“Momentum” might be one of the most common theorized causal agents in the sports lexicon. Watch any given hockey game and you’re sure to hear the color analyst talk about which side has all the momentum or the potential momentum change that could be caused by a penalty kill or timely goal. The assumption is that psychological alterations brought on by big, unique or potentially outcome determining plays can improve or inhibit the performance of athletes to a significant degree. Players on a team with momentum with feel good, confident and focused. Players competing against the flow of momentum may feel less confident or deflated. Momentum should therefore not only be something positive for teams to strive for, but predictive in terms of success.

The reality doesn’t really back the assumption, however. Take the Calgary Flames recent string of games for instance. In Montreal, the Flames rapidly fell behind the Canadiens 4-0 by the middle of the second period. With each subsequent goal, the Habs seemed to take further control of the game. The assembled Canadiens fans began singing their familiar “Ole Ole!” gloat after Miikka Kiprusoff was pulled. They seemed to have all the momentum.

That is, until the Flames reeled off three consecutive goals of their own, before tying the contest three minutes into the third period. Somehow, they had turned the tide of momentum to their favor…just long enough to lose the game in overtime. Despite the loss, the team and Flames fans were energized by the unlikely comeback. The club took the momentum gleaned by Montreal game and promptly lost 6-0 to the Minnesota Wild in the very next contest. They were so crushed by the let down of that defeat that they went on beat the Dallas Stars 7-4 two nights later.

As you can see from this small slice of games, neither inter nor intra-game momentum seemed to exist. Or, if it did, it was so overwhelmed by other factors so as to be utterly irrelevant.

Of course, anecdotes don’t equal data. Folks have tried in vain to identify a true momentum effect in sport through larger scale studies, again without finding much. One of the most famous was the research paper done by Dr. David H. Romer entitled “Do Firms Maximize? Evidence From Professional Football”. Although the enduring conclusion of the paper was that NFL coaches as a group are far too conservative, Romer dedicated a section of the study to momentum and it’s effects on the game:

it is possible to obtain direct evidence about whether outcomes differ systematically from normal after plays whose outcomes are either very bad or very good. To obtain a reasonable sample size, for very bad plays I consider all cases in which from one situation to the next (where a situation is defined as before), possession changed and the ball advanced less than 10 yards. For very good plays, I consider all cases in which the offense scored a touchdown. These criteria yield 636 very bad plays and 628 very good plays. I then examine what happens from the situation immediately following the extreme play to the next situation, from that situation to the next, and from that situation to the subsequent one.

The results provide no evidence of momentum effects. All the point estimates are small and highly insignificant; the largest t-statistic (in absolute value) is less than 1.3. Moreover, the largest point estimate (again in absolute value) goes the wrong direction from the point of view of the momentum hypothesis: from the situation immediately following a very bad play to the next, the team that lost possession does somewhat better than average.

So not only did Romer find no real momentum effect, his data actually suggested there is a reverse-momentum effect: that is, teams tended to rebound slightly after bad plays rather than performing worse. Something similar occurs in hockey. Stats-focused writers have recently dubbed it “playing-to-score effect” which refers to the tendency for teams to decrease their attack while leading or increase their attack while trailing. Gabriel Desjardins outlines playing-to-score here. The tendency appears throughout the data – teams with the lead generate fewer shots. Teams playing from behind generate more shots. The bigger the lead or deficit, the larger the effect. It’s so consistent and pervasive that most quantitative hockey analysts are starting to consider possession and scoring stats with the lead tied as a truer indication of a team’s ability.

Why does this happen? Playing-to-score is probably caused by a number of things on both ends: the leading team sitting back and taking fewer risks while the trailing team shortens it’s bench and takes more risks, for example. Of course, this is technically the opposite of the theorized effect of momentum on performance: a team with the lead we can assume has the momentum on their side, while the trailing team doesn’t. And yet, the data in football and apparently in hockey suggests that’s not really the case.

Why is “momentum” so prevalent in sports reporting and analysis then? Probably because it lends a narrative arc to the game/season/playoff series, etc. When explaining events after the fact, if someone says “but then TEAM X scored a goal and swung the momentum”, the audience can anticipate the end of the tale. It is the turning point in the rising action, the thing that predicts the final resolution of the conflict. Besides, when we see something like a a big hit or an end-to-end goal that seems to spark a comeback or deflate the opposition, the drama of the outcome is too compelling to ignore. Thanks to confirmation bias, that’s the example we think about the next time the color guy talks about the next potential momentum swing and not the hundreds of other instances when an exciting or calamitous play failed to change or predict the course of the contest.