There is a story mixed martial arts likes to tell about itself: that it started as raw, chaotic violence and became a refined striking sport, where technically brilliant athletes knock each other spectacularly unconscious. The submission belonged to the past. The knockout was the future.
We went looking for that story in our own archive — 4,287 individual bouts across 623 documented UFC events, from the first tournament in 1993 through to early 2026 — and found something considerably more interesting. Half of it is true. The other half is exactly backwards.
The submission really did collapse
Start with the part everyone gets right. In the UFC's first seven years, 46.2% of fights in our archive ended in a submission. Royce Gracie's tournament wins were not an anomaly — they were the sport working precisely as its early audience understood it. A grappler who knew something nobody else knew could take your back and end the evening.
By 2016–2020 that figure had fallen to 18.9%, and it sits at 19.4% today. That is a drop of nearly two-thirds.
The submission did not become less effective. It became less surprising. Once every fighter on the roster had a competent guard and a coach who understood position, the specialist's edge evaporated. What had been secret knowledge in 1993 was a white-belt fundamental by 2010.
But the knockout never took over
Here is where the accepted story falls apart. If submissions collapsed, something must have replaced them — and the assumption is always striking. Better boxing coaches, better kickboxing, heavier hands.
The data says no. Knockouts and TKOs accounted for 35.5% of bouts in 1993–2000. In the current era, they account for 35.3%. Across three decades, five eras and more than four thousand fights, the knockout rate has barely moved. It has never risen above 37.8% or fallen below 35%.
Strikers did not inherit the sport. The knockout has been a remarkably stable feature of professional MMA since the beginning.
What actually replaced the submission was the judges
The share that grew — and grew enormously — is the one nobody celebrates. Fights going to the scorecards rose from 17.7% in the earliest era to 45.1% today. That is not a drift. It is a two-and-a-half-fold increase, and it is the single largest change in how UFC fights end.
Some of the early figure is structural: the first events were tournaments, often without judges or time limits, where a fight simply continued until somebody could not. But the trend holds long after that stopped being true — 38.7% in 2001–2010, 40.3%, 42.8%, and now 45.1%. Each era has been more likely to end at the scorecards than the one before it.
And the finishes that do happen come later
If fights are getting harder to end, we would expect the finishes that do occur to take longer to arrive. They do.
In the earliest era, 76.3% of all finishes came in the first round. Today it is 44.2%. Fighters are not merely surviving to the judges more often — even when they lose, they last longer before it happens.
Knockouts arrive on average in round 1.75, submissions in round 1.83. The two are closer together than the folklore suggests: the idea of the early knockout and the late, grinding submission is not really borne out.
What survives of grappling is strangles
One last detail about the submissions that remain. Of the 288 in our archive where a specific technique is recorded, 86.8% are chokes. Joint locks — the armbars, kimuras and heel hooks that filled early highlight reels — account for just 5.9%.
The reason is positional rather than technical. Limb attacks require an opponent to make an error you can exploit at range; strangles follow naturally from control. Modern fighters make fewer errors, so what survives is the submission that comes from simply being in a better place than the other person.
So what changed?
Not violence. Competence. The professionalisation of MMA did not produce a more explosive sport — it produced a sport in which everyone is much harder to beat decisively. The skill gaps that made 1993 so dramatic have closed, and a closed skill gap produces close fights, and close fights are scored rather than finished.
Whether that is good is a matter of taste. That it happened is not.
How we did this, and what we got wrong
These figures come from our own archive of 623 UFC event reports, parsed programmatically into a structured database of 4,287 bouts. Every percentage above is computed from that data and can be reproduced from it.
Two honest limitations. Round data is recorded for 2,380 of 4,287 bouts, so the round-based findings rest on roughly half the sample rather than all of it. And 2026 is a partial year, so the most recent era carries fewer events than a full six-year period would.
A correction worth stating plainly. An earlier version of this analysis reported that knockouts had risen sharply — from around 18% to over 30% — and that decisions were collapsing. Both were wrong, and the cause was our own parser. It read each bout's result from a fixed window of text that frequently overlapped the next bout on the card, and it recognised the word "decision" but not the abbreviations "UD" and "SD" that our more recent reports use. The combination deleted modern decisions and misattributed neighbouring knockouts, manufacturing a dramatic trend that did not exist. We found it before publishing, rebuilt the database, and the corrected figures are what you see above. We mention it because a number is only as good as the method behind it, and ours was not good enough the first time.
This is our own documentation rather than official promotional records, and where the two disagree, the official record should win. If you spot an error, our contact page is the fastest way to tell us.