I just came across this old post by Bill Barnwell at Grantland about momentum, or as he calls it, nomentum, because he believes there’s no such thing. As is the case with the hot hand theory, which I wrote about before, there’s a striking disconnect between players, commentators, and fans, who seem to put a lot of stock in momentum, and stats guys, who flatly deny its existence.
I wonder if further research into momentum might follow a similar pattern as the hot hand theory: initial popular assumption that it’s real and important -> rudimentary research uncovers no evidence of the effect -> more nuanced research suggests that the effect is real, but less drastic than commonly believed, and offset by compensatory effects.
Part of the difficulty is that it’s not at all clear what momentum is. The word is used to describe a variety of different phenomena. In order to wrap my mind around momentum, I tried to nail down a definition. At its most basic level, momentum is the idea that what happened before is going to keep happening. But isn’t that just evidence-based reasoning? We all use the past to try to predict the future.
What makes momentum special is the extraordinarily high weight it puts on recent events. Again, this makes sense, up to a point. In trying to predict the result of the Super Bowl this weekend, we should clearly put more weight on the Patriot’s victory in the Conference Championship two weeks ago than on the team’s performance in 1974, which featured no players or coaches currently on the team. But how much more should we value the Conference Championship than the Patriots’ performance in week one of the regular season?
It seems to me that momentum isn’t really a unique idea, but a form of evidence-based reasoning that puts an especially high value on recent events. Therefore, a better question that “Is momentum real?” will be “How heavily should we weight recent events?” In most cases, I suspect momentum believers value recent events to a degree that can’t be justified.