ESPN commentators will often insist that such-and-such team has such-and-such "momentum," so really, there's not much of a question when it comes to tonight's game.
Behavioral economists have long attempted to figure out whether this clichéd concept that sports fans have long endeared themselves to can actually, statistically, be a predictor of a team or player's success. The results have proven inconsistent.
A 2006 review concluded that momentum is more likely to present itself in sports "where there is minimal or non-existent opportunity for opponents to interfere with player performance."
One study demonstrated that pro golfers can indeed go on "non-random streaks." Another recent study that looked at in-game momentum in the NFL discovered no evidence of it after turnovers and allowed scores, but did see an advantageous effect after a team successfully converted on a fourth-down. A 2006 review concluded that momentum is more likely to present itself in sports "where there is minimal or non-existent opportunity for opponents to interfere with player performance."
Kevin M. Kniffin, a researcher at Cornell University's Dyson School of Applied Economics and Management, and his assistant Vince Mihalek, a recent graduate and four-year player on the school's hockey team, have added yet another entry into the empirical sports momentum literature by analyzing the results of 458 collegiate two-game hockey series (916 games total) played in the Western Collegiate Hockey Association (WCHA). The pair wanted to figure out if performance in the first game would statistically predict the outcome of the second. In the study, which will appear in the March edition of Economics Letters, they contend that this particular data set works perfectly for studying momentum.
Of the six conferences for Division 1 men’s hockey, one of them–the Western Collegiate Hockey Association (WCHA)–uniquely schedules series in which a given team will host another team for two consecutive games in the same location with uniform amounts of time (i.e., 1 day) between the games. Given that teams take turns across the course of the season hosting and traveling to compete against other squads, there is also a balance in the scheduling to guard against any effect of home ice advantage. The conditions in the WCHA consequently make it ideal for testing whether momentum exists within series from the outcome of Game 1 in relation to Game 2.
When the researchers controlled for the quality of the team by taking into account end-of-year conference records, they discovered that a win in the first game, no matter how excessive in terms of outscoring, did not predict a triumph in the second. "In fact, our finding that a team that wins the first game by 2 goals is slightly less likely to win Game 2 is affirmative evidence that positive momentum, at least, does not exist in our sample of series," the researchers conclude. So when it comes to collegiate hockey in the WCHA at least, momentum doesn't seem to exist. Sorry hockey fans.
Of course, these conclusions aren't generalizable to other sports, or leagues for that matter. And the findings also support earlier suggestions that momentum's power doesn't extend to contests, like hockey, in which an opponent can directly "respond to offensive maneuvers." Bowlers on a streak needn't worry—yet.