Interesting Article on Making Money in Fantasy Sport.

The New York Time posted a story about a guy who has made a reasonable profit on Fantasy Sports sites. Here is a link to the Site.

I have done work in this area both for fun and profit. Years ago Drew Conway started a machine learning fantasy football league which I played in. I think it is fair to say that the results of that league were unimpressive even though it had within it some the most successful predictive analytics guys in the country. Later we did some work on picking inefficiencies in the sports betting market which was successful. It made a little bit of money week after week. It was a interesting project, but it simply was not profitable enough to justify the work involved.

Over time we learned a little about which sports are easier to do prediction in. The hardest by far of the pro sports was the NFL. The relatively few games in the NFL is a big reason for this.  However the NFL is not equally difficult throughout the season. The first three and last three weeks are the one that are the hardest to predict with the middle weeks being reasonable. Interestingly enough we found the NBA with its numerous games, small number of playesr and other factors was consistently the best data to work with either on a player level for fantasy or on a team level for sports betting. Although we also found that the sport betting market for the NBA was an efficient one compared to other sports so any change needed to be taken advantage of quickly. I found it interesting that the NBA is the sport that the New York Times article focused on because it is where I would focus on as well.