lifestyle leader

Last December I wrote about Nassim Taleb's The Black Swan and proceeded to tell anyone who would listen to Read This Book. You will have to forgive my boyish enthusiasm. It was probably a classic case of getting really excited about that-which-you-have-read-most-recently. Happens all the time.

Since then, however, I've read a number of criticisms - concerning both Taleb and his ideas. None, however, was nearly as comprehensively devastating as this one by Eric Falkenstein. Although written some time back, it's new to me and in the interest of balance I feel obliged to link to it here. It's certainly worth a read if you've read Taleb or are at all interested in questions of probability and prediction. For instance:

"Taleb is consistently amusing because his criticisms of others apply so neatly to himself: he claims he is an empiricist yet supports his points with anecdotes. The Black Swan makes fun of ‘experts’ with credentials, but he states he does not deign to engage with anyone not sufficiently expert; he states he is not interested in being a speaker-bureau commodity , but routinely travels the rubber chicken circuit; he derides forecasters who don't give a full accounting of their prior forecasting history, yet delinks old remarks about Value-at-Risk, and recategorized his extinct Hedge Fund as a hedge, not a fund; he claims to prize humility, yet is most immodest; he argues against applying the law of large numbers, and also of inferring too much from small samples; people apply models to reality in biased manner, people naively extrapolate data without the appropriate theory; forward thinking is adaptive, forward thinking is error-laiden. Some people think inconsistency is a sign of genius; I think it just reflects confused thinking."

"Black Swan argues that standard statistics is flawed because it is backward looking — it uses ‘historical’ data — and argues that standard measures of risk like the normal distribution are ‘frauds’. I too prefer future data, but it is hardly a practical alternative. The Gaussian distribution is common in theory because it is so analytically tractable; it often creates closed form solutions that allow one to see how one variable affects another, and has nice properties, such as the fact that two Gaussian random variables added together is also a Gaussion random variable. In practice, no one actually believes in this view, and makes ad hoc adjustments.... Non-economists often giggle at the term ‘fat-tailed’ or homoskedasticity, but indeed most real world distributions are not ‘Normal’ or Gaussian, they simply have fatter tails than average. Does this imply statistics is a fraud? Well, if you mistake the map for the territory, indeed, this is news."