Attn! MMA math model has Sara McMann winning 52% of time

[ibetkellybro]

just reran my model in wolfram alpha's mathematica and McMann's implied probability is 52. Bet accordingly

[Ron_Paul_2012]

just reran my model in wolfram alpha's mathematica and McMann's implied probability is 52. Bet accordingly

...

[jacktheknife]

This oughta be a good one.

[Ron_Paul_2012]

This oughta be a good one.

Yeah. I had posted a rather funny comment & video. However, this thread is such low hanging fruit that I decided to erase the post & video. I already know that everyone is going to pile on. I just couldn't be a part of it.

[NunyaBidness]

Isn't mathematica and wolfram alpha separate things? I haven't used Mathematica since college but I know it wasn't capable of what Wolfram|Alpha is.

Love to take a look at your model, though.

[NunyaBidness]

Yeah. I had posted a rather funny comment & video. However, this thread is such low hanging fruit that I decided to erase the post & video. I already know that everyone is going to pile on. I just couldn't be a part of it.

Why would we pile on sider? Everyone likes him and he actually posts his bets unlike others around here.

[Vaughany]

My wolfram alphamale model has Cody Mckenzie winning the LW title by 2015 at 83%

[Educ8d Degener8]

Isn't mathematica and wolfram alpha separate things? I haven't used Mathematica since college but I know it wasn't capable of what Wolfram|Alpha is.

Love to take a look at your model, though.

I bet you took a Turbo Pascal or Fortran programming course...

[mirinquads]

Well that sounds like another all in then

[marzwoody]

What does that fancy mathematica tell you about rhondas win % ?

[MD]

Isn't mathematica and wolfram alpha separate things? I haven't used Mathematica since college but I know it wasn't capable of what Wolfram|Alpha is.

Love to take a look at your model, though.

They're separate but related. Alpha is based on Mathematica, which were both created by Wolfram Research. Sider probably either meant to call it "Wolfram's Mathematica", or to say that he ran it through Wolfram Alpha. Likely the latter.

[ibetkellybro]

What does that fancy mathematica tell you about rhondas win % ?

48 percent

[ibetkellybro]

They're separate but related. Alpha is based on Mathematica, which were both created by Wolfram Research. Sider probably either meant to call it "Wolfram's Mathematica", or to say that he ran it through Wolfram Alpha. Likely the latter.

yeah the program not the search engine. "Wolfram.Research.Mathematica"

[ibetkellybro]

Yeah. I had posted a rather funny comment & video. However, this thread is such low hanging fruit that I decided to erase the post & video. I already know that everyone is going to pile on. I just couldn't be a part of it.

[ibetkellybro]

we will see, i might publish the source code to git or sell this model for millions one or the other

[marzwoody]

I want to lick that nose.

[NunyaBidness]

They're separate but related. Alpha is based on Mathematica, which were both created by Wolfram Research. Sider probably either meant to call it "Wolfram's Mathematica", or to say that he ran it through Wolfram Alpha. Likely the latter.

Thanks Gunshard.

[MD]

Thanks Gunshard.

Did you read your own post? No reason to think you knew that, based on what you said.

[NunyaBidness]

Thought it was pretty clear, personally. If the desktop version of mathematica had the functionality of wolfram I'd be very happy, I hate using online interfaces and subscription products.

[Hannibal]

I cannot fathom how a mathematical model could predict an mma match more accurately than educated analysis.
Sounds like a lot of pretentious dick measuring

[ibetkellybro]

I cannot fathom how a mathematical model could predict an mma match more accurately than educated analysis. Sounds like a lot of pretentious dick measuring

you might be right, very small dick here

[ibetkellybro]

Thought it was pretty clear, personally. If the desktop version of mathematica had the functionality of wolfram I'd be very happy, I hate using online interfaces and subscription products.

i dunno about windoze but the linux version is up on nzbs

[NunyaBidness]

I cannot fathom how a mathematical model could predict an mma match more accurately than educated analysis.
Sounds like a lot of pretentious dick measuring

Really? I'm sure people said that about baseball, football, and every other sport in the past as well. Feel cappers are getting killed every year in the sports.

I have some rudimentary MMA models that I use to assist my capping, but eventaully someone much brighter than me will solve it. I'm pretty sure at this point you could beat MMA doing some rough modeling although the variance would be high.

[ibetkellybro]

going to be a star studded event

Clarissa Chun @ClarissaChun Jan 9 @helen_maroulis I'm going too! 😁 Go Sara and Cormier! 💪

going to be a big night wonder how many judo people will be there?

