Monday, March 28, 2011

Foxes vs. Hedgehogs

Dan Gardner's Future Babble:
Tetlock dubbed his experts “foxes” and “hedgehogs.” Foxes beat hedgehogs. Tetlock’s data couldn’t be more clear. On both calibration and discrimination, complex and cautious thinking trounced simple and confident. By cross-checking other factors in the data, Tetlock also found that hedgehogs who are ideologically extreme are even worse forecasters than others of their kind. He even found that when hedgehogs made predictions involving their particular specialty, their accuracy declined. And it got worse still when the prediction was for the long term. Put all that together and there’s a very clear lesson: If you hear a hedgehog make a long-term prediction, it is almost certainly wrong. Treat it with great skepticism. That may seem like obscure advice but take a look at the television panels, magazines, books, newspapers, and blogs where predictions flourish. The sort of expert typically found there is the sort who is confident, clear, and dramatic. The sort who delivers quality sound bites and compelling stories. The sort who doesn’t bother with complications, caveats, and uncertainties. The sort who has One Big Idea. Yes, the sort of expert typically found in the media is precisely the sort of expert who is most likely to be wrong. This explains one of the most startling findings to emerge from Philip Tetlock’s data: The bigger the media profile of an expert, the less accurate his predictions are.

No comments:

Post a Comment