High Percentage martial arts takes data from hundreds of fights, competition matches, and self defense situations to answer the question: What's truly high percentage?


We Recorded 500 Submissions to Get Real Data on What's High Percentage

We Recorded 500 Submissions to Get Real Data on What's High Percentage

This is the largest study of BJJ submissions that we’ve ever done. It took over two months to complete.

This is the largest study of BJJ submissions that we’ve ever done. It took over two months to complete.

The numbers are in! This is the most exhaustive submission study we have done so far on this site. High Percentage’s mission is to use science and data to focus our game on what the cold hard numbers say is high percentage. Luckily, the data from these matches is practically screaming at us: A small, powerful group of submissions outperform all others. As well as some other surprises.

What We Did

We watched five hundred matches that contained submission finishes. One hundred of these were white, another hundred blue, another hundred purple, and so on. Half of all matches at each belt were in a gi, the other half nogi.

We marked every submission finish and timestamped them based on when they occurred in the match. We came up with common definitions of every submission, often preferring to clump variations of the same submission together (ex. inside heelhooks and outside heelhooks are just “heelhooks”).

Three of us compiled the data and reviewed portions of each other’s work. In the rare cases of disagreements, we opted for a rule that if there wasn’t unanimous consensus, we would omit that particular match. This mainly happened in cases of “double submissions” such as a triangle/armbar where it was unclear what caused a tap.

Hungry for more? We’ve partnered with Gold BJJ to deliever an entire BJJ Game Theory Course. Go to goldbjj.com/highpercentage to learn more.

Hungry for more? We’ve partnered with Gold BJJ to deliever an entire BJJ Game Theory Course. Go to goldbjj.com/highpercentage to learn more.


  • There are always unavoidable limitations in collecting this data. We do the best we can to be accurate and efficient. We are a team of about three, and we check each other’s data and come up with shared definitions to avoid subjective judgments. For a full breakdown of how we record the matches, here’s a short podcast episode where we went into details.

  • Technically, the term “high percentage” is a little misleading because we did not record attempts (although we have done that before). Measuring submission attempts is challenging and often highly subjective, so we are cautious about doing so.

  • We did not factor in rulesets in this study. We will have other content that dives into specific rulesets, but we didn’t do it for this one. For the most part, the gi matches were probably overwhelmingly IBJJF rules, with some outliers. For nogi, there is a larger variety of rulesets, often not always apparent. Taking the time to double check the various rulesets of a promotion is a grind on top of a grind. Ultimately, it primarily affects the numbers of certain submissions such as heelhooks, which were almost certainly illegal moves in many matches we observed, although it’s difficult to put a number on it. We’ll address that later.

So What’s The Top Finisher?

Rickson by armbar. Er, make that: everybody

Rickson by armbar. Er, make that: everybody

For regular readers of our site and listeners of our podcast, it will come as no surprise that the armbar is, without question, responsible for more finishes, at more belt levels, in both gi and nogi, than any other single submission. Period.

It is the quintessential finisher. Responsible for 123 (25%) of ALL submissions across the board. It leads the pack in gi matches and is in a virtual tie for first with the RNC in nogi matches.

Just Three Submissions Account for Half of ALL Finishes

Although we’re fans of innovating new techniques, your favorite cliche about “basics” is still safe to use. In fact, the Armbar, the Triangle, and the Rear Naked Choke combined accounted for 51% of all finishes. They were so dominant, they outperformed the other twenty recorded submissions combined.

There Is A Very Different Set of High Percentage Nogi Submissions

gi vs nogi submissions.png

If you’re going to be competing in nogi, you definitely want to sharpen up on some alternative attacks. The most notable is the RNC, which accounted for less than 5% of gi finishes but a staggering 19% of nogi. In fact, the RNC beat out the armbar as the most common nogi finish by just a single point.

Other powerful nogi subs were the guillotine. In a gi, it finished just five matches. In nogi, it finished twenty five. Similarly, foot/ankle locks (which we combined for this study because it’s nearly impossible to distinguish just by watching) were dominant in nogi. Rounding out the top nogi submissions were heel hooks and arm triangles.

The Chances Of A Submission After Four Minutes is Virtually Zero.

We don’t advise the whole “wait for your opponent to get tired” strategy in competitive jiu-jitsu (but in self defense, it might be a good idea). A mere six matches ended with submissions after the four minute mark. If these matches are representative of the sport as a whole, or anywhere even close, your odds of a late submission are just 1.2%

So you’re saying there’s a chance. . .

There Are No Special Black Belt Submissions

The second image above this text is black belt submissions only. You can see, it’s close to a mirror image of the total population. The majority of black belt submissions (53%) are still armbars, triangles, and RNCs.

If anything, black belts were even more laser focused on the top three submissions. A larger percentage of black belt submissions were armbars than any other group.

Also notable to us was that black belts used significantly less kimuras, bow and arrows, and heelhooks than colored belts.

So What About Leg Locks?

This gets tricky. Foot and anklelock submissions are allowed, generally, across the board at all levels. So their data should be considered sound. Kneebars, toeholds, and heelhooks have wildly different conditions, heelhooks being banned even in nogi, black belt level IBJJF competition. Because of this, it’s perfectly reasonable to assume that heelhooks, toeholds, and kneebars may be underrepresented because in so many of the matches that we watched, they were illegal.

But consider this: Heelhooks currently are the 12th most frequent submission. To break into the top five submissions, they’re true frequency would need to be triple what it is now.

We definitely plan on doing leglock specific research later this year, so hold on!

What’s Next?

The easiest place to take this data next is recording what positions each of these submissions occurred from, with special attention to guards. We have strong data on armbars that they are much more successful from the top than bottom, but that data is only for white belts in a gi.

We are also going to prioritize some new studies on submission only matches, we might enlist your help in the most efficient way to do that.

Finally, we have another article coming next week taking a closer look at when submissions are most likely to happen. But we already have some great data on this topic.

Think about jiu-jitsu in a car? You might like The High Percentage Podcast!

Podcast: How Do We Record The Matches?

Podcast: How Do We Record The Matches?

Podcast: The Law of Large Numbers

Podcast: The Law of Large Numbers