This is a review of the NCAA tournament bid research conducted by B. Jay Coleman, J. Michael DuMond, and Allen K. Lynch applying probit stepwise regression.
The NCAA Tournament garners more than 60 million Americans completing a tournament bracket. It is estimated that anywhere from $60 – 70 million is bet legally every year with another $3 billion wagered through pools. ESPN held a tournament projection contest in 2014, which gained more than 11 million participants. Obviously, the NCAA Tournament is the focus of attention from the time the playoff bracket is announced until the final championship game.
During the time prior to the announcement of which 68 teams will be competing players, coaches, directors, media and fans devote a lot of time and effort into attempting to determine which teams will be invited to participate. Thirty-two teams are invited based on their winning record while the NCAA Selection Committee, which consists of ten people, namely athletic directors and conference commissioners, determines the remaining teams. They not only determine which teams will be invited but they also seed each team and place them in the playoff bracket.
Gaining an invitation to the tournament is of major importance to the teams. Many view being invited to the tournament as the indication that they have had a successful season. In addition, there are many benefits to be gained for schools, players, and coaches. Each school receives $1.58 million, paid out over six years, for each tournament game their team plays in.
The Selection Committee holds closed session when determining their selections and do not typically reveal why certain teams are chosen while others are not. It is not known which factors the committee considers to be of highest importance in their decision making each year. While the factors they focus on are not known, there is a list of all the factors that the NCAA provides to the committee for their consideration.
This research is designed to improve on earlier processes used to predict which teams which will be selected. It is based on the Rating Percentage Index which is the weighted average of: the team’s winning percentage versus Division I teams, the team’s opponent’s winning percentage, and the winning percentage of the team’s opponents’ opponents. The RPI is determined for each team using a formula that weights all wins and losses equally. The outcome is measured based on a binary variable of whether or not a team was invited to play at the tournament.
The first step involved performing a probit stepwise regression. These results were used as a basis for further analysis. The probit stepwise regression analysis selected the following predictors: RPI ranking, in-conference losses below 0.500, wins against the RPI top 25 teams, wins against the RPI teams ranked 26-50, games above 0.500 against teams ranked 51-100, as well as the number of road wins.
Next, two additional model selection processes were used to sort through the data to create alternative factors for comparison. Neither alternatives were better predictors and therefore the original method was retained.
The obvious use of this statistic is that everyone involved will gain a greater ability to determine who is likely to be selected to join the tournament. Coaches can decide which factors they need to focus on in order to increase their chances of being chosen.
Finally, the fact that this model is able to accurately duplicate the decision of the Selection Committee over the past years is a clear indication that the process employed by the committee is very consistent.
Find out how Sports Analytics Expert Victor Holman can give your team the competitive advantage.
How mature is your team’s analytics program? Take the Sports Analytics Maturity Assessment.
Discover the Groundbreaking Sports Analytics Application and Framework coaches and sports analysts are talking about!
Learn all about sports analytics in Victor Holman’s Sports Analytics Blog.