The Myth of Momentum

Posted with permission from myself (here). Blizburgh told me to post this on the front page but neglected to tell me that I don't have front page privileges. He's stupid. This is a post about Momentum, specifically, about how the word gets thrown around a lot. I am of the opinion that Momentum is a myth, though readers are encouraged to disagree with me. If you don't care about momentum, you should definitely not read this. Even if you do care, you should probably not read this, as it is boring, poorly written, and lacks a clear organizational structure, which is typical of the things I produce on my keyboard. Enjoy [but you probably won't],

There it is:

force or speed of movement; impetus, as of a physical object or course of events: The car gained momentum going downhill. Her career lost momentum after two unsuccessful films.

Len Pasquarelli almost got it (inadvertently, as you'll see):

More than 40 years ago, then-Supreme Court justice Potter Stewart, speaking on pornography, conceded that he could not adequately define the issue being debated in front of him and his eight colleagues. "But," Stewart said, "I know it when I see it."

 

Same can be said for momentum, which is arguably even harder to identify.

I say almost right because he adds that "arguably" into it and suggests people can identify momentum if and when they see it. He provides one example:

But if you were around the Dallas Cowboys on Sunday evening, squeezing through big bodies and stepping over shoulder pads and rolled up athletic tape in the hole-in-the-wall niche that serves as the visitor's locker room at Giants Stadium, you could actually see momentum. It was there, tangible, palpable, real and undeniable, for everyone toting a tape recorder or a minicam to witness.

Of course by "Sunday evening" he meant December 3rd, 2006 after the Cowboys beat the Giants 23-20. The, er, tangible, palpable, real and undeniable result of that momentum was a 42-17 loss to the New Orleans Saints a week later, their worse loss of the year. After that, er, visible momentum the Cowboys went 1-3 culminating in their humiliating one point loss to the Seattle Seahawks on a botched extra point hold by Tony Romo.

So, thanks Len, Mr. Pasquarelli ultimately did have something very meaningful to say about momentum, namely, even the people who are allegedly most qualified to identify it can't do so. Yet it is always assumed to be present; Dan Wetzel has an excellent article in Yahoo Sports (hat tip: Chris Mottram @ The Sporting Blog) discussing the merits of the two point conversion (genius!):

There are so many variables in a football game – momentum, weather, matchups, fatigue, play calling, etc. – that each 2-point conversion should be treated as an isolated action and not part of some statistical trend compiled by other teams against other opponents in other situations often in other seasons.

Although one might wonder why barely-if-at-all-observable "momentum" is placed in the same sentence as weather, matchups, fatigue, and play calling, since all these things are hardly abstract concepts.

I'm not picking on anybody. I talk about momentum. You talk about momentum. We all talk about momentum. I can recall using the word "momentum" as it relates to football just a few days ago. I was using the vague concept to try and justify my two-point conversion strategy; I was saying that teams have "momentum" after scoring a touchdown and thus their odds of successfully converting were better. This is of course absurd, since the data already accounted for "momentum." I am stupid.

The main problem with momentum is that, because we can't quantify it, it means nothing and everything at the same time. Did Mike Shanahan beat the Chargers because he had "momentum"? Sure, why not. Huh, my fever abated. Must've been that eye of newt and toe of frog brew I drank. This is how Lisa Simpson keeps tigers away with a rock.

If it can't be identified with any great accuracy, then what's the point of discussing it? Even if there is some mystical force called "momentum" affecting the outcome of games, we're as likely to notice its presence as miss the boat, so why even try and account for the results by guidance of what amounts to witch's brew?

Actually, it's worse than that. If the only problem with the concept of "momentum" was that it was nebulous in the sense that identification of it was nigh impossible, then fine, whatever, at least it gives those of us who write about sports for fun an opportunity to play make believe about the whys and hows of football. Good print and all that, etc.

But, er, it isn't merely muddily defined, in fact, it might actually be imaginary. Momentum doesn't exist. There is no such thing as momentum.

I was reading through University of California Herman Royer Professor of Political Economy David Romer's epically awesome paper: Do Firms Maximize? Evidence from Professional Football. His basic premise is that NFL coaches routinely depart from the rational decisions that would maximize their potential for winning by kicking field goals when they shouldn't or punting when they shouldn't. It is precisely the kind of statistically interesting and word-by-word justified (43 pages later I'm nodding my head and saying: he just covered it all) you won't find here, there, or anywhere in the NFL world. Where Len Pasquarelli is saying cmon you can just feel it touch it sense it so it must be there Romer is one of those party poopers who says: I will prove this for you.

