Your weight class kind of doesn't matter until black belt.

A data-driven look at how weight differences in jiu-jitsu matches affect win rates

Thanks to IBJJFRankings.com, I got to analyze more than 3,000 IBJJF open class matches to see if weight matters. Spoiler alert, weight matters (statistically), but not as much as you might think after a closer look. A quick summary in case you’re not going to look at all of my cool charts:

• Among most colored belts, going up one weight class (~12 lbs) makes virtually no difference in the probability of winning.

• Among black belt adult competitors, every weight class difference matters a lot, especially for women.

When we look at matches across all ages and all ranks, it’s safe to say that one weight class isn’t that big of a deal — the lighter athlete wins 45 percent of the time, almost a coin toss. That means, in general, maybe you don’t need to cut that extra five pounds for your next tournament.

It seems like things start to fall apart at a four weight class (~50 pound) difference. It didn’t matter if an opponent was four, five or six weight classes heavier, the lighter opponent won 30 percent of the time. A seven weight class difference is kind of a death sentence. We found 18 matches where this was the case, and only one was won by the lighter opponent. We also found three recorded matches where there was a 8 weight class difference (rooster v. ultra heavy). They went the way you’d expect.

Blue belts don’t need to worry about weight.

We looked at open class matches among adult colored belts, and blue belts stand out the most. When there was a one weight class difference, it literally didn’t matter who was lighter or heavier. Out of 163 matches, the lighter person won 83 (as close to half as possible). Apparently, it also didn’t matter when there was a six weight class difference. We can probably chalk that up to fewer matches happening at that weight discrepancy, only 14 matches in our data.

We don’t see a totally linear relationship between weight differences among the colored belts, even brown belts, whose lighter athletes do better when pitted against an opponent two weight classes higher than against an opponent one weight class higher.

The 50 percent success rate for the lighter athlete at a five weight class difference is impressive , albeit data made of 18 matches. Seven out of nine wins were featherweight women defeating super heavy women.

At purple belt, one weight class didn’t really make a difference, but two weight classes negatively affected the lighter athlete. Maybe weirdly, the lighter athlete saw more success with a three and four weight class difference, almost 43 percent wins over more than 100 matches.

One possible explanation for the variance in colored belt success despite weight differentials would be that the technical ability in higher weight classes might not catch up to the the lighter weight classes until black belt. We’ll see shortly below that heavier black belts don’t mess around.

Weight matters more at black belt.

At the black belt adult level, weight mattered a lot more, especially among female athletes. Even with just a single weight class differential, the lighter female athlete won only 30 percent of the time, compared to 44 percent of lighter male athletes.

We can see that the relationship between weight and winning is much more linear and “makes sense.” The heavier, presumably stronger, athlete wins most of the time.

Master division women (colored belts) are animals.

When we look at how open class has gone in masters divisions, master colored belt women defied the odds. In more than 50 open class matches where the weight difference was two or three classes, the lighter athlete won more than half the time. The breaking point was four weight classes, 15 percent success rate for the lighter athlete.

Unfortunately there weren’t enough matches among master black belt women to make a comparison.

Are all master “heavy”-ish black belts the same?

Among master black belt men, those with the courage to venture into open class found almost equal success among one, two, and three weight class differences. A majority of these matches were the “heavies” battling it out: medium heavies beating heavy, super and ultra heavy athletes; heavies beating super and ultra heavies; super heavies between ultra heavies. With winning rates of greater than 45 percent, one might wonder how much weight cutting is necessary if you’re that big, that old, and that good. Only more data will tell!

About the data

Match data was collected by IBJJFRankings.com (@ibjjfrankings on Instagram), an initiative to establish ELO rankings for IBJJF athletes. According to its founders, IBJJF will run 140,000 matches per year. I’m looking forward to revisiting some of analysis when the dataset is literally 40-50x bigger. What this means is that you’ll see the margins of error in the first chart of this article get smaller, and the results will be even more reliable especially when we divide the data up into demographic cross-sections (various groupings of athletes).