Bayesian Logic

[garyoz]

Interesting article in the Economist about using Bayesian Logic in decision making when using limited data. The article is also critical of the "frequentist" approach of relying upon large samples and assuming normal distributions. Also suggests a basis for correlation/causation confusion.

http://www.economist.com/science/displaystory.cfm?story_id=5354696

[Overlay]

Whenever I hear the term "Bayesian", I think about the joint research project that a fellow horseplayer and I collaborated on in 1984 as the final requirement for a master's degree in systems management. We did an analysis of betting patterns (morning-line odds versus post-time odds) on two years' worth of harness races (the only game in town) at Quad City Downs in East Moline, Illinois, to see if we could uncover any trends that might lead to consistent profitability. (The fact that I've played the thoroughbreds exclusively since then speaks to our degree of success -- or lack thereof.) Anyway, my friend did most of the number-crunching, and I wrote the draft of the text, which he then reviewed and made suggested changes to. One of his additions that has particularly stuck in my memory for some reason was his characterization of the public's odds as "Bayesian probabilities by popular acclaim". I always wanted to work that phrase into a conversation sometime, and, by golly, now I have!

[highnote]

Interesting article in the Economist about using Bayesian Logic in decision making when using limited data. The article is also critical of the "frequentist" approach of relying upon large samples and assuming normal distributions. Also suggests a basis for correlation/causation confusion.

http://www.economist.com/science/displaystory.cfm?story_id=5354696

That was interesting. Thanks.

[traynor]

From Economist article: <How the priors are themselves constructed in the mind has yet to be investigated in detail. Obviously they are learned by experience, but the exact process is not properly understood. Indeed, some people suspect that the parsimony of Bayesian reasoning leads occasionally to it going spectacularly awry, with whatever process it is that forms the priors getting further and further off-track rather than converging on the correct distribution.>

I thought I picked the wrong horse; my losing tickets were in fact the result of the parsimony of Bayesian reasoning going spectacularly awry. That makes me feel a lot better.

"Priors" are an interesting topic. Any thoughts, comments, opinions in that area?
Good Luck