And wouldn't you know it, in his analysis he takes the time to address a number of concerns, among them the concept of "MOMENTUM" as it relates to his strategy. As this analysis is absolutely momentous, I quote the short section in its near entirety (some emphasis added, just scroll down to the bold for the conclusion, if you're not interested in checking the data):

Momentum. Failing on fourth down could be costly to a team's chances of winning not just through its effect on possession and field position, but also through its effect on energy and emotions. As a result, it might be more costly for the other team to have the ball as a result of stopping a fourth-down attempt than for it to have the ball at the same place on the field in the course of a normal drive or as the result of a punt. In this case, the analysis would understate the cost of a failed fourth-down attempt.

There are two reasons to be skeptical of this possibility... Second, studies of momentum in other sports have found at most small momentum effects (see, for example, Gilovich, Vallone, and Tversky, 1985; Albright, 1993; and Klaassen and Magnus, 2001[*]).

More importantly, it is possible to obtain direct evidence about whether outcomes differe systematically from normal after plays whose outcomes are either very bad or very good. [Romer discusses sample size and his definitions of very bad and very good plays, click link to see for yourself, although data set is 636 very bad plays vs. 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. In each case, I ask whether the realized values of these situations one situation later differ systematically from the V's for those situations. That is, I look at hte means of the relevant [math sign I cannot reproduce here] (always computed from the perspective of the team that had the ball before the very bad or very good play).

The results provide no evidence of momentum effects. All the point estimates are small and highly insignificant; the largest t-statitistic (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.

(All typos from above are my own and not Professor Romer's.)

Long winded as that is, the point simply being: Momentum? Does not compute.

The issue with momentum is not that it exists and we can't find it, but that it doesn't exist at all! Whatever it is Len Pasquarelli, you, me, every single person who has spoken about football from a broadcast booth or television studio thinks we can just see, feel, know or gleen, it ain't momentum (I have used the term in 15 different storiess at Hogs Haven; impressively, readers have only fallen prey to its allure in one fanpost).

I'm putting momentum to bed and promise to try my bestest not to use the term in the future. In fairness to the coaches Romer is roasting in the above paper -- the ones he thinks should go for it more often -- I think it's only fair that they get an opportunity for response. A while back Greg Garber did just that at ESPN. Here's what they thought of this stupid paper, evidence, science what's that, etc.:

"This is a professor from Cal-Berzerkely?" asked Giants head coach Jim Fassel, in the true tradition of a former Stanford man.

Fassel turned a sheet with the equation on it sideways, then upside down in a humorous attempt to absorb its subtleties.

"What does the professor coach?" Fassel asked. "Maybe," he added, "he needs a few more classes to teach. Too much free time?"

"The crowd is going 'Go for it,' and they're just drinking beers and just going for it," Mariucci said. "Sometimes you get swayed a little bit. So you've got to block them out and you've got to make sense of it all.

"So then you start thinking about that article the guy from Cal wrote and then you say, 'Well, what would he do in this situation?' "

Hey this sounds familiar, emphasis added:

"If I don't get the first down, what are the repercussions?" asked Packers head coach Mike Sherman. "Are they moving the football? If you're on the road and don't get that fourth down the momentum is going to change over to the other team."

Momentum, according to coaches (see sidebar), is a matter of some consequence in fourth-down situations. But Romer -- a man as serious as an economist can be -- doesnt' pretend to offer an infallible system, just a guideline. That's why he couches his conclusions with the words "on average."

(Which raises the obvious question: Did any of the people interviewed read the entire paper? There it is, in big bold letters, an entire section on the Great Myth of Momentum under Section IV Complications -- great thing about academics is they have to actually address contrary opinions as opposed to merely stating them.)

Brian Billick, the cerebral head coach of the Baltimore Ravens, isn't so sure about all of the professor's numbers.

"There are only two numbers," Billick said. "And those are 50-50. You either make it, or you don't."

And at least one guy who appreciates the good Professor's work:

Bill Walsh, another Bay Area professor with some serious tenure, is a believer.

"To this day, I can close my eyes and see 22 players," he said in his San Francisco 49ers office, where he still serves the team as a consultant. "I can see them moving and I can see the equation. My indicators would be somewhat different than (Romer's), but this (equation) is what it takes."

Huh. Jim Fassel. Steve Mariucci. Mike Sherman. Brian Billick. Bill Walsh. Which one of these is not like the others?

* If you're interested in these papers the cites are:

Gilovich, Thomas, Robert Vallone, and Amos Tversky. 1985. "The Hot Hand in Basketball: On the Misperception of Random Sequences." Cognitive Science 17 (July): 295-314.

Albright, S. Christian. 1993. "A Statistical Analysis of Hitting Streaks in Baseball." Journal of the American Statistical Association 88 (December): 1175-1183.

Klaassen, Franc J. G. M., and Jan R. Magnus. 2001. "Are Points in Tennis Independent and Identically Distributed? Evidence from a Dynamic Binary Panel Data Model." Journal of the American Statistical Association 96 (June): 500-509.

So much for momentum.