Mathematical model for predicting baseball outcome


Well-Known Member
Jul 19, 2009
Croatia, Zagreb
e46 330xd Touring
My friend, who loves baseball and programing, decided to make a program, mathematical model, that would predict the outcomes of games. I'm just here to spread the word, maybe some here would be interested in this and have some use for it or you can help him out by giving him some feedback on all of this. I'll just quote him :

Hi there,

I'm an engineer and mathematician who decided to combine passion for going over MLB statistics and betting on it with control systems engineering and math. That resulted in a mathematical model written in >1000lines of Python (data acquisition included) which spits out a bunch of variables being crunched and uses some math (discrete distribution, Poisson mainly) and a bunch of heuristics to determine whether a game is worth betting on and who to bet on. I call these "best picks". Ofcourse, lesser certainty picks are chosen also but I use those only to tweak the model when I see fit (filtering out some statistical anomalies etc.). Model is about 80% successful all together.

I worked on this for about 2-3 weeks straight, using past this season data to get the model working, spent another week evaluating it and put everything online a few days ago. You can find it all at It is currently at 7/8 for best picks, and 8/14 for lesser picks (but as I said, don't bother with those).

I don't have a lot of time currently to put more detailed info online, I hardly get time to put game picks online, but in short it uses statistical analysis of team streaks, all available pitcher data and some sabermetrics and a lot of heuristics I came up with while making the model and before, when I bet on baseball "by hand". I still have to put ball park factor into everything and team batting, but that won't happen at least for a week or two so don't hold your breath :)

I put quite a lot of effort into this and I'm quite pleased with it. I don't plan on charging people for using this, everything is freely available online (I just won't publish the model itself)... I'm only thinking about placing a paypal "donate" button so if someone makes money out of this and wants me to continue putting it online and expanding the thing (thus spending an hour a day on it) he can :)

What do you think? Should I continue this little project of mine?

ps: on the tech site, it uses beautifulsoup4 for fetching stuff from web and numpy/scipy for math stuff.
Mathematical model for predicting baseball outcome

Hi Boobo-oobo,

I was wondering if your friend ever continued to work on this project and it it is available online anywhere? If so please let me know because I would love to use it and maybe even add some of my own stuff to it as I am interesting in doing something similar myself. Anyways, my email is I would love to hear back from you soon so I can continue to get some more feedback on my own work and look at your friends work for ideas. Please let me know as soon as you have a chance. Thanks so much.