Why doesn't bjj have like a elo ranking system?

It would perfectly, we'd know who was the best and the elo ranking system would be awesome. Forgive me if this has already been done.

 

1. nice work on reading wikipedia.

2. the ibjjf recently introduced a ranking system.

http://www.ibjjf.org/ranking.htm

 

I see you nick, I do think the elo raking is pretty good, also the tenis ranking system could be applied, as a matter of fact, any ranking system will be better than the actual non existence ranking system lol...

 

I wasnt aware of this....

 

Wikipedia? ELO is common knowledge. They use it in chess, hence my username.

 

I felt that it wasn't coinsidence that this was brought up at a similar time to a Magnus Carlsens ranking being on the wikipedia front page.

 

Why does BJJ even need a ranking system?

So people can sit at their keyboard and talk about rankings and who is better, just like MMA?

 

If anyone wants BJJ to be a pro sport, yes, rankings are needed so tournaments can be seeded, or contenders lined up against a champion.

 

The big problem with Elo ranking in combat sports is that combat sport results are not transitive. There is too much style vs style matchup influence.

If A beats B, and B beats C, then you'd think A would beat C even more easily. However, that doesn't hold up. This messes up an Elo ranking system pretty badly.

 

Someone actually did an elo system for MMA a few years back. It didn't work very well. It came up with fairly absurd results IIRC.

 

dude you just posted with your real account, the dude was answering to your whitebelt troll account I would guess, although, I didnt think this topic was a troll thread...

 

ELO is very bad for combat sports for a lot of reasons. Lack of transitivity, not enough activity etc...

 

didnt think on this, elo for MMA would be stupid.

 

I think I remember this. These were the rankings that had Jeff Monson as HW GOAT?

 

No ranking system is perfect. The seeding in tennis gets a bit whacky when a relative unknown gets a top 10 ranking for playing and winning a ton of small tournaments, but against less than stellar competition.

It's still worth doing though, as every sport does do it.

 

Yeah, something like that.

There is a new version out now, that is not as bad... but I am pretty sure it is not a pure elo ranking system, and it is still pretty bad IMO: http://www.mma-elo.com/ranks.php

 

That can be true in chess as well, or tennis. They use the system with little trouble. It tends to average out, anyway...if you're a top level guy, you'll have trouble with some guys and other guys will have trouble with you. Doesn't invalidate the system.

 

thats why tennis doesnt use an elo system... it does totally invalidate the system for combat sports, because as we all know, mmath does not work, and an elo system is basically based on mmath (if it was used for MMA)

 

No, but I am pretty sure the elo ranking system is based entirely on what has already been done. For sport rankings to be of any real use, they have to also consider potential and the obviousness of an up and comer's skill, even if it hasn't been proved yet over the course of many matches. Otherwise the rankings don't have much predictive value or don't give us "contenders" in the way we want them.

 

It does invalidate the system when there is a true loop that breaks transitivity. The problem is only averaged out if it's a fluke and not representative of the underlying probabilities.

If A beats B most of the time, and B beats C most of the time, and C beats A most of the time, they will all tend to end up with the same Elo ranking if they all fight repeatedly. But this invalidates the system because the ranking is supposed to statistically predict the outcome. It will say that A has an equal chance against B, B has an equal chance against C, and C has an equal chance against A when we just stated that the true probabilities were much different than this.

There isn't really enough transitivity to have it make sense within say a top 20 black belt bracket. You end up with really weird proxy world champion results that screw the rankings to nonsense. Just try applying Elo rankings to the results of the major IBJJF tournaments last year, and I think you will get some very strange results accordingly.