Soccer Statistics

This software application was developed for a consortium of individuals who had created a mathematical soccer betting model. The problem consortium members faced on a daily basis was that the data required to feed the model was obtained from multiple Internet sources and had become very time consuming. When the consortium approached me members were spending two to six hours each per day collecting the required data. Today these same consortium members spend 30 minutes each morning overseeing the software's automated processes.

Primary Results Source
The consortium's primary source of results was www.soccerway.com. The time taken to navigate to each league's results page and cut and paste the results data into their spreadsheets was very time consuming, and prone to error. Now with a single click the software's automated website navigation and data collection abilities means that the user just sits back and oversees the data quickly and accurately being collected

 
(click for full size)


Secondary Results Source
Occasionally the Primary results source (www.soccerway.com) would be delayed in the publishing of results. This meant that a secondary source of data was needed to cover any missed results from the primary source. The site chosen was www.futbol24.com and as with the primary source any missing results are collected with a single click.

(click for full size)


Primary Odds source
The complex mathematical model that I coded to the consortium's specifications required that accurate odds of the soccer market be known after the matches have been played so they can be fed into the system. The Primary odds source was www.betexplorer.com and with a single click the software's automated processes quickly and accurately imports the odds from all the required leagues.

(click for full size)


Secondary Odds source
As with the primary results source sometimes the primary odds source would delay publishing the odds so a secondary source was chosen. The selected secondary source was www.oddsportal.com and as with the primary source a single click is all that is need to collect any odds data missed by the primary source.

(click for full size)


Bayesian Rating page
The rating system is based on a Bayesian normalised distribution statistics engine and required a complex set of formulas to solve (see here). The formula complexity problem was solved by using and online mathematics source www.easycalculation.com. With a single click the software's automation processes submit the required inputs derived by the software and record and save the results for the next step in the overall mathematical model.

(click for full size)


Regression Analysis
Before the consortium started using this software this was the most time consuming part of the mathematical modeling of the data. With many hundreds of matches taking place, on some days each individual match required it's data to be analysed using spreadsheet software. Now a single click will analyse over ten years of past performance data for each league and compare the predicted result with odds being offered by Betfair a major online odds exchange an then advise whether the match should be bet as a home win, away win or a draw.

(click for full size)


Bet Placement
After regression analysis bet placement was the second most time consuming activity of the consortium's betting model. Betting sites require navigation to each event's betting form and the entering of required information such as stake and bet type. The software's use of a betting sites application programming interface (API) if available or automatic navigation and form submission when not available. Means that placing a bet with all it's associated information is now a single click event.

(click for full size)


Profit/Loss Analysis
Each consortium member monitors their own betting profit loss account via their associated betting site accounts but an overall profit loss performance of the entire betting model was required. This was previously a tedious and lengthy chore but now is achieved by a single click providing an at a glance graphical interface to provide a very accurate overall health check of the entire betting model over time. Supporting information to the graphical display is provided in the form of a statistical Chi Squared Test. As well as a break down of individual leagues financial performance.

(click for full size)