# Survive the Round Source: https://fightstats.online/insights/survive-the-round Published: 2026-08-08 > A knockdown looks like the end of a fight, and 86% of the time it is — but the fighters who hear the bell before the referee waves it off are in a very different fight from the one the highlight reel suggests. --- Knockdowns · 3,005 fights · 1994–2026 A knockdown looks like the end of a fight, and 86% of the time it is — but the fighters who hear the bell before the referee waves it off are in a very different fight from the one the highlight reel suggests. - 86.1% the knockdown ends it there and then - 34.4% get up, hear the bell, and win anyway Is a knockdown the end of the fight? Eighty-six percent of the time, yes. But that figure hides the only part of it a fighter can do anything about: if the round ends before the referee does, the fighter who was on the floor goes on to win better than one time in three . The number everyone quotes ### Eighty-six percent, and a hidden seam Across every UFC fight where the statisticians recorded a knockdown, the fighter who scored the first one went on to win 86.1% of the time. On 3,005 fights that figure is about as solid as this sport produces — the range consistent with it runs 84.9% to 87.3%. It is also two completely different numbers wearing a coat. Some of those knockdowns ended the fight immediately: the referee stepped in, and the knockdown was the finish. The rest didn’t. The fight carried on, and in some cases carried on for another twenty minutes. Averaging those together produces a statistic that describes neither. The bell ### Everything depends on whether the round ends first Split those 3,005 fights by one question — did the fight get past the round the knockdown happened in? — and the single statistic comes apart into two that barely belong on the same page. Red: the fighter who scored the knockdown. Bone: the fighter who was on the floor. If the referee lets it go and the round runs out, the knockdown stops being a verdict and becomes a scorecard entry. Ninety-four point nine percent falls to sixty-five point six. Both intervals are tight and nowhere near each other, so this is not a subtlety — it is the largest single split anywhere in these stories. Put the other way round: better than one in three fighters who get dropped and hear the bell go on to win the fight. Three hundred and thirteen of them, in a sport that treats a knockdown as a full stop. Nothing about the knockdown changed. What changed is that somebody got ten seconds. The 313 ### A quarter of them come back and finish it The obvious assumption about a survivor’s win is that it is a grim, clinging, points-based affair. Most of them are. Not all. - Won by decision 179 57.2% of survivors’ wins. Dropped, recovered, and out-worked them over the remaining rounds. 57.2% of survivors’ wins. Dropped, recovered, and out-worked them over the remaining rounds. - Won by KO/TKO 83 26.5%. Knocked down, got up, and knocked out the fighter who had just put them there. 26.5%. Knocked down, got up, and knocked out the fighter who had just put them there. - Won by submission 49 15.7%. Two more wins came by other routes, which is where the last 0.6% goes. 15.7%. Two more wins came by other routes, which is where the last 0.6% goes. Composition of the 313 fights won by a fighter who was knocked down and survived the round. Eighty-three fighters have been knocked down, got back up, and then stopped the person who dropped them. That is not a rounding error in a dataset of three thousand — it is roughly one such fight every four months across the promotion’s history. A trap worth naming ### More knockdowns, fewer wins Knock a fighter down once and you win 87.4% of the time. Twice, and it falls to 79.8%. Three or more, and it falls again to 77.6%. Read carelessly, that says knocking someone down repeatedly is bad for you. It says nothing of the sort. A fight only contains a second knockdown if the first one failed to end it — so this table is not measuring the effect of extra knockdowns, it is measuring the fights where knockdowns weren’t working. The causation runs backwards through the number. It is included here because it is exactly the shape of finding that gets written up as an insight, and the arithmetic that would catch it — the difference between 87.4% and 79.8% is real, 7.6 points with a range of 3.5 to 11.7 — does not catch it at all. A real difference can still be a meaningless one. Win rate for the fighter who scored the first knockdown, by how many the fight contained. Scale runs 60% to 100%; the mark is at 70%. When it lands ### The round barely matters A knockdown in round one is worth 85.2% to the fighter who scores it. Round two, 87.9%. Round three, 86.7%. Three intervals that sit almost entirely on top of one another — there is no evidence here that an early knockdown means less than a late one, which is not what most commentary assumes. Among the survivors there is a hint of something. A fighter dropped in round one goes on to win 36.8% of the time; dropped in round two, 29.5%. That gap is 7.3 points and its range runs from about half a point to fourteen — it clears zero, but only just, and it rests on 627 and 251 fights. Worth one sentence, not a section. Rounds four and five carry 19 and 15 fights between them. Nothing can be said about championship rounds from that, and this page does not try. Win rate for the knockdown scorer once the fight went past that round. Scale 40% to 90%; the mark is the overall survivor-fight figure of 65.6%. The emblem ### Four times off the floor, four times the winner Nobody in the database has been knocked down and still won more often than Michael Bisping — four times, from a UFC record of 30–9–0. He shares the mark rather than owning it: Dan Henderson and Diego Sanchez are also on four, and no one has managed five. A shared record is the right shape for this story. The point of the 313 is not that a handful of freaks can absorb a knockdown; it is that surviving one is an ordinary, repeatable thing that happens roughly a third of the time the bell allows it. Three fighters reaching four apiece, and nobody reaching five, is what a common event looks like at the tail. Behind them, on three: Neil Magny, Cezar Ferreira, Mateusz Gamrot, Eryk Anders, Matt Hughes. Between them these eight fighters account for twenty-seven of the 313 — under a tenth, which is the point. UFC Analytics · Archive Edition #### DAN HENDERSON HITS THE FLOOR TWICE, WINS TWICE The first two recoveries in the promotion’s records happen on the same night, to the same man, in the same tournament. UFC 17 gives the archive its earliest recorded knockdown-and-recovery, and then immediately