protop Posted April 9, 2008 Share Posted April 9, 2008 Anyone got any views on using the poisson model for predicting 1X2 for football matches. Do you think it's a profitable way to go Do you use it yourself How about a modified version All thoughts both positive and negative most welcome Quote Link to comment Share on other sites More sharing options...
Grex Posted April 9, 2008 Share Posted April 9, 2008 Re: Poisson Model Pardon my limited knowledge of Poisson´s Distribution, but I understood it expresses the probability of a number of events occuring in a fixed period of time if these events occur with a known average rate and independent of time since the last event (with acknowledgement to Wikipedia:clap). I don´t see how this theory can be "adjusted" to take into account relative strengths of teams and other dependable variables that affect the result of a football match. But I'm sure the "statisticians" among us will correct me! Perhaps it´s a case of one man´s poison is another man´s.... (joke) :ok Quote Link to comment Share on other sites More sharing options...
muppet77 Posted April 10, 2008 Share Posted April 10, 2008 Re: Poisson Model have a search on the PL mate for poisson. i looked at it ages ago as did a few others. Quote Link to comment Share on other sites More sharing options...
muppet77 Posted April 10, 2008 Share Posted April 10, 2008 Re: Poisson Model in fact here it is. http://www.punterslounge.com/forum/showthread.php?t=8171&highlight=poisson&page=1 Quote Link to comment Share on other sites More sharing options...
protop Posted April 10, 2008 Author Share Posted April 10, 2008 Poisson Thanks muppet77 have had a look throuh the threads, interesting read. Did you abandon the idea or improve upon it. The reason for asking is I have spent the last couple of years or so playing with differnt ideas and have come up with a model which will hopefully bear fruit. A few stats I have collected since the 4th March to date; 180 Fixtures rated: - a small sample I know! Using the poisson model to predict Homes and Aways (Average 63%) Using same model to predict Just Homes (Average 64%) Using same model to predict - Just Aways (Average 56%) Actual Results : Homes and Aways - 62% Homes - (63%) Aways - (58%) Another interesting point if i had placed 1 unit on each game without taking in to account Exchange commission I would have lost 1.8 units which is exactly ROI -1% So it seems to me that for the moment at least the model is predicting the games with some accuracy and secondly overall so are the Bookies/Exchanges. However that is overall, amongst the selections there are a reasonable amount of value bets according to the model and these are the ones of interest. A couple of points: The model is only inteseted in the outcome of games from leagues that fit the model this is dynamic - No Italian A or B games at the moment, for example. All selections must have a predictive percentage of at least 50% There also many filters in place which ignore certain fixtures: Like no games are rated that have a cup game coming up within 7 days that kind of thing. Anyway enough rambling, just thought I would add to the debate. Good luck with your own betting. Protop Quote Link to comment Share on other sites More sharing options...
muppet77 Posted April 10, 2008 Share Posted April 10, 2008 Re: Poisson Model i sacked it off after it produced unexciting results. i remember discussing the fact that it underestimates the 1-1 score and also treats goals as separate discrete events/probabilities. in truth a team that goes 3-0 up works less hard trying to score than when it was 0-0. this is one issue. Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.