Numbers and scepticism

Forget astrology and reading palms, tea leaves or minds. The closest we come to understanding the present and predicting the future is with statistics.

I was once a statistics sceptic, believing numbers could tell you something about simple linear relationships but, when things get complex, I used to think numbers were useless. Game theory and complexity theory have changed my view since then, but it took the publication of Freakonomics, the social science best seller by Levitt and Dubner, to convince me that linear stats, sensibly used, could be useful.

Freakonomics is still among best pieces of social science research I've seen. It's not just the clever application of stats that impresses; it's the imagination to use them in a wide variety of problems, like uncovering corruption in sumo wrestling and falling crime rates.

This article in American Scientist caught my attention precisely because it's critical of a subject I'm fond of. It's worth reading because the authors reveal mistakes gleaned from the Freakonomics formula, and renew the call to scepticism when numbers are concerned. Perhaps most importantly for investors, they remind us that even if we are devoted to a particular source, we shouldn't be shy about questioning it.