After rooting through a box of papers for an article on baseball pitching probabilities, Gilbert Fellingham returned to his office chair surrounded by three computer monitors and enough number-filled books to make anybody nauseous - an unlikely habitat for a sports nut, but just the place where academia and athletics can be reconciled after sitting at different cafeteria lunch tables for so long.
"Statisticians who do sports stuff are kind of the redheaded stepchild in statistics," said Fellingham, a BYU professor of statistics. "You're kind of the egghead to the sports guys and you're kind of the dumb guy to the statisticians."
Utah Jazz coach Jerry Sloan might agree.
"Statistics screw me up," he said. "I wasn't smart enough to pass arithmetic. I just watch games; I never saw a statistic yet."
BYU football coach Bronco Mendenhall, on the other hand, has a different philosophy.
"We have three pillars that we've established that are based completely on statistics," Mendenhall said. "We did our own study to determine what three statistics determine whether you win or lose college football games at BYU over the last 20 years. We've come up with our own, and I don't share them. They've been great predictors. I believe facts are our friends."
In any case, Fellingham uses his statistical know-how in sports as often as he can, including work for the U.S. national volleyball team. He and Shane Reese, a fellow BYU statistics professor by day and sports fan by night, recently teamed up with former BYU student Garritt Page for a study of the NBA that wound up in the Journal of Quantitative Analysis of Sports.
"We associated the categories in the box score with the final point spread in the game," Fellingham said. "We essentially paired up the players, and said, 'Point guard against point guard -- who got more assists? Who scored more points? Who got more rebounds?' And we used that to model the final point spread."
Together they devised a statistical formula that would factor in different coaching philosophies, talent levels and even home court advantage, and then ran it across 1,163 box scores from the 1996-97 NBA season.
"Gathering the data was problematic," said Page, now a graduate student at Iowa State University. "Data is a hot commodity. We had to hire somebody to parse the web for it."
Their chief finding was the importance of team assists, which, by the numbers, outranks even total field goals in the winning formula.
"To me, the most important message from that study is that assists are more highly correlated with score spread than anything else," Fellingham said. "That would indicate that ball movement matters a lot. Even in the NBA, where you think about star power, it's ball movement and guys who can deliver the ball that really have value ... If you can out-assist the team you're playing, you really increase your probability of winning the game."
"That's why the Jazz are so much fun to watch," BYU basketball coach Dave Rose said. The Jazz, next to the Phoenix Suns, are at the top of the NBA in team assists and both teams are playoff-bound.
The study's other conclusions were less intuitive. Assuming each position is more or less fulfilling its fundamental role, the statisticians found a large impact at the small forward position, where assists were more beneficial, and turnovers were more detrimental, than almost any other contribution on the box score. Even more unexpected might be the value they saw in steals at the center position and defensive rebounds from the guards.
This might explain how Cleveland Cavaliers small forward LeBron James can carry his team so well. Not only does he lead the NBA in scoring with 30.2 points per game, but he also leads all small forwards with 7.3 assists per game. On the other hand, James also leads all small forwards with 3.36 turnovers per game. Given that, the Detroit Pistons' Tayshaun Prince just might be the most valuable player in the league, considering he has the highest assist/turnover ratio, at 3.18, of all starting small forwards.
"It takes a lot of energy, especially from the forward, but if you're aggressive and active, it's doable," Jazz small forward Andrei Kirilenko said of helping move the ball at the small forward position.
Nevertheless, there are certainly limits to the applicability of the study. The basketball box score, for example, has no column for things like hustle, tipped passes or contested shots.
"Nowadays in a box score the main thing people look for are points," said Jeff Rich, a BYU junior and a lifelong Jazz fan from New Jersey. "But guys like [BYU's Chris] Collinsworth or [the Jazz's] Ronnie Price may not get a ton of points or minutes, but they bring a ton of energy on the floor."
"You kind of have to work with what they record ... you value what you grade," Fellingham said. "If you look at pay for linemen in the NFL, you'll see that left tackles started to become highly paid shortly after sacks started to be counted."
