David Frohardt-Lane's Talk on "Predictive Modeling in Sportsbetting"
PROMYS 25th Summer Celebration
July 6th, 2013 at 10am

Abstract: In this talk, I will talk about the process of building a prediction model. The model in question will be designed predict the outcomes of NFL games, using only box score statistics in previous games, and built with the intention of identifying profitable wagering opportunities. The focus of the talk will be on how to extract as much information as possible from a limited data set. This should be easy to follow; I won't be assuming prior familiarity with statistics or football.

Bio: David spent two years at PROMYS, in '94 and '95. After PROMYS, he went to Carleton College where he majored in mathematics and then went on to get a MS in statistics from the University of Chicago. In 2004, he ended up at a proprietary trading company called GETCO, and spent the next 8 years there, working out of their Chicago, NY and Singapore offices and trading a wide variety of financial products. As of July 1, he has started a new job with 3Red, a small trading firm in Chicago. David have always been interested in sports analytics and for a period of time (2003), he was making the majority of his income from betting on baseball. While he no longer gambles (the legal situation has changed since then) he continues to dabble with sports prediction models as a hobby. Recently he has started consulting with an Major League Baseball team to help them project high school and college players for the amateur draft. A common theme across these endeavors is building quantitative models to make predictions.

Loading more stuff…

Hmm…it looks like things are taking a while to load. Try again?

Loading videos…