Sean McHale says Ed Miliband should learn from Barcelona. In doing so, he draws attention to why politics is so flawed a discipline.

Sean says that, in the Real Madrid game, Victor Valdes continued to play short balls out of goal despite the fact that doing so had gifted Madrid a goal early doors. He advises Miliband to similarly stick to one track rather than fumble around in the dark.

This misses the huge difference between Valdes and Miliband*. Valdes could draw on vast experience which shows that playing it short is, for Barcelona, a successful strategy. But Miliband has no such large evidence base to draw upon. And this makes his job far tougher. Valdes could rationally interpret a failure of a single short ball as an exception to a generally good strategy, because he had played countless such balls successfully before. He therefore had a strong Bayesian prior that short balls work, on average. But Miliband, lacking such relevant precedents, can have no such strong priors. If he sees a policy statement fail to win support, he cannot infer, Valdes-style, that this is a rare exception to a successful overall strategy. Instead, a rational Bayesian would be more likely to infer that he is on the wrong track.

In this sense, Sean's call for Miliband to stick to his instincts might be mistaken. What caused Valdes to continue playing it short was not instinct, but rather a rational judgment based upon experience. Miliband, however, is not in so happy a position.

Miliband is not unique in this. Politicians very often find themselves in positions where they have few precedents from which to form judgments - a problem exacerbated if they are ignorant of history or of other countries' experiences. It will, therefore, be far harder to find the right strategy, and even if they do an early failure might cause them to wrongly but rationally change course.

In this sense, politicians are more like entrepreneurs than managers. A defining feature of the entrepreneur - as distinct from the manager - is that, as Israel Kirzner said, he is a person who must act on the basis of limited precedent and knowledge. Such actions are liable to fail.

But there's a massive difference between entrepreneurship and politics. We have found a way of improving the odds of entrepreneurial success. We allow many entrepreneurs to compete against each other, to see who succeeds; this is why free entry into markets and access to capital are so important. In politics, however, competition is much more limited and entry restricted. So the natural selection we have in markets operates much less well.

In this sense, political activity offers us the worst of both worlds. It has neither the body of experience, evidence base and precedent that sportsmen, engineers, bureaucrats, lawyers or some artists can draw upon. Nor does it permit the ruthless natural selection that well-functioning markets do. It is, then, small wonder that, as Enoch Powell said, 'all political lives end in failure.'

* OK, there are other differences, not least of which is that Valdes is surrounded by geniuses. I'm confining myself to just one.