Thursday, April 4, 2013

Sports Analytics and Big Data

I am a huge fan of fantasy sports and sports analytics so when I hear of the MIT Sloan Sports Analytics Conference I immediately delve into their history of speakers. This conference hosts numerous panels spanning several sports and facets of sports as well as the evolution of sport and research papers. One panel that I found most interesting was the fantasy sports analytics panel. This panel featured Peter Schoenke, Jonah Keri, Joe Bryant, and the man with my dream job Matthew Berry. Each of these guys expertise is in fantasy sports. A common theme throughout their session was finding the "new stat" to get a leg up on your competitors. Matthew Berry, fantasy expert for ESPN, mentioned a "Points Per Touch" statistic for each NFL player. Fortunately for Berry he has the ESPN statistics department at his disposal. This is an interesting approach to look at. Sports statistics are easily attainable and span back many seasons.,, and one of my favorite sites provide all stats spanning back many seasons.

Creating these new statistics for further insight into players can be very useful for fantasy drafts. The hardest part is finding which stats and situations will provide the best opportunity for fantasy points. As the panelists mention during their session football is hard to predict because there are so many factors to look at including defenses, offensive schemes, weather conditions, etc. as opposed to bowling where the conditions are always the same and involves a single player at a time. I have come up with an "Opportunity Factor" which is essentially the amount of involvement a football player is within his teams offense. For example a running back's opportunity factor would equate to his (number of receiving targets plus number of rushing attempts) divided by (the teams total passing attempts and rushing attempts.) I hope to find this as a great indicator of useful players in fantasy football and also encourage others to find that great new statistic. Below is the link to the video of the panel I referenced to. It's a great watch for anyone who is very involved in fantasy sports.

1 comment:

  1. There are a lot of sport betting companies all over the world. They play with a huge amount of money. First of all create a pool of sport game results and any detail which might have affected the results; whether condition, date of the game, team coaches, all the information about players, performance histories, etc. By using data mining techniques they can predict the result of any game with 95% accuracy rate and so they can make money easily. Dr John Goddard from University of Wales Swansea published a paper called ‘Modelling football match results and the efficiency of fixed-odds betting’ which gives interesting details about how a statistical model help to predict the result of a sport game. Here is a brief abstract of the paper:
    “An ordered probit regression model estimated using 15 years’ data is used to model English league
    football match results. As well as past match results data, the significance of the match for end-ofseason league outcomes; the involvement of the teams in cup competition; the geographical distance between the two teams’ home towns; and the average attendances of the two teams all contribute to model’s performance. The model is used to test the weak-form efficiency of prices in the fixedodds betting market, and betting strategies with a positive expected return are identified. “