This talk will cover two research areas. The first examines predictive modeling for the NCAA tournament. Each year, more than $3 billion is wagered on the NCAA Division I mens basketball tournament. Most of that money is wagered in pools where the object is to correctly predict winners of each game, with emphasis on the last four teams remaining (the Final Four). Joel and his colleagues have developed a combined logistic regression/Markov chain (LRMC) model for predicting the outcome of NCAA tournament games given only basic input data. Over the years, the model has been significantly more successful than the other common methods such as tournament seedings, the AP and ESPN/USA Today polls, the RPI and the Sagarin and Massey ratings, and a modified version of the model is used to provide rankings to the NCAA tournament selection committee.

The second research area will examine decision support for the MLB draft. Preparing for the annual Major League Baseball draft is a difficult task. With 1500 players selected each year, teams need to evaluate and rank many hundreds of potential draftees. To evaluate the players, teams send out scouts, baseball experts who make qualitative and quantitative observations and report their opinions to the team. However, scouts often disagree in their opinions. Joel and his colleagues worked with a Major League team to model and solve the problem of suggesting a consensus ranking of all players scouted by the teams representatives. The methodology can also make in-season recommendations for dynamic scout scheduling based on the level of information each scout is likely to provide on each player, and the uncertainty in the correct overall ranking of each player. The team has been using the optimization tool for the past two years, and a second Major League team has asked to evaluate their ranking data as well.