Tuesday, 14 October 2008

Taleb talks

I went to see Nassim Nicholas Taleb talk at LSE last night. Interesting bloke, some of the presentation was over my head (there were, like, graphs and equations and stuff) but the message was generally the same as usual - mathematical models for measuring risk are flawed, we know a lot less than we think we do, one big event can completely reverse a previous trend.

To some extent I wish I had understood some of the more techie bits better, as there is obviously some very intersting stuff in there. He made the point that in many markets a single observation can account for a huge chunk of the return. I'd really like to get my head around more of that kind of stuff, but the presentation assumed a bit more knowledge than I (and probably quite a few of the several hundred other people there) had. Also it was a bit of shame that he didn't really go into the philosophical side of his ideas more, which I find really interesting.

The Q&A was great though (unless you are involved econometrics...) as he came across very much like the iconoclastic person you imagine from his books. He is extremely sceptical about the abilities of the current and former chairs of the Fed, and was scathing of the merits of economics as a whole.

PS. There's an S&M post on Taleb today too.

2 comments:

Nick Drew said...

saw Taleb on the box for the first time the other day, and it seems he's as annoying in the flesh as he is in print !

he does a lot of gratuitous 'blinding by science' and he is entirely wrong to say that Value-at-Risk should be outlawed: in the correct hands it's a useful tool, albeit that several other tools are needed in the armoury

equally, though, everyone should hoist onboard his Extremistan concept

Tom P said...

I quite like his style (I know I'm in a monority here judging by reviews of his books). But I'm out of my depth when it comes to whether he is actually on the money on the techie stuff.