Saturday, May 22, 2010
Is there such a thing as Conditional Probability?
Andrew Gelman says "no". I have been using conditional probability for ever and even teaching it. Now this individual alleges that taking a measurement mixes up everything, that there is some kind of uncertainty principle (Heisenfeld? Eisenburg? Eygenfuck? I am not certain), so conditional probability is false.
In another place he proposes that a large number of measurements do not arrange themselves in a gaussian distribution, so a probability calculation operates as an attractor that destroys information, ergo, the results are false.
Should we be in the Soviet Union in 1938 and I was Stalin in my current mood, he would find himself mining coal in Vorkuta (pic). People who puts in doubt the established believes of an old man should be punished. Next somebody will write that the vodka I am drinking is no good for the circulation.