Thread:Jkalm/@comment-24577221-20150313023719/@comment-24577221-20150314040110

You may also be interested in Belthazar's essay on binomial theory and its application to the statistics of Ghazt breeding.

Your "model" predicts that with each successive failure, the chance of success on the next try increases. I am rather sure that this is not the case, though of course I don't have access to the algorithms used by the MSM game. It isn't provable from the data, of course, because all the data can show is that after some string of failures, there is eventually a success. But for the sake of argument, you're free to flip coins a bunch of times and see if there is a pattern that matches your prediction that the result of one coin toss depends on the results of the preceding tries.

The experimental data match my formulae, in part because one of the crucial parameters of the formulae (the chance of success on each try) is derived from the experimental data.