Mathematical metaphysics of randomness (Q1275005): Difference between revisions

This is a very long paper which tries to give a meaning to many loose notions (two instances are ``randomness'' and ``impredictability'') for infinite sequences of trials. The reviewer feels that some arguments and statements are not very convincing, based as they are on the well-known (and often challenged) von Mises' conceptions. In general, these kinds of researches look like an effort of giving a sort -- in a sense -- of ``quasi-deterministic'' interpretation (based on ``computational'' ideas) of some probabilistic results. Let us just mention as the authors state the law of large numbers: ``the ratio \(n1/n\), where \(n1\) is the number of 1's in the sequences' initial segment of length \(n\), tends to 1/2 as \(n\) tends to infinity''. Another issue that makes clear the difficulty of treating probabilistic phenomena in the framework of computer science (such as the Kolmogorovian idea of complexity) is the statement, by the authors, of the following ``principle'': ``no event of probability 0 can happen''. In the final philosophical supplement the authors try also to discuss a notion of ``surprising'' sequence, which should have (as they claim) direct connection with basic problems of mathematical statistics.