UFC 118 Penn-Edgar 2 Toney-Couture

[Luke]

Dana White says James Toney is one and done with the UFC (video)

http://bit.ly/cIFL6u


I am shocked . I thought he had a mutli fight deal

[Svino]

Dana White says James Toney is one and done with the UFC (video)

http://bit.ly/cIFL6u


I am shocked . I thought he had a mutli fight deal

Too bad Dana let Kimbo go. Kimbo vs. Toney would have sold mad PPVs.

[zY|]

Dana White says James Toney is one and done with the UFC (video)

http://bit.ly/cIFL6u


I am shocked . I thought he had a mutli fight deal

I'm shocked as well...

[Mr. IWS]

Dana White says James Toney is one and done with the UFC (video)

http://bit.ly/cIFL6u


I am shocked . I thought he had a mutli fight deal

Knowing Dana, Toney will be fighting Brock for the title for the super bowl show.

[Mr. IWS]

-1.8 units for the night
-12.05 units for the year

[Mr. IWS]

^^^^^^^^If you take away the two BJ/Edgar fights, Im up around 1 unit for the year.

[poopoo333]

-12u for the year for just MMA? Or all things?

[SPX]

^^^^^^^^If you take away the two BJ/Edgar fights, Im up around 1 unit for the year.

Fuck you, BJ Penn!

[SPX]

We need to develop a computer system where we can just input various information, like what their grappling credentials are, who they've beat, their striking level, TDD abilities, etc. and then have it compute the likelihood of that fighter beating another fighter.

I wonder if that's possible. I bet with enough thought it would be.

[Luke]

We need to develop a computer system where we can just input various information, like what their grappling credentials are, who they've beat, their striking level, TDD abilities, etc. and then have it compute the likelihood of that fighter beating another fighter.

I wonder if that's possible. I bet with enough thought it would be.

You cant even calculate odds silly ,how are you going to develop a computer program?

[Luke]

Knowing Dana, Toney will be fighting Brock for the title for the super bowl show.

Nah since the Superbowl is football he'll probably sign Hershell Walker for that task

[SPX]

You cant even calculate odds silly ,how are you going to develop a computer program?

I said "we," motherfucker!

[zY|]

It's done for other sports, so I'm sure it's possible. The question is how accurate could you make it?

[Luke]

It's done for other sports, so I'm sure it's possible. The question is how accurate could you make it?

fairly accurate imo.

You could use a simple formula and use excel for it . It wouldnt have to be some complex computer program.
You would just need to assign values to everything and then when the difference is over a certain amount make it a bet

[Svino]

We need to develop a computer system where we can just input various information, like what their grappling credentials are, who they've beat, their striking level, TDD abilities, etc. and then have it compute the likelihood of that fighter beating another fighter.

I wonder if that's possible. I bet with enough thought it would be.

If you could do that with any sport, I think it would be baseball. As for MMA, I always wanted to see someone try to implement a serious ranking system. Something like the Elo system they use for chess would probably work OK, although the the fact that guys fight so few times per year would make it iffy.

But even this wouldn't be able to say anything about style matchups. I don't think you'll ever be able to put anything together that's actually better at picking fighters than a person who knows what they're doing.

There is a good article by a guy who worked on Microsoft's "TrueSkill" algorithm. It's a little mathy, but it starts from the basics.

http://www.moserware.com/2010/03/comput ... skill.html

[SPX]

It's done for other sports, so I'm sure it's possible. The question is how accurate could you make it?

That's a good question. I really don't know. It would basically have to work off of a shitload of statistics where it basically looks at previous fights where two fighters had very similar credentials/characteristics/abilities and then determines how often one the fighter with profile A won/lost to fighter with profile B.

For things like takedown defense and what not, I figure you could assign it a number that represents a particular level. Like levels 1 through 5, with 1 being James Toney and 5 being Machida. Obviously this is something that an educated human would have to estimate and enter into the program, so there's a certain human error factor that plays into it, especially, for instance, if you're trying to estimate TDD for a guy with only a few fights or a guy who has fought mostly strikers and guys with poor wrestling ability.

I have no idea how to do it, but it's an intriguing idea.

[zY|]

It's done for other sports, so I'm sure it's possible. The question is how accurate could you make it?

fairly accurate imo.

You could use a simple formula and use excel for it . It wouldnt have to be some complex computer program.
You would just need to assign values to everything and then when the difference is over a certain amount make it a bet

I think it would be extremely tough to account for style matchups as Svino says. Perhaps it could be done, but I dunno. It would take a ton of trial and error.

And Svino, http://www.mma-elo.com/

[Svino]

And Svino, http://www.mma-elo.com/

Well shit. Thanks.

[zY|]

It's done for other sports, so I'm sure it's possible. The question is how accurate could you make it?

That's a good question. I really don't know. It would basically have to work off of a shitload of statistics where it basically looks at previous fights where two fighters had very similar credentials/characteristics/abilities and then determines how often one the fighter with profile A won/lost to fighter with profile B.

For things like takedown defense and what not, I figure you could assign it a number that represents a particular level. Like levels 1 through 5, with 1 being James Toney and 5 being Machida. Obviously this is something that an educated human would have to estimate and enter into the program, so there's a certain human error factor that plays into it, especially, for instance, if you're trying to estimate TDD for a guy with only a few fights or a guy who has fought mostly strikers and guys with poor wrestling ability.

I have no idea how to do it, but it's an intriguing idea.

I'm not sure if just assigning numbers like that in a blanket fashion would account for all the subtle nuances of the MMA game. I think there would need to be a ton of carefully selected variables to get an accurate assessment of all the intricacies of MMA, which we still don't really understand fully.

Another thing, for instance, grappling credentials isn't a reliable determinant of success in MMA. Some guys translate it well, others don't at all. It seems like it would have to only rely on in-fight performance, and this would take awhile to do.

[Mr. IWS]

-12u for the year for just MMA? Or all things?

Just mma