Big Data Finds What Football, Basketball and Hockey Have in Common
What do basketball, football and hockey have in common? On the surface, not very much. After all, one sport is played on ice and the other two are played on either a grassy field or a wooden court.
But according to University of Colorado computer science professor Aaron Clauset, when the games are analyzed from a Big Data perspective, patterns and similarities between the sports begin to emerge.
“These games look a lot less complicated than most people think,” Clauset says in a recent Slate article.
As a Big Data superstar, Clauset’s opinion carries weight, and his recent paper, “Scoring dynamics across professional team sports: tempo, balance and predictability,” which he submitted to the Journal of Quantitative Analysis in Sports in October 2013, reveals the game mechanics at work in each sport.
“[In all three sports], events occur randomly (a Poisson process). Which team wins the points is coin flip (a Bernoulli process) that depends on the relative skill difference of the teams on the field,” Clauset and his co-author, Sears Merritt, write.
Clauset and Merritt also do a bit of myth busting as they tackle the popular concept of teams gaining momentum in a game.
“[G]ameplay is largely [a] sequence of roughly independent, short-term optimizations aimed at maximizing near-term scoring rates, with little multi-play strategic efforts and few downstream consequences for mistakes or miscalculations,” he writes.
Translation: There’s no such thing as momentum or a hot streak.
What the Data Say About Team Sports
While Clauset and Merritt found a lot of things the three sports had in common, the researchers also found some differences. Between college football and professional football, they found that college games were more unbalanced than NFL games. This makes NFL games more unpredictable, because the skill level is equal and a team’s win or loss is largely left up to chance.
“CFB games are much less balanced than NFL games, suggesting that the transition from college to professional tends to reduce the team skill differences that generate lopsided scoring,” Clauset and Merritt write.
Of the three sports, basketball is the most unpredictable. This is due to the unique pattern of lead sizes, or spreads, in NBA games.
“Pro basketball is the only sport where the spread tends to shrink. In football and hockey, the spread tends to grow over time,” the authors write.
In his review of Clauset and Merritt’s work, Slate writer Joel Warner points out that these theories were actually put to test, and they worked:
Clauset and Merritt developed a mathematical model that, after observing just a few scoring events, predicted game outcomes for college and pro football, the NHL, and the NBA with surprising accuracy.
Their model proved more accurate than the simple metric of looking [at] who was in the lead at a given time, and it outperformed SportsbookReview.com’s pregame betting odds while more or less matching the accuracy of the live-betting site Bovada. Impressive results, considering Clauset and Merritt spent just three months analyzing the data and coming up with their model.
The success of their model suggests that tuning in to the frequencies of tempo and balance in sports produces predictable results across the board.
The data that Clauset and Merritt have compiled gives sports fans plenty to chew on, but it doesn’t reveal anything we didn’t already know about team sports: that winners are, for the most part, determined by skill and luck. The other stuff — team curses, losing streaks, etc. — doesn’t really matter.
