Friday, November 12, 2010

Linearity And Predictability

As you've probably noticed, if you read this blog by visiting the site rather than through an RSS feed, The Black Swan has been on FLG's current reading list for what seems like forever. Well, over at Buttonwood's blog, there's an interview with Nassim Taleb in which he makes a point that FLG has been arguing for a while now -- non-linear variables don't lend themselves to prediction.

FLG's point is more along the lines of most people project linearly, and this leads them to horrible and silly mistakes. Taleb goes even further, it seems, and asserts that one cannot predict in a non-linear domain. Consequently, in those cases, one must use either organic protection (redundant systems, for example) or rely upon tradition. This insight has led him to read, wait for it, classical authors.

FLG really needs to read his books.


Andrew Stevens said...

Taleb is overrated. He is currently being hailed as a prophet (and accepting those laurels) when his whole point used to be that there are no prophets. And, in fact, he predicted absolutely nothing and has no actual solutions to offer for how to guard against the risks we don't know about.

Yes, his hedge fund apparently did well in 2008 (though he is very cagey about exactly how well), but his previous fund had lost money in 2001, a perfect year for his whole "Black Swan" philosophy, and he wound it up in 2004-2005 after extremely modest results before trying again with his latest effort, which is apparently doing better, but somebody is always going to make a lot of money when the market is swinging wildly. I've seen no evidence that Taleb was good rather than lucky.

I'm just not sure what I'm supposed to take away from Taleb except that some people have too much confidence in their ability to predict the future and I didn't need Taleb to tell me that.

FLG said...

Well, there's something to be said for understanding how we commonly go wrong in predictions.

Also, I've read a few articles about him and his fund strategy. It's not something I'd follow. That's for sure.

Andrew Stevens said...

He doesn't seem to have much to say about how we commonly go wrong. His major thesis is that future events might include things that didn't happen in the past, which we all already knew. We use past experience as a proxy for probabilities of future events because what else are we going to use? As far as I know, Taleb has never given an answer.

Andrew Stevens said...

Mostly I object to Taleb because he loudly proclaims his brilliance and originality while saying things that had been thoroughly discussed over and over in the quantitative finance literature back when he was a toddler.

Caribbean Fool said...

Well, I'll defer to the other commenter's because I don't know anything about hedge fund management or care about what his returns were. (Shit, following the logic he sets out there wouldn't seem to be much point in investing anyway...)

I did find all of his books interesting more for his POV than for his specific conclusions. Not sure he is any more brilliant than anyone else but he does have some insight into the impact of non-linear projections even if the projections themselves are largely useless. Anyhow, if you do read them I'd be interested in your response.

Not sure if you've already read it as it came out a few years ago, but 'The Omnivore's Dilemma' by Pollan is pretty good. Lots in there to think about. Take it easy duder.


Creative Commons License
This work is licensed under a Creative Commons Attribution-No Derivative Works 3.0 United States License.