Analyzing the SuperContest

This fall, the Westgate Las Vegas Hotel Casino is hosting the prestigious SuperContest, in which entrants pony up $1500 to pick 5 games a week against the spread. At season’s end, standings are judged by which entrants have the highest number of correctly picked games.

Last year’s winner was David Frohardt-Lane, who took home about $550k for first place. Perhaps not surprisingly, Frohardt-Lane is a statistician!

One of the neat aspects about the SuperContest is that the picks of all 1,403 of the 2014 entrants are posted immediately after the games begin (here). Even better, the data is fairly clean and easy to play with (at least so far).

That said, there are some very intriguing and potentially difficult questions to answer from a statistical perspective: How to account for the fact that people only pick 5 games per week? What about bye weeks? Do entrants pick each week seemingly by chance or are there trends over time? Are there teams that are picked more or less often? And how about success – do entrants perform better or worse than we would expect by chance? Lastly, does success beget more success?

I’ll hope to answer a few of these questions over the course of the season, and I encourage any interested readers to do so, too! Here’s a link to the google doc where I’ll post the picks (csv file). From what I can tell, the hotel only posts picks for the previous week, so I’ll do my best to update the file weekly.

For a taste of what one could do, I decided to plot Week 1 and Week 2 picks by team. The size of the circles in the graph below are proportional to the number of entrants which chose that team, and the colors (for Week 2 only) represent whether or not that team had covered the spread in Week 1 (Green is yes, red is no).

For reference, 38% of participants took the Patriots in Week 2; only 6% of participants had Carolina in Week 1. 

My hypothesis was that teams which covered in Week 1 would be perceived as “hotter” and would be picked more often in Week 2. Hard to tell if that’s the case, although perhaps if you remove the Patriots as the outlier, you could make an argument that the green circles are larger than the red ones. It’s worth pointing out that lines are fixed on the Wednesday prior to games beginning, and so news of Adrien Peterson’s transgressions perhaps moved entrants to pick the Patriots at a higher rate. 

Also, it’s amazing how many entrants lost with Tampa Bay in Week 1 and didn’t pick the Bucs in Week 2.

Thanks for reading, and if you have any ideas for graphs or analyses, please feel free to share below or send me an email!