When looking at data we need a sizeable data set with a large sample size. The problem is that a BJJ black belt match can last 10 minutes with no submission, of course it can potentially be shorter with a submission. Even if a match were to last 6 minutes with a submission end 100% of the time. This would still be 10 hours of footage, perhaps half of which would end with a joint lock, leaving us with no data on high percentage gi chokes.
The following data set comes from the IBJJF 2019 Worlds. This data is for all belts and both genders. Out of 630 fights, 150 were finished with a choke.
To start off with a disclaimer, this is not a fantastic data set, a lot of the fights will be at lower belts (that said, this is the world championship so quality will be high). In addition to this, 630 fights is not too many, for a proper analysis I would actually like around 5 – 10x this data, but it is not really possible given the time it takes to count these submissions.
Table of Contents
Finishes
| Submission | Successful Chokes | Unsuccessful Chokes | Finishing Percentage |
| Collar Choke From The Back | 66 | 23 | 74% |
| Bow & Arrow | 17 | 2 | 89% |
| Rear Naked Choke | 10 | 2 | 83% |
| Triangle | 28 | 17 | 62% |
| Cross Collar | 9 | 1 | 90% |
| Ezekiel | 4 | 2 | 50% |
| North South | 4 | 0 | 100% |
| Lapel (E.g Peruvian Necktie) | 3 | 0 | 100% |
| Baseball | 2 | 2 | 50% |
| Crucifix | 2 | 0 | 100% |
| Head & Arm | 2 | 2 | 50% |
| Darce | 1 | 0 | 100% |
| Anaconda | 1 | 0 | 100% |
| Guillotine | 1 | 0 | 100% |
| Loop | 1 | 4 | 20% |
| Pressure from mount | 1 | n/a | 100% |
What is a high percentage gi choke?
The data alone can be a bit misleading as percentage (even with a high sample size) does not show us everything. To be honest “high percentage” moves is a bit of a silly phrase in BJJ right now. Not many people know what they are talking about as they don’t have the data, in addition to this, just because something is high percentage doesn’t necessarily mean it’s good for a few reasons.
- Low percentage may actually be better in some circumstances. If we look at pressure from mount, we already know pressure from mount is good since it holds a dominant position with no stalling penalty. We have a minimum of 4 points from being in mount, if the opponent taps we get a neck pressure mount choke, which may be a 1 in 500 submission, but we have given up nothing by attempting it. A similar argument can be made for mounted Ezekiel which is only 50% successful. If performed correctly you surrender nothing by attempting it.
- By the same logic, high percentage can possibly be bad. If you are winning by 1 point in a dominant position with 45 seconds left, it may be unwise to go for a 95% successful gi choke, since 5% of the time you will miss is and potentially lose the position, causing you to lose the match. If you were to hold the dominant position this gives you a 100% chance to win.
- Certain moves may not be done to finish so will have low finish percentages. We all know IBJJF matches are often won by advantages. We will often see competitors at the high levels attempt toeholds on each other with no intention of trying to finish these, only playing for an advantage. Thus toehold will be a low percentage attack, which is exactly what it’s meant to be.
A further look in to the numbers
Overall I am quite disappointed with the data, other than one glaring factor I will mention later, I do not think it shows too much.
All 3 main chokes from the back are the most applied and have success rates varying from 74% – 89%. These chokes are all often used in conjunction with one another, meaning if one fails you will often revert to the other. Bow & Arrow has the highest percentage success, but with this being said I don’t think it should be looked at like this.
As mentioned earlier, the bow & arrow is the riskiest of these 3 chokes, so whilst having a higher percentage completion, I certainly wouldn’t regard this as a high percentage move in comparison to the collar choke. This is because if someone escapes your bow & arrow, you may at best find yourself in a scramble. If someone escapes a collar choke, they may end up in a completed bow & arrow.
The Glaring Findings
Perhaps the most worrying finding from the table is the 62% completion rate of a triangle choke. Failure of a triangle choke is possibly the worst kind of failure on the list. A failed triangle will cost you 3 points a lot of the time, since if your opponent escapes your guard is very compromised and a good guard passer will take advantage of this.
Based on these findings I would recommend only using the triangle if this is your expert area. Once again the data does not show the entire story as there are seasoned competitors with excellent triangles such as Chad “the beast” Hardy and Ryan Hall. It would be very stupid of me to tell people such as these that the triangle is a low percentage gi choke. However I think that in this case for the average competitor, the stats do not lie, it is a very big risk to take if compared with any other gi choke.
Conclusion
In summary, a lot more work needs to be done. The sample size is very small, not only this but there are limited details about the success and failures of a choke. For example, a failed rear naked choke is recorded in the same way as a failed triangle choke. One of these results in you maintaining back control, the other can result in your guard being passed.
This is a massive disparity and yet both are recorded in the same way, which is really quite ridiculous. With that being said, until we have some kind of artificial intelligence to watch and record BJJ matches, we will probably not truly know what a high percentage gi choke truly is.
We can say with some certainty that getting to the back of the opponent is a truly high percentage way to choke the opponent in gi (and im sure for no gi too). However this was already obvious to begin with. Perhaps in the future we will be able to know what are the most efficient ways to finish an opponent from all of the various guards and top positions once we have a very large data set.
