Tell 05
Back to Back
Two KOs in a row reads like a fighter falling apart. We checked whether the spacing says anything the count doesn't.
Tight or spread out, about one fight in four ends in another KO. The rhythm adds nothing to the count.
- Sample
- Every UFC bout since 2005, 8,980 of them
- Confidence
- 31% less likely to 25% more likely
- The catch
- Half the tight pairs had no match
Re-run 27h ago, after UFC 334: Gane vs. Hokit.
What everyone believes
Justin Gaethje got stopped by Eddie Alvarez in December 2017 and by Dustin Poirier in April 2018, back to back. That's the sequence the sport reads as a fighter coming apart. Then Gaethje fought thirteen more times in the UFC and got stopped by strikes in one of them.
You could call that an exception. The folk version would. Two KOs close together is supposed to be a different animal from two KOs years apart: the first is bad luck, the second in quick succession means the chin is gone and the next one is coming. Commentators say it, and the betting chat fades it on sight.
Chin Decay already found that the first KO loss marks a real change, and that the count after it doesn't compound in any way we could measure. That tested how many. This tests how close together.
The comparison
Share of fights that end in a KO or TKO loss from strikes, for fighters already stopped at least twice.
Spaced out
26.5%
Tight pair
24.5%
8% less likely-31% to 25%
- Age
- Spaced out33.6
- Tight pair33.7
- UFC fights
- Spaced out12.6
- Tight pair12.0
- KO losses
- Spaced out2.5
- Tight pair2.7
- Fights since last KO
- Spaced out1.5
- Tight pair1.7
What we measured
Every UFC fight since 2005, 8,980 bouts across 797 events. A KO loss is a KO or TKO from strikes, the same definition Chin Decay and The Odometer use. Every fight counts as a fight when measuring a gap, so a hole in the stats can never make a pair look tighter than it was.
A tight pair is two KO losses with one fight or fewer between them. A spread out pair has three or more. A gap of exactly two sits in neither group, so the two groups describe two different rhythms.
Each fight is compared against fights from fighters with the same number of KO losses, the same number of fights since the last one, a similar age, a similar UFC record and a similar amount of damage taken. 531 fights after a tight pair, from 207 fighters, against 435 matched fights after a spread out one.
All of it was written down and committed before it ran.
Every way of asking
Four definitions of a tight pair, each against its own matched comparison. The line at 1 is no difference.
Tight vs spaced
531 fights
No buffer
616 fights
By days
375 fights
Speeding up
262 fights
Why there's nothing here
Once a fighter has been stopped twice, they get stopped again about a quarter of the time, and it doesn't matter much how the first two were spaced. Tight pairs run 8% less likely on the headline comparison. Drop the buffer and it's 1% less likely. Measure the gap in days instead of fights and it's 6% more likely. Look only at fighters with three or more KO losses whose latest gap is shorter than the one before, the version that should show it if anything does, and it's 4% more likely across 262 fights.
The count is doing the work. Whatever a second KO says about a fighter, it says it whether it came the next fight or five fights later. A chin doesn't seem to keep a calendar.
Two KOs is two KOs. The gap between them doesn't add a third.
What this can't tell you
The catch is in who got compared. A spread out pair takes at least five UFC fights to build, so fighters who got stopped twice early in their run had nobody spread out to match against. That left 497 fights after a tight pair, 48% of them, with no comparison at all, even after the matching was loosened by a rule written in advance for exactly that. The answer above covers fighters with enough of a career to compare. The young fighter stopped twice in a row in their first few UFC fights is mostly outside it, and plenty of those got cut before they could be measured.
The same survivorship runs through everything else. Every fight after a tight pair belongs to a fighter the promotion kept booking, and the ones it kept are the ones it believed in.
And a null is not proof of absence. The interval runs from 31% less likely to 25% more likely. What it rules out is the confident version, the one where back to back KOs mean the next one is due.
Since we published
Nobody new has fallen into this one yet. Check back after the next card.
Research log
- Re-runRe-ran after UFC 334: Gane vs. Hokit. Verdict no signal. Effect 8% less.
- CardUFC 334: Gane vs. Hokit went into the record. 1 fighter already flagged fought again, and it went the way this Tell says.
- PublishedPublished.
- CardUFC Fight Night: Cortes Acosta vs. Walker went into the record. 4 fighters already flagged fought again, 0 went the way this Tell says.
- CardUFC Fight Night: Moicano vs. Nolan went into the record. 2 fighters joined. 3 fighters already flagged fought again, 1 went the way this Tell says.
- CardUFC 333: Volkanovski vs. Evloev went into the record. 1 fighter joined. 3 fighters already flagged fought again, 1 went the way this Tell says.
9 earlier
- CardUFC Fight Night: Buckley vs. Malott went into the record. 1 fighter joined. 5 fighters already flagged fought again, 2 went the way this Tell says.
Count the KOs. Don't bother with the calendar.