There was a neat link in TrueHoop’s bullets today about the benefits of being clutch. I will now rant about clutch.
I guess before I start ranting, you should read the article by Ian Levy. The definition of clutch time is straightforward: it’s the stuff that happens when teams are within 5 points with five minutes or less to go in the game. The idea is that when a game is close, players have to rise to the occasion and lead their team to victory. The idea that clutch exists is one of those things that stats guys love to talk about: most fans believe in it strongly whereas the numbers are much more murky on the subject. I, for one, don’t believe in it. Why? Because it isn’t a repeatable action. It isn’t something that players can just do. If players could just decide to play better, they would do it more often, right? It just raises my hackles.
So Ian’s article examines the connection between how much a team overachieves in the win column, comparing actual wins to expected wins (from point differential), to how much they overachieve in crunch time, comparing their clutch rating to their overall efficiency. And so far this year, as well as last year, teams that overachieve in the clutch (play better than they generally play) also tend to overachieve their point differential (win more games than they ‘should’). Ian concludes that clutch play is obviously valuable, and that there is value for teams to try and become 4th quarter juggernauts.
The article is pretty interesting; I don’t want to knock the whole thing because it’s a neat finding. But how in the world would a coach or GM attempt to make his team clutch? The most consistent finding about clutch play is that clutch play is not consistent. And you can actually see that in Ian’s data. If you take the clutch differential for each team in 2011 and correlate it with the differential for 2012, you get a correlation of .09. If you run a regression so you can try and predict each team’s clutchness, the equation squeezes everyone down to within 2.5 points of 0 even though in both years the actual swing is more like 20 points in either direction. In general, knowing who is clutch now tells you next to nothing about who will be clutch next year. A team can happen to play well when the game is close, and thus be clutch, but they can’t seem to do it on purpose.
So, does being clutch help? Sure it does. Can you actually do anything to ‘be clutch’? No. Or, to quote Brian Burke, “Although clutch performance may not be a persistent skill in players, there undoubtedly exists clutch play itself”. If a player or whole team happens to play well at the end of close games, they will undoubtedly win more than they should (see also, Tim Tebow). But there is no reason to then call them clutch or expect them to keep doing it in the future. If you want to win a lot, you should try being good.