[screnn]

Hey guys, i've searched the internet and it looks like this is the forum where people actually know their shit. So I figured this is the place to come with the questions I have.

I'm a junior at university, pursuing an applied mathematics degree. What I want to know is what mathematical methods you guys use to make predictions. I know its based in probability in statistics obviously, but specifically what techniques are being used?

I have a solid background in programming so I'm really looking to jump in head first here and start doing some predictions.

Also just to be clear, I'm not pursuing a math degree just to do sports betting, i'm more so looking to do it on the side.

[Waterstpub87]

Good luck with your math degree, absolutely one of the best majors to graduate with. As far as my personal handicapping has gone, I haven't used more the averaging, dividing, subtracting ect. Some of the things I've used recently have been
Pythagorean odds (GF^2/GF^2+GA^2) to get winning percentages, in addition to some of the saber metric figures for baseball. I'd like to experiment with using monte carlo for prop bets, but I haven't got into that yet.

An understanding of basic statistics, goodness of fit models, correlations ect. can help you to reinforce your opinions about the correctness of your models. Learning some of the methods that industry professionals use to price financial products might help you to understand the methodology of pricing uncertainty even if they don't directly factor into your betting.

Many of the more advanced math I'm sure would be useful, but I haven't gotten around to using it. Most of the financial math I've had to learn (Black-Scholes, ect) measures contains a component that measures a change to time, which in my experience, isn't as much of a factor in sports games.

This is a good forum to post on, but many people will give you answers that may seem alittle less polite then you would like. Don't let it bother you, I've always appreciate the brutal honesty that posters have here. I would much rather someone tell me that what I am trying to do is stupid, rather then bet it and loss the money.

[frongi]

i'm in a similar boat as you are, i'm currently a college math major. one thing you should know is that people are unlikely to give away their secrets on this forum, because if they truly have a great model that enables them to win money, they sure as hell aren't going to start giving it out or even how to get there. feel free to pm me if you want to talk about modeling sports because i'm just getting into it and i'm sure my model isn't great but it's a start, and i could probably help you out a little.

[screnn]

fongi - my post count is too low to send a pm, you should shoot me your email in a pm

[VLR100]

Depends on how complex you want to go. At the basic level there is simple regression analysis (OLS, logit, poisson, whatever depending on the data) and simple probability. If you want to go more advanced, use Markov Chain Monte Carlo methods to simulate entire events. It's all about the data though; the more variables and observations you have, the more markets you can analyse.