On Fri, Mar 30, 2012 at 11:23 AM, Russ Abbott <[email protected]> wrote:
> The naive strategy for predicting coin tosses is anti-reductionist in John > Kennison's terms. There is even a rationale. We "know" that in the long run > (given a fair coin) the number of heads will be approximately the same as > the number of tails. Therefore, when one count grows larger than the other, > predict that the other will occur to even things out. Many people think > that way. > > *-- Russ * > Just to make sure I'm up with the conversation: The point you're making is that someone following the "predict the least occurring value" strategy would indeed even things out but not because of their strategy, but because of the nature of binomial distributions, i.e. strings of H/T with the same number of each are more probable. But the impact on the individual is to falsely deduce that they had an effective strategy, rather than a good understanding of probability. And worse, there is no way to prove them wrong .. well maybe by simply choosing the "all heads" strategy and showing it to be equally effective. -- Owen
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
