In his study of whether to go for it on fourth down, Dr. David H. Romer of the University of California checked whether momentum was a factor. It was not. His discussion of his momentum analysis was only five paragraphs long, but I thought it merited its own separate article.

Studies of momentum in other sports

He says that three studies of momentum in other sports indicated there was little or nothing to the notion of momentum. Those studies are:

• Gillovich, Thomas, Vallone, Tversky “The Hot Hand in Basketball: On the Misperception of Random Sequences” Cognitive Science journal July, 1985
• S. Christian Albright “A Statistical Analysis of Hitting Streaks in Baseball” Journal of American Statistical Association December 1993
• Klaasen and Bagnus “Are Points in Tennis Independent and identically Distributed? Evidence From a Dynamic Binary Panel Data Model” Journal of American Statistical Association June 2001

In baseball, the notion of “clutch hitting” is similar to momentum. According to baseball analyst Bill James, “…clutch hitting does not exist (because measures of clutch performance independent of actual batting skill have zero reliability from season to season)…”

Definition

To analyze a problem you must first define it. Romer defines momentum as superior performance by a team after a very good thing happening to them or lousy performance after a very bad thing. My words, not his.

He defined a very bad thing as a change of possession where the field position changed ten yards or less, e.g., turnover on downs, fumble. And he defined a very good thing as a touchdown. Each definition gave him over 600 NFL plays to examine the aftermath of.

Reverse momentum

Not only did he find that the team that suffered the very bad thing did not do lousy on the next several plays, he found that they actually did a little better than average. In other words, the notion of momentum is not only untrue, it is, if anything, the opposite of the truth. To state it in broadcaster terminology, teams that lose momentum actually do a little better after the momentum-losing play.

Counterintuitive

I recognize that this is counterintuitive. We have all experienced the let-down of a bad play either as fans or as players or coaches. Surely, our team must do worse when they have those let-down feelings running around their brain.

Well, if it were so, it would show up in these various analyses. In fact, there are random streaks in any observable data. For example, if you flipped a coin 100 times, there would be some heads streaks during the hundred flips. Is that because the coin has momentum? Was the coin starting to feel good about itself after a couple of heads? No. It’s because stuff happens randomly but sometimes looks non-random.

Anger

Also, very bad plays also anger players. They may play harder after their team screws up. Then the question would be whether the very bad play’s ability to engender discouragement exceeded its ability to engender anger.

Counter-momentum training

Athletes, even at the amateur youth levels, are experienced and trained in counter-momentum thinking if you will. They have experienced hundreds of very bad plays by the time they reach high school. They long ago were told to forget the last play and focus on the next one. Don’t worry about it. Get the next one. At the time it happens, both their teammates and their coaches instantly launch into this verbal routine.

Instant change of personnel

In football, very bad plays are typically committed by either the offense or the defense. As soon as the very bad play occurs, the unit that made the mistake is removed from the field and the other unit comes onto the field. The guys in the game for the next several plays are innocent of the very bad play. They are probably angry about it, but their anger is directed toward their teammates on the other unit and towards the other team, not at themselves or their own unit.

Beginning athletes only

There may be momentum in tee ball, but by the time athletes reach high school, they have been thoroughly trained and experienced in the one-play-at-a-time mentality. If the data is any indication, they are sufficintely trained and experienced that whatever effect momentum may have had is nullified.

My attempt at capitalizing on momentum

I once tried to capitalize on momentum. We got a devastating pass interference call against our opponent. It gave us a key first down and great field position. So I immediately call 32 pop pass figuring the opponent would be demoralized and playing at a subpar level. They intercepted it and we lost the game by two points.

That’s only one play. You need more data than that to prove there is no momentum. But it is at least one concrete example that belies the abstract notion that momentum occurs whenever something bad happens to one team on the football field. We’ve all seen dramatic turnovers in games that were followed on the very next play or two plays later by another turnover going the other direction.

If this Football Think Tank Web site is about one thing it is relying on the data, not intuition. If the intuition cannot be supported by the data, which it cannot in the case of momentum, it is a childlike mistake to stick with intuition.

To contribute an idea or comment to this Football Think Tank web site, either email to johnreed@johntreed.com or fax to 925-820-1259 or snail mail at 342 Bryan Drive, Alamo, CA 94507.

Best wishes,

John T. Reed