Friday, February 29, 2008

Going through the bins

I'm currently reading a brilliant book called "The Black Swan" by Nassim Taleb where he exposes just how poor humans can be at forecasting. Taleb was a quant on Wall Street so he should be well aware of the power or otherwise of forecasting.

One interesting observation I took from the book was just how poorly we forecast because we limit the set of information we consider before drawing conclusions. So how do the guys in the city make their money?

One explanation came my way the other day. The "top" guys look for the big unexpected (by others) events. How do they do this? Well according to my source, one method is to hire private detectives to go through the rubbish of directors from the firms and industries that they track. Why bother trying to forecast share prices with dubious assumptions on productivity gains etc when you can just go through the rubbish and find out whether the company has a massive law suit creeping up on them. Genius forecasting if you ask me!

1 comment:

rjs said...

I read some of it and really liked it too, then read a review of it in the Royal Statistical Society mag ("significance") that said its proposition was nonsense because the standard pdf does take account of outliers too. So haven't finished it but will take it on holiday next time.