This is a review of the Probit Regression Model and Least Squares Regression Model research conducted by N. David Pifer, Timothy D. DeSchriver, Thomas A. Baker III, and James J. Zhang.
Every March the men’s Division I teams from the NCAA compete in the NCAA Men’s Basketball Championship, which a six-stage, single elimination tournament. The pressure on teams and individual players is immense as there is a great deal at stake including funding and possible future careers. There are also the internal pressures of wanting to win and a strong desire to demonstrate your best skills to a large national audience. The standard belief is that the experienced player, whether in terms of playoff experience or in terms of age, will rise to the occasion while inexperienced players will falter. This assumption has not been tested so this study looks at a statistical analysis to analyze the statement.
This study used two different measures of experience. The first was prior experience in March Madness games determined by the number of minutes a team member had played in previous March Madness competitions. The second measure was the age of the player, determined by their class ranking. Other factors included were the win-loss percentage, strength of schedule, offensive and defensive ratings, coaching experience, and player height.
To test the impact experience has on winning a game, a probit regression model was developed. In the early stages of the tournament, the strength of schedule had the greatest impact on a team’s probability of winning with win-loss percentage, offensive and defensive ratings having a significant impact as well. Prior March Madness experience did not significantly influence the probability of winning in the early or late rounds. In fact, results show that experience, or age, actually has a negative effect in the later rounds of the tournament. Actually, a strong defense had the greatest effect on winning in the later rounds of the tournament.
An ordinary least squares regression was developed to look at the impact experience has on the margin of victory. Again, win-loss percentage, strength of schedule, offensive, and defensive ratings were the only variables having a significant impact on the margin of victory. Greater experience has a slight positive impact, which could prove helpful, especially in closely contested games in the later rounds.
Overall, the assumption that age and experience increase a team’s probability of winning the tournament appears to be false.
Analysts can use this information to re-evaluate their predictions regarding who will ultimately be the winner of March Madness. Teams can use this information when forming their teams, taking a closer look at the younger players to determine if they have the ability to increase their probability of winning. Obviously, while some experience is definitely necessary it is important to ensure that a team does not consist entirely of seniors, but includes a strong selection of younger players as well.
This study demonstrates the importance of testing what appears to be the most logical of assumptions to ensure their accuracy.
Analytics Used: Probit Regression Model, Least Squares Regression Model
————————
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.