The Other Recency Effect

Another day at the Guardian football pages, another transfer rumour about Roque Santa Cruz. He seems to be a pretty hot commodity. I appreciate that he may have had a good last season at Blackburn, but every time I read that kind of thing, I think: "People, haven't you followed his career at Bayern Munich? He was awful!"

Sports fans tend to think that the recent performance of an athlete or team is a particularly good predictor of the next performance. Words used in this context are streak, flow or the hot hand. (And look, here's even a blog called The Hot Hand in Sports.) Theoretically, this makes sense. After all, a good performance should instil confidence in you, which in turn should improve your performance.

The thing is that research tends to find that recent performance is a bad predictor for future performance; taking into account performance over a longer period is much more predictive. And courtesy of The Sports Economist, along comes this:

An economist examined performance and salary data for all hitters in Major League Baseball who signed free-agent contracts between 1985 and 2004. He found that many teams systematically over-estimated the predictive value of a player's performance in the most recent season, relative to earlier seasons, in setting salary levels.
(The cite is: Healy, A., "Do Firms Have Short Memories? Evidence from Major League Baseball," Journal of Sports Economics (August 2008). I have not read that paper.)

Psychologists speak of the "recency effect", but that simply denotes the fact that individuals are particularly good at recalling recent information. Naively, one might think that this explains the excitement about Roque Santa Cruz, to stick with the example. But I think prospective clubs have ways of assessing his performance with Bayern. I think the problem is that persons in charge consciously and systematically overweigh last season's performance relative to other ones.

Assuming all of this is correct, three points:

1. Do we see this bias in other walks of life, too? For example, if your partner is generally happy-go-lucky, but is in a bad mood today, will you overestimate the likelihood that he or she will be in a bad mood tomorrow? Something to ponder.

2. Traditionally, economists have operated with a model of hyperrational humans. Then psychologists, and later behavioural economists, came along and pointed out many ways in which this model was wrong. The comeback from more traditional economists was: "We-hell, you may be able to find these anomalies in lab experiments, but when the stakes are high, people are going to act more rationally." This is true to some extent. But in professional sports, the stakes are certainly very high, both in monetary and nonmonetary terms, yet people appear to act irrationaly.

3. Relatedly, an old, quasi-evolutionary argument in economics is that firms may behave stupidly, but these firms soon disappear from the market. Indeed, one should expect that successful sports clubs have developed mechanisms which keep individual irrationality in check, but the study cited above and my anecdotal evidence from football suggest otherwise.

1 comment:

J Thomas said...

Maybe it's better for the game when people get excited about recent wins.

For one thing, a player who's developed an average track record can expect nothing better. But if people ignore the averages, there's a chance he'll have a streak of good luck that gets him a lot of attention and also maybe a sum of money. He might get his 15 minutes of fame.

Spectators pay for excitement in sports. A spectator who makes bets and loses money gets the excitement of the betting, which probably matters more to him than the money. When he loses on average it doesn't mean he has to stop betting, it only limits the size of his bets. And if he won he probably wouldn't use his winnings to bet higher stakes, he'd buy something he enjoyed, something that reminded him of his victory. Maybe sports memorabilia.

Evolutionary arguments don't necessarily work in this environment, just as they don't in gambling. If losing gamblers were consistently shot then quickly nobody would bet except a small minority who thought they couldn't lose. But when all they have to do is go home and make more money so they can bet some more, they persist.