Hi there
I need some feedback to my website that I have created. The website is my try in creating a free of charge, easy to use betting system for the major European sites.
I need your help, though, for feedback towards what people would you would like to see in the site - and a general opinion in the usability of my website.
Some background knowledge
I have collected data from 10 seasons worth of matches and selected various parameters that I have found affects the outcome of a football match. The parameters are stuff like goals scored (duh!), corner-kicks, shots on goal etc. From these I have created a model over home team victory - one that can predict the probability of a home team victory based on the previous x matches played (by both teams). From here on I calculate what I call "FairOdds", which is the minimum odds (according to my system) you as a betting player should have from a bookmaker if you want to be profitable in the long run.
What can I do with this?
When the "FairOdds" is obtained for a given match, I extract Odds from bookmakers (at the moment, I only have coded one bookmaker - but I will include more as I progress with this project ), and find the highest odds. If this odds is higher than the FairOdds we obtained, the betting amount is calculated (based mostly on the relative difference between FairOdds and the odds from the bookmaker).
The point is, that as you use this system throughout the season (and only bet when and what the system tells you), you will statistically have a profit at the end of this system... That is, if the model works .
Does it work?
Yes it does, although I have some ideas up my sleeve to improve the system, it actually does create a profit. I fitted my model with data up to August 2010, and ran all matches throughout the 2010-2011 season in the following leagues:
English Premier League
German Bundesliga
Italien Serie A
From a total of 1066 matches in the three leagues, the system advised to make 471 bets for hometeam victory, and the overall earnings divided by leagues taken through the entire season were:
Premier league: 28,2%
Bundesliga: 14,7%
Serie A: 10,5%
That is the percentage of pure profit if you use the system the entire season. I expect the Italian league to perform even better in the 2011-2012 season, since the level of training data in this league has been incriesed by 18%.
Until now, only one round in Bundesliga have been plaid, and 4 out of 5 bets have been predicted correctly. That meant a return of a staggering 129.5%. That's just one round, though - and I do not expect the bundesliga prediction model to perform that well through the entire season
.
I need feedback!
As I wrote in the top, I do need your feedback as of how the usability of the site is. Is it obvious when you look at the site what to do? Is it obvious that it is only home team victory I have modelled??
What would you like to see? More statistics, filtering options?
Both the Spanish and French leagues will be added within a month or so.. The Italian is already Added, but as I only show the next week of matches, Serie A won't show up in the table (starts late August).
What filtering options would you find favorable? For example bookmaker filter (so only ones you have an account with is shown?) As I said, right now - only two bookmakers is there atm., but I will add more in the following week. Any preferences in what bookmakers should be there?
The link for my website is: *** Link removed, links are allowed only in signature when they are advertising links.
A little about me: I have a MSc. in Biomedical Engineering, where I'm trained in developing models for decision support systems for medical devices. In my current job, however, I have very little to do with modelling and statistics, and have been missing it a bit (yea - I'm weird), so I decided to create this "little" hobby project of mine - it's been roughly around 100 hours of work, but I like to do it, and my programming skills in c# have improved as well . Making models for football prediction is actually surprisingly similar to make them for diagnosis systems .
Thanks for your time, I hope to see some feedback.
PS. I did read the rules on this, And I don't believe me linking to my website is a violation of it? It's a non-commercial website and I don't earn anything of it - no registration is necessary (or even possible), it's simply surf there and pick ya numbers.
PPS. Sorry for wall of text
I need some feedback to my website that I have created. The website is my try in creating a free of charge, easy to use betting system for the major European sites.
I need your help, though, for feedback towards what people would you would like to see in the site - and a general opinion in the usability of my website.
Some background knowledge
I have collected data from 10 seasons worth of matches and selected various parameters that I have found affects the outcome of a football match. The parameters are stuff like goals scored (duh!), corner-kicks, shots on goal etc. From these I have created a model over home team victory - one that can predict the probability of a home team victory based on the previous x matches played (by both teams). From here on I calculate what I call "FairOdds", which is the minimum odds (according to my system) you as a betting player should have from a bookmaker if you want to be profitable in the long run.
What can I do with this?
When the "FairOdds" is obtained for a given match, I extract Odds from bookmakers (at the moment, I only have coded one bookmaker - but I will include more as I progress with this project ), and find the highest odds. If this odds is higher than the FairOdds we obtained, the betting amount is calculated (based mostly on the relative difference between FairOdds and the odds from the bookmaker).
The point is, that as you use this system throughout the season (and only bet when and what the system tells you), you will statistically have a profit at the end of this system... That is, if the model works .
Does it work?
Yes it does, although I have some ideas up my sleeve to improve the system, it actually does create a profit. I fitted my model with data up to August 2010, and ran all matches throughout the 2010-2011 season in the following leagues:
English Premier League
German Bundesliga
Italien Serie A
From a total of 1066 matches in the three leagues, the system advised to make 471 bets for hometeam victory, and the overall earnings divided by leagues taken through the entire season were:
Premier league: 28,2%
Bundesliga: 14,7%
Serie A: 10,5%
That is the percentage of pure profit if you use the system the entire season. I expect the Italian league to perform even better in the 2011-2012 season, since the level of training data in this league has been incriesed by 18%.
Until now, only one round in Bundesliga have been plaid, and 4 out of 5 bets have been predicted correctly. That meant a return of a staggering 129.5%. That's just one round, though - and I do not expect the bundesliga prediction model to perform that well through the entire season
.I need feedback!
As I wrote in the top, I do need your feedback as of how the usability of the site is. Is it obvious when you look at the site what to do? Is it obvious that it is only home team victory I have modelled??
What would you like to see? More statistics, filtering options?
Both the Spanish and French leagues will be added within a month or so.. The Italian is already Added, but as I only show the next week of matches, Serie A won't show up in the table (starts late August).
What filtering options would you find favorable? For example bookmaker filter (so only ones you have an account with is shown?) As I said, right now - only two bookmakers is there atm., but I will add more in the following week. Any preferences in what bookmakers should be there?
The link for my website is: *** Link removed, links are allowed only in signature when they are advertising links.
A little about me: I have a MSc. in Biomedical Engineering, where I'm trained in developing models for decision support systems for medical devices. In my current job, however, I have very little to do with modelling and statistics, and have been missing it a bit (yea - I'm weird), so I decided to create this "little" hobby project of mine - it's been roughly around 100 hours of work, but I like to do it, and my programming skills in c# have improved as well . Making models for football prediction is actually surprisingly similar to make them for diagnosis systems .
Thanks for your time, I hope to see some feedback
.PS. I did read the rules on this, And I don't believe me linking to my website is a violation of it? It's a non-commercial website and I don't earn anything of it - no registration is necessary (or even possible), it's simply surf there and pick ya numbers.
PPS. Sorry for wall of text