[Thor4140]

http://forum.sbrforum.com/boxing-mma-ufc-betting/2769266-ufc-170-rousey-vs-mcmann-february-22-2014-a.html

[MD]

Really? I'm sure people said that about baseball, football, and every other sport in the past as well. Feel cappers are getting killed every year in the sports.

I have some rudimentary MMA models that I use to assist my capping, but eventaully someone much brighter than me will solve it. I'm pretty sure at this point you could beat MMA doing some rough modeling although the variance would be high.

I feel like MMA would be one of the hardest sports to beat using just statistics, though, because of the lack of constants. Ignoring grappling, a guy like Cerrone has ridiculously bad results against strikers fighting from south paw, while Martins is very accurate with his straight left. It's a stretch, however, to compare Martins landing at will on cans to Cerrone getting chewed up by Pettis and Diaz. Both stylistically and in terms of level of competition. Maybe I'm wrong, but I assume that would be a problem inherent to both MMA and boxing. The nature of the sports dictates low sample sizes for fights, combined with very little comparable data. I've heard of plenty of NFL models, plenty of NBA models, even a few MLB models, but very few people seem to be beating boxing or MMA with modelling. How many times a year do NFL teams rematch each other? Most MMA fighters won't even have mutual opponents. I could very well be off-base here, it's just something I thought about when I started 'capping fights.

I suppose it helps that everyone doing MMA analytics at the moment is bad at it. The closest thing to an MMA statistician is Reed Kuhn, and he's terrible and misleading.

[Hannibal]

I dont mean to insult..I was trying to express my lack of understanding.

Correct me if im off..but doesnt a mathematical model depend on the data it is fed?
How to you quantify qualitative data?? And if you only input the available quantitative data, how accurate would it be based on the small pool of data? Many fighters have less than a dozen fights. I could see a model being morr effective in the sports you describe because there are soo many games to sample data from.

[NunyaBidness]

I dont mean to insult..I was trying to express my lack of understanding.

Correct me if im off..but doesnt a mathematical model depend on the data it is fed?
How to you quantify qualitative data?? And if you only input the available quantitative data, how accurate would it be based on the small pool of data? Many fighters have less than a dozen fights. I could see a model being morr effective in the sports you describe because there are soo many games to sample data from.

I wasn't insulted.

That's why I'm saying the variance would be large. You could draw general inferences from the data, which might put you in some bad spots here and there but would in general beat the market.

[NunyaBidness]

How many times a year do NFL teams rematch each other?

Snicker.

[NunyaBidness]

I feel like MMA would be one of the hardest sports to beat using just statistics, though, because of the lack of constants. Ignoring grappling, a guy like Cerrone has ridiculously bad results against strikers fighting from south paw, while Martins is very accurate with his straight left. It's a stretch, however, to compare Martins landing at will on cans to Cerrone getting chewed up by Pettis and Diaz. Both stylistically and in terms of level of competition. Maybe I'm wrong, but I assume that would be a problem inherent to both MMA and boxing. The nature of the sports dictates low sample sizes for fights, combined with very little comparable data. I've heard of plenty of NFL models, plenty of NBA models, even a few MLB models, but very few people seem to be beating boxing or MMA with modelling. How many times a year do NFL teams rematch each other? Most MMA fighters won't even have mutual opponents. I could very well be off-base here, it's just something I thought about when I started 'capping fights.

I think all of this assumes that every fighter is a unique flower and that certain generalizations cannot be drawn, where in I would say that most of it boils down to variance.

I initially started posting on this board because I was doing some modelling work and thought I might find some likeminded people on here. Lol.

[Vaughany]

Ever used the "Wiki-winnerz model"?

[Educ8d Degener8]

Nuns, do you datamine from BFO? I did for a bit to run some troglodytical analyses but it was tedious as hell... It was tolerable when I was doing it on company time though LOL.

[ibetkellybro]

Nuns, do you datamine from BFO? I did for a bit to run some troglodytical analyses but it was tedious as hell... It was tolerable when I was doing it on company time though LOL.

you scraping it becuz i dont think there is a api for it? BFO misses a lot of movement. Plus dimes is just retarded with their feed. best bet is get a api key to pinny and hook in to their api and have a virtual server logging it in to a sql database for you. could do it easy with python or php

[ibetkellybro]

btw i fuking hate php

[NunyaBidness]

Nuns, do you datamine from BFO? I did for a bit to run some troglodytical analyses but it was tedious as hell... It was tolerable when I was doing it on company time though LOL.

It's way easier to scrape from the books themselves. Hard part is getting the names to line up.