gives it a second. Henderson is dropped in the first round by Allan Goes and wins on the judges’ cards. Later the same night he is dropped in the first round by Carlos Newton and wins again, this time by split decision. Two knockdowns, two bells, two wins, one evening. Kevin Randleman follows eighteen months later at UFC 23, dropped in the first by Pete Williams and still standing at the end of the fifth. What looks in the highlight reel like the end of a fight has, from the very first night it was written down, been something a fighter could walk out of. Editor’s note: earliest recorded recoveries. Round-by-round statistics exist from March 1994 onwards and cover 8,602 of 8,641 decided fights — but a knockdown appears in only 43 of the 244 scored fights before 2001, against 2,000 of 5,590 since 2015. Whether early crowds saw fewer knockdowns or early statisticians wrote down fewer of them, this page cannot say. One thing that looks like a finding Recovery rates by division run from 59.3% at Middleweight to 72.4% at Light Heavyweight — thirteen points, which would make a tidy paragraph about heavier fighters being harder to put away for good. The paragraph would be unfounded. Those figures rest on 118 and 58 fights, and the range consistent with a thirteen-point gap between them runs from about minus one to plus twenty-eight. It contains zero. Every division here overlaps every other one, and the ordering below is presented as a list rather than a ranking for that reason. Win rate for the knockdown scorer in fights that continued. Scale 40% to 90%; the mark is the overall figure of 65.6%, which every division’s interval contains. A knockdown is a judgement, not an event. There is no bright line between a knockdown and a slip, a stumble, or a fighter dropping level for a takedown and getting clipped on the way down. The count reflects whoever was scoring that night, and every figure on this page inherits that. Coverage. Round-by-round statistics cover 8,602 of 8,641 decided fights, from March 1994 onward, and 3,051 of those fights contain at least one recorded knockdown. The analysis uses the 3,005 of them that produced a winner. The early years are thin in a specific way. Only 43 of the 244 scored fights before 2001 carry a knockdown, against 2,000 of 5,590 since 2015. That is a large enough difference in recording rate that era comparisons are not attempted anywhere on this page. Survival is not the same as recovery. “The fight went past the round” is a fact about the referee and the clock as much as about the fighter. A knockdown with ten seconds left on it and a knockdown with four minutes left are both in here, and this data cannot separate them. Men and women ### Too few women’s fights to say — which is the answer Women reach a knockdown far less often than men, so the subset this whole story rests on is very thin in their divisions. The honest report is the width of the interval, not the number in the middle of it. #### Scorer wins once the round is survived 860 fights. This is where the 34.4% survival figure comes from, and it is solidly measured. #### Fifty fights An interval running 57.3 to 82.7 — wide enough to contain the men’s figure, and wide enough to contain almost anything else. The difference is −4.7 points with a range of −17.8 to 8.4, so nothing here separates the two games — but nothing here would have. Fifty fights could only have detected an enormous divergence, and the honest statement is that the women’s game has not been tested rather than that it agrees. Everything below was measured in the men’s divisions. It is likely to carry across, because the mechanism is a referee and a clock rather than anything about the fighters, but this page has not shown that. What this changes ### Drill the ten seconds, not the knockdown The knockdown itself is not the trainable part. The stretch immediately after it is, and it is worth more than almost anything else on this page. #### Surviving the round as its own skill Once a fight gets past the round the knockdown happened in, the fighter who was dropped wins 34.4% of the time. That is a specific, rehearsable scenario — covering, moving, tying up, buying the clock — and 313 fighters have converted it. Almost nobody trains it deliberately, and it is the highest-value ten seconds in the sport. #### If you score one, the window is now A knockdown that ends the round ends the fight 94.9% of the time. One that doesn’t falls to 65.6%. Nearly thirty points of outcome sit in the seconds between those two states, which is an argument for finishing discipline in exactly the moment adrenaline makes it hardest. #### That the survivor is just hanging on Of the 313 fighters who got up and won, 83 did it by knockout — a quarter of them stopped the fighter who had just dropped them, and another 49 submitted them. Surviving a knockdown is not a slow, clinging loss deferred. In half of these fights it turned into a finish going the other way. One limit worth respecting before any of this becomes a drill: “the fight went past the round” is a fact about the referee and the clock as much as about the fighter. A knockdown with ten seconds left in a round and one with four minutes left are both counted here, and this data cannot separate them — so the 34.4% is the value of getting to the bell, not proof that any particular fighter could have got there. --- ## Data Structured claims: https://fightstats.online/insights/survive-the-round/data/findings.json Underlying query results: - https://fightstats.online/insights/survive-the-round/data/by-the-round-it-happened.json - https://fightstats.online/insights/survive-the-round/data/coverage-by-era.json - https://fightstats.online/insights/survive-the-round/data/coverage-round-by-round-data.json - https://fightstats.online/insights/survive-the-round/data/did-the-bell-save-them.json - https://fightstats.online/insights/survive-the-round/data/does-a-second-knockdown-settle-it.json - https://fightstats.online/insights/survive-the-round/data/emblem-most-often-dropped-and-won.json - https://fightstats.online/insights/survive-the-round/data/exact-counts.json - https://fightstats.online/insights/survive-the-round/data/how-the-survivors-win.json - https://fightstats.online/insights/survive-the-round/data/men-vs-women.json - https://fightstats.online/insights/survive-the-round/data/origins-earliest-recoveries.json - https://fightstats.online/insights/survive-the-round/data/recovery-by-weight-class.json - https://fightstats.online/insights/survive-the-round/data/survivors-by-round-they-were-dropped.json - https://fightstats.online/insights/survive-the-round/data/the-headline.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.