Most penalty kicks result in a goal; this is why soccer players go to such comical lengths to draw a penalty. The distance between the ball and the goal is so short that goaltenders don't have time to react; they have to commit to a strategy as the ball is being kicked. The usual strategy is to try and guess where the ball will be kicked and to jump in that direction. Surprisingly enough, the tactic of not jumping - that is, guessing that the ball will be kicked down the middle - is under-utilised.
An interesting study from a group of Israeli researchers (Bar-Eli et al, 2007) offers a plausible explanation: 'action bias'.
Goaltenders jumped in more than 90% penalty kicks in the sample: the frequency of staying put was only 6.3%. Kickers, on the other hand, distributed their targets in roughly equal proportions.
The goaltenders' strategy was not wholly ineffective: when the kicker aimed left or right (71.4% of the time), goaltenders guessed correctly 6 times out of 11. But the fact remains that the frequency of the goaltender staying put (6.3%) is much smaller that the frequency of kicks aimed down the middle (28.7%).
This under-reprentation is even more puzzling when you look at the save percentages. This is a reproduction of Table 2:
Unsurprisingly, when kickers aim left or right, goaltenders never stop the ball when they guess wrong before jumping. What may be surprising is that even when they jump in the correct direction, the save percentage is less than 30%. On the other hand, on the rare occasions in which goaltenders (correctly) stay put, the save percentage is 60%.
- The frequency with which goaltenders stay in the centre of the goal is much lower than the frequency with which kickers aim for the centre.
- Correctly staying in the centre has a much higher save percentage than jumping in the correct direction.
This doesn't mean that goaltenders should never jump. What it does mean is that goaltenders jump far too often. Why?
Bar-Eli et al suggest an explanation: 'action bias'. This is presented as an example of Kahneman and Miller's (1986) 'norm bias'. Goaltenders believe it is less bad to follow the 'norm' (i.e., to jump) and fail than to not jump and fail. In other words, goaltender think that jumping and missing is less costly than not jumping and missing.
Which brings us to economic policy. Politicians are continually demanded to 'do something' about a kaleidoscopic array of grievances, and the norm in these cases is to promise to do something. As far as politicians and most voters are concerned, doing something is better than doing nothing - even when doing nothing is the correct response.
In many cases - possibly the majority of cases - doing nothing is the smart move. A recent example is the concern about the so-called 'skill shortage'. When firms complain that they can't get the workers they want at the wages they are willing to pay, the correct response is to do nothing: the market response to a labour shortage is to let wages increase.
But doing nothing is almost always bad politics: it is invariably interpreted as a lack of concern, and this percieved indifference will be pounced upon by other political parties. A politician who promises to act polls better than one who promises to do nothing.
Goaltenders jump because jumping is expected of them: they fear to look silly by not jumping, even if not jumping is the smart move.
Politicians promise action because action is expected of them: they fear to look uncaring by not acting, even if inaction is the smart move.
Note: I started writing this *before* Nick's post about the macroeconomics of 'doing nothing'. Really!