UC Master's Candidate Adds a New Twist to Sports Talk: Monday Morning Quantitative Analysis

Paul Bessire, graduate student in the University of Cincinnati’s College of Business, thinks he has the answers.

Bessire graduated in 2004 from the College of Business Carl H. Lindner Honors-PLUS Program with a BBA in finance and quantitative analysis. After receiving his undergraduate degree, he was fortunate enough to have enough quarters of scholarship to attend UC for grad school to study for a Master’s of Science degree in quantitative analysis (MSQA). The MSQA degree combines applied mathematics, statistics and computer applications in a business environment. Including what is often called operations research or management science, as well as applied statistics, this unique MS education helps students excel in all corners of the business world — not to mention sports.

“My love of numbers and the fact that I could apply them to sports led me to the MSQA program. On June 10, I will graduate with my second degree and take a full-time job doing data analysis for WhatIfSports.com, a sports simulation company,” says Bessire.

On June 3, Bessire presents the results of his research in his paper “Measuring Individual and Team Effectiveness in the NBA Through Multivariate Regression.”

“We believe that much of the variation found in a basketball team’s success can be explained mathematically through looking at the interactions of the five players on the court and not just individual player abilities,” says Bessire. For this project, Bessire examined several methods for rating individual NBA players and used multivariate regression analysis to assist in building successful NBA teams.

Multivariate regression is a mathematical approach to comparing many variables that have an effect on the outcome of a situation.

“Multivariate regression is probably much simpler than it sounds,” says Bessire. “We try to predict a dependent variable, in this case, the average scoring margin per minute for a five-player lineup, based on a series of independent variables, the attributes of the five players in the lineup. The average scoring margin per minute illustrates how effective a lineup is (e.g., a positive value denotes a team that usually outscores its opponent). The attributes of the players are measured by commonly kept basketball statistics. What makes the regression model unique is the allowance for interaction between the dependent variables. Having a point guard who can pass well on the court with a shooting guard who can shoot well may have a different impact than summing the impacts of having either attribute, but not the other. We find the mix of player interactions that most accurately predicts average scoring margin per minute to produce our final model.”

That’s simple??!

“In terms of the analysis, baseball is a sport that has been widely researched and modeled,” Bessire says. “It is difficult to find new, unique and fresh problems to model in baseball. While  the problem associated with my paper, modeling teamwork in basketball, is not new, very little has been done to explore it mathematically. In fact, applying statistical analysis to basketball is just now coming into vogue and only began within the last ten years.”

Applications of the model include examining which players should play at each position, predicting the lineups that should have the greatest team success and specifying which skill areas the coaching staff should seek to improve through the annual NBA draft, free agency and trades. For example, the paper applies the model to some trades that occurred in February 2005.

“For instance, it shows that a trade that sent Chris Webber, Matt Barnes and Michael Bradley of the Sacramento Kings to Philadelphia for the 76ers’ Brian Skinner, Corliss Williamson and Kenny Thomas benefits both teams,” Bessire says. “The 76ers get offensive help on the interior, something they needed immensely, and the Kings receive interior defense in return. Previous to the trade, the Kings had two similar players starting down low with Brad Miller and Webber. The 76ers started Thomas and Samuel Dalembert, also two similar players. Redundancy is inefficient so both teams improve.

Conversely, in the trade of the Warriors’ Speedy Claxton and Dale Davis for the Hornets’ Baron Davis, neither team is helped (on the court). The analysis explains that the point guard position is overvalued in the NBA. Any team that has a competent point guard, one that can pass well — which is almost every team — should not worry about an upgrade at that position. NBA teams do not differentiate themselves at the point guard position. Passing ability is important, but it is generally more important for a lineup to upgrade passing ability at the power forward and swing forward positions.”

OK, surely Bessire did not gain all his sports knowledge at UC. He grew up in Wisconsin, home of the Green Bay Packers. With two brothers, he frequently played pick-up games of basketball. When it came time for him to choose a college, he narrowed his list to Duke, Arizona, Utah, UC and William & Mary.

“Four of those schools were ranked in the top 25 in college basketball,” Bessire points out. “UC and Duke were Nos. 1 and 2. I chose UC because of the great honors business program, but I probably would not have known about the school had it not been for the national attention brought upon it from its success in basketball.”

So did Bessire have any predictions for the playoffs?

"Unfortunately, the model cannot be used too effectively in predicting the NBA playoffs,” he says. “The limited number of games allow for too many other factors to play a role in the outcome. The analysis should provide a very good indication of a team lineup’s effectiveness over the course of the season. It can also help rank players for the NBA draft and provide a tool for NBA General Managers looking to improve through free agency.”

If you want to know more about quantitative analysis of basketball, visit Bessire’s Web page: www.WhatIfSports.com.

About the Master’s of Science in Quantitative Analysis