# Blood, Sweat and Submissions Source: https://fightstats.online/insights/they-stop-trying Published: 2026-08-13 > Submissions dry up as a fight wears on — round one produces four times as many as round five — and everyone blames the sweat. The attempt rate falls by more than half; the finish rate per attempt falls by six points. The choke did not stop working. Fighters stopped reaching for it. --- Submissions · every UFC fight since 1994 Submissions vanish as a fight wears on — round five runs at a quarter of round one’s rate — and everyone blames the sweat. The sweat is real, and it is nowhere near big enough. What actually empties out of the late rounds is something a fighter chooses. The question ### Are submissions easier to finish early in a fight, when there is less blood and sweat? The received wisdom As a fight draws on, the fighters become bloodier and sweatier, and your chances of finishing a submission diminish greatly. TLDR; Quick Facts 10.23 → 2.48 submissions per 100 fights still live · round one to round five The thing everybody has noticed, and it is real. A fight still going in the fifth ends by submission at a quarter of the rate it would have in the first. 32.6% → 26.9% attempts that finish · rounds 1–2 against 3–5 · z = 2.73 So the wisdom is not wrong. A submission attempted late does convert less often than one attempted early — by 5.7 points, which is a real difference. 5.7 points the entire size of the sweat effect But look at what that is next to. Conversion falls by about a sixth. Submissions themselves fall by three quarters. The slippery-neck theory cannot carry the gap. 0.259 → 0.111 submission attempts per fight · round one to round five Here is what actually collapses. They stop trying. Attempt volume more than halves while the chance of any given attempt working barely moves. 27.1% round three conversion, against 32.3% in round one Which makes the late rounds underpriced rather than closed. The choke in round three is nearly as dangerous as the one in round one. It simply arrives less often. In short Blood and sweat are real and they are minor. Attempts late in a fight do convert less often, by about five points in a hundred — but submissions fall away four times faster than that, and the difference is a decision rather than a physical limit. Fighters stop looking. The neck does not get much slipperier; the fight gets treated as a striking contest being scored. Round one against round five ### One live fight in ten. Then one in forty. Yes — a little. An attempt made late genuinely is harder to finish than one made early, and that part of the wisdom survives the arithmetic. It is the size that does not. Nowhere near enough of the collapse in submissions is failure; almost all of it is fighters attempting fewer. First, the collapse itself. Counting submissions against the fights that actually reached each round — because fewer fights get to round three, and a raw count would fall for that reason alone: **Submissions per 100 fights still live, by round: round one 10.23, round two 8.87, round three 5.77, round four 4.83, round five 2.48.** - Round 1 10.23 792 of 7,743 - Round 2 8.87 482 of 5,434 - Round 3 5.77 238 of 4,124 - Round 4 4.83 16 of 331 - Round 5 2.48 7 of 282 That is a real and steep decline — from roughly one submission in every ten live fights to one in forty. The question is what causes it. The sweat theory, weighed ### Six points off the finish. More than half the attempts gone. A submission needs someone to go for it and then to finish it. If sweat and fatigue were the story, the second of those would be where the damage shows: same number of attempts, fewer of them working. It is the first that moves. - −57% fewer attempts per fight, round one to round five 0.259 attempts per fight in round one, 0.111 in round five. - −5.7 points of finish rate, rounds 1–2 against rounds 3–5 32.6% of attempts finished early, 26.9% late. z = 2.73 — real, but modest. Both are true, and the sizes are what matter. The chance that a submission attempt works falls by about six points across the fight — a genuine effect, and the closest thing here to evidence for the sweat theory. But the number of attempts falls by more than half. The overwhelming reason you see fewer submissions in round four is that almost nobody is trying one. Attempt counts come from per-round statistics, which are only reliable from 2015, so that side of the comparison is held inside the modern era. The round-by-round submission counts above use the full history, because method and round are recorded for every fight back to 1994. UFC Analytics · Archive Edition UFC 31: Locked and Loaded · Atlantic City, New Jersey · 3 May 2001 ### CARLOS NEWTON FINDS THE BULLDOG IN THE THIRD The first late submission anyone wrote down, and it took a welterweight title off Pat Miletich. Two minutes and fifty seconds into the third round, Carlos Newton locked a bulldog choke on the ground and Pat Miletich’s reign ended. It is the earliest round-three-or-later submission in this database — the first time the record shows a fight going long and the choke still arriving. Twenty-five years on it remains the exception it was then: only 261 of the men’s submissions have come in round three or beyond, against 1,274 in the first two. Editor’s note: records before 2008 are thin, so “first recorded” is not always “first”. Method and round, though, are held for every fight back to UFC 1. Charles Oliveira · 2010 to 2021 ### Five late submissions, and nobody else has more Five submissions in round three or later — more than anyone in the history of the promotion. The first came in September 2010. The most recent came in December 2021. Eleven years of being the man still hunting a finish at a point in the fight where the rest of the division has quietly switched to winning rounds. He is not evidence that the late submission is easy. He is evidence of what the rest of this story implies: the opportunity is still there in round three, and it mostly goes unclaimed because nobody reaches for it. Men and women ### A women’s five-round fight is rare. A submission inside one is rarer. The women's divisions show the same shape — submissions concentrated in the opening rounds and thinning out after — but the counts behind the late rounds are small enough that the round-by-round figures cannot be separated from one another with any confidence. A women's five-round fight is rare, and a submission inside one rarer still. So the honest position is that the headline finding is established in the men's divisions and consistent with, but not demonstrated by, the women's. Where a sample cannot carry a claim, it should not be made to. The Training Room What this means for your training ### Train for the third-round choke nobody is expecting Blood and sweat are real, and they are minor. What collapses late in a fight is not the chance of a submission working — it is the number of times anybody tries one. #### Do not write off the back half Your attempt is about six percentage points less likely to finish in rounds three to five than in rounds one and two — 26.9% against 32.6%. That is a real penalty and it is a small one. More than a quarter of late attempts still end the fight, and you will be trying it against a division that has largely stopped defending it because it has largely stopped seeing it. #### Weight the late rounds toward the ground, not away from it The instinct in a tired third round is to treat the fight as a striking contest being scored. The finish rate says the choke is nearly as dangerous as it was in round one; what has changed is that it arrives less often, which is exactly what makes it a surprise when it does. #### What not to change: your grip work The sweat-and-blood theory would justify hours of training on holding positions when everything is slick. The evidence for it here is a six-point drop in conversion — present, but far too small to justify reorganising a camp around. The thing that actually collapses is attempt volume, and that is a decision, not a physical limit. Which is the useful version of the received wisdom rather than the flattering one. Nobody is being stopped by a slippery neck. They are being stopped by deciding, somewhere in the third round, that the fight is now a different kind of fight. --- ## Data Structured claims: https://fightstats.online/insights/they-stop-trying/data/findings.json Underlying query results: - https://fightstats.online/insights/they-stop-trying/data/attempts-and-conversion.json - https://fightstats.online/insights/they-stop-trying/data/early-versus-late.json - https://fightstats.online/insights/they-stop-trying/data/emblem-late-finishers.json - https://fightstats.online/insights/they-stop-trying/data/freeze-stamp.json - https://fightstats.online/insights/they-stop-trying/data/hazard-by-round.json - https://fightstats.online/insights/they-stop-trying/data/origin-first-late-submission.json - https://fightstats.online/insights/they-stop-trying/data/women-hazard-by-round.json --- Analysis and prose licensed CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Underlying fight records sourced from ufcstats.com. Method: /data.