they won.
Bars: fights contested at each age. Line: win rate. Ages 20–43, minimum 20 fights.
Thirty-one is the first age at which UFC fighters lose more fights than they win — 49.4% across 1,552 of them. It is tempting to call that the tipping point, and it would be wrong. On 1,552 fights the interval around 49.4% runs from 46.9% to 51.9%, which comfortably contains a fifty-fifty split. So does twenty-nine’s, and thirty’s, and thirty-two’s. Every single year from twenty-seven to thirty-five is consistent with a fighter winning exactly half.
The crossing is real — the under-25s win 57.2% and the over-40s 33.6%, and no amount of uncertainty bridges twenty-four points. What cannot be done is naming the year it happens in. The honest statement is “somewhere in the early thirties”, and anyone who tells you a fighter falls off a cliff at a specific birthday is reading a single year’s wobble as a milestone.
The tail deserves the same scepticism. Forty-one comes in at 39.7% off 58 fights and forty-two at 31.8% off 44 — the wobble there is sample size, not a late-career resurgence. What survives is the direction, which never reverses across two decades of ages, and the size of the fall from one end to the other.
The exception
Five of the fifteen oldest wins belong to one man
Sort every win in UFC history by how old the winner was that night, take the top fifteen, and Randy Couture’s name is on five of them — including the top three, which nobody else has ever come near.
At forty-seven he submitted James Toney at UFC 118 — the oldest win in the database, three years clear of the next name that isn’t his own. He was still winning by decision at forty-six, still finishing by knockout at forty-four. The story every other number on this page tells is the one he spent a decade refusing.
The five oldest winning performances on record. Three of them are the same man, in the same year.
The night in question · the oldest win in thirty-two years of records
Every division, same direction
The effect is not a quirk of one weight class. In all eleven divisions with at least 200 fights on record, the younger fighter wins more often than not. Ten of the eleven do so by a margin their sample can actually support; Women’s Bantamweight, on 225 fights, is the one that cannot rule out an even split.
The order below is not a ranking. Seven points separate the top of the list from the bottom, and the range consistent with those samples runs from zero to fourteen — so the divisions are sorted here, but nothing establishes that age costs a lightweight more than a middleweight.
- Lightweight 59.9%
- Women’s Strawweight 59.2%
- Flyweight 58.4%
- Featherweight 58.3%
- Heavyweight 58.1%
- Bantamweight 58.0%
- Women’s Flyweight 57.9%
- Welterweight 56.2%
- Light Heavyweight 55.3%
- Middleweight 54.7%
- Women’s Bantamweight 52.9% ±6.5
Dimmed: the one division whose interval still contains 50%.
What this is not. A fight is not a controlled experiment. An older fighter is usually also the more experienced one, often the more damaged one, and matchmaking is not random — the promotion decides who meets whom, and it does not decide at random. These numbers describe what happens when a younger fighter meets an older one. They do not isolate why.
What’s missing. Dates of birth are recorded for 2,619 of 2,723 fighters — 96.2%. Fights involving anyone without one are absent from every figure here. Detailed strike and takedown statistics only begin to be kept reliably after the promotion’s first decade, so the tape section leans on the modern era far more heavily than the win rates do.
The thin end. There are 265 fighter-appearances at forty and over, against 6,625 in the thirties. Every figure for the oldest band rests on that smaller number and should be read with it in mind.
Men and women
Age charges both games the same
Of every comparison run across this site, this is the closest two numbers have come to each other.
Younger wins, five-year gap or more
2,600 fights with both dates of birth recorded and a gap of at least five years.
One tenth of a point apart
331 fights. The difference is −0.1 points, with a range from −5.6 to 5.4.
A tenth of a point is closer than the samples could reasonably promise, and some of that is luck — the interval still spans five points either way. But the direction and the size agree completely across two populations with different weight ranges, different eras, different finishing rates and thirteen years against thirty-two of history. Whatever ageing takes from a fighter, it appears to take the same amount regardless of which of these two sports they are in, which makes the tape findings above more likely to be about physiology than about style.
What this changes
Train what leaves, not what stays
The instinctive plan for an ageing fighter is to protect the output. This data says output is the one thing that does not need protecting.
Damage avoidance, not work rate
Fighters past forty absorb 44.2 significant strikes a fight against 29.4 for the under-25s — a fifteen-strike gap far too wide to be sampling error. Their own output holds at 34.5, level with the youngest band rather than below it. Head movement, exits and defensive positioning are what the ledger says decays; conditioning for volume is defending something that is not under attack.
Age is charged in damage, not in rounds
Against an older opponent, the younger fighter wins 60.2% of fights that end in a knockout but only 54.8% of those that reach the cards — a real gap of 5.4 points. If you are the younger fighter, the finish is the higher-percentage route. If you are the older one, a fight that goes the distance is the version that suits you.
Takedown defence, quietly
Roster-wide takedown defence runs 63.7% in the late twenties and 57.2% past forty. It is the clearest measurable skill on this page that fades, and it fades without announcing itself — nothing about the shell feels different until it stops holding.
Read this as a description of the roster rather than of a career. These are different fighters at different ages, not the same fighter followed through time, so “this is what happens to you” is an inference the data does not strictly license. Matchmaking is not random either. And the oldest band rests on 265 fighter-appearances against 6,625 in the thirties.