Forecasting Hard Questions
By Ben Golden on December 09, 2015
I've been trying to pick NFL
game winners. I'm not using any complex analytical model; rather,
I'm making decisions the way most sports bettors do--I watch some games,
read the news, and use my judgment. I make each of my picks on SportsCast, which allows me to
track my
performance, interact with other forecasters, and track the performance of
the prediction market--that is, the collective performance of all the
forecasters on SportsCast.
Picking winners is hard--note that we initialize prices for these questions based on online odds. Most people who try to beat the Vegas sports book end up losing money. Those who succeed usually do have a complex analytical model, or a system. The level of challenge is much higher than other types of forecasting, and is comparable to beating the stock market. I don't know whether my forecasts will ultimately be winners or losers, though I suspect in the short run, they'll cost me. To avoid losing all my (fake) money, I'm making fairly small forecasts, while continuing to make larger investments elsewhere.
I'm closely tracking my results (so far I'm down), but I'm much more interested in 1) whether I can get better over time, and 2) whether the prediction market as a whole outperforms the initial odds. There aren't a ton of users currently forecasting on these questions, but as SportsCast continues to grow, it's plausible that the prediction market becomes highly accurate, even if many individuals aren't.
Through the rest of the football season, I'll publish updates on my accuracy, market accuracy, and how they compare to the Vegas baseline. And if the prediction market does well, maybe we'll eventually take our "system" on a road-trip to Vegas (and compensate our forecasters for their contribution).
Join me in forecasting NFL games by signing up for free at SportsCast. If you liked this post, follow us on Twitter at @cultivatelabs, and sign up for our monthly newsletter.
Ben Golden/@BenGoldn is an Engineer at Cultivate Labs.
