This week I read my copy of “The Economist” August 18th-24th, and I found some articles about the most recent financial turnmoil which I’d like to comment, since there’s some info related to something I wrote here and here.
The Game is up (pp.63-64)
“Quantitative funds have been hardest hit […]. Goldman Sachs admitted as much when it said that its funds had been hit by moves that its models suggested were 25 standard deviations away from normal. […] That is silly.”
Of course is silly!, however, ex-post the probability of this whole thing happening is 100% (as the article explains, probability 1 is certainty).
Behind the veil (pp.65-66)
Talking about hedge funds:
“Insiders say losses were caused by a mixture of human and computer-driven error.”
How’s the split???. Was the error to use them models? (as a math believer, I want to believe NOT).
“…the complex models that drive them were up ended by the extreme market volatility. […] the models went awry.”
Of course, they already said 25-sigma!!!.
“The saga has damaged the image of computer-driven funds.”
It’s a pity this could damage the image of whole OR as well. But that’s not the case…
“People aren’t going to give up their computers and go back to insider information and tips.”
And, I’d like to add up, at the end of the day the models were right, at least ex-ante ; ) .
“…highly unusual gyrations were ‘way out of whack’ with their models. Clearly the models were flawed.”
Hopefully, in another section of the magazine we can find an article about statistics and climatology; nothing seemingly related to these events but for a good insight about the difference between Bayesian statistics and Pascal statistical concept. Thought’d be interesting for understanding were the models could failed.
Remember the small print: “Historical performance does not assure future performance”. What a pitty!.