On Wed, May 15, 2013 at 9:01 PM, Kevin O'Malley <kevmol...@gmail.com> wrote:
> No, you got it wrong again. To use your dice analogy, it is as if someone > went ahead and rolled the dice 6*14,720 times and they yielded 14,720 > hits. But along comes a skeptic who says that all of those hits were > misreads. The chance of those misreads is 1/3 (If you want to establish > that the chance is higher, then make the case for it -- but it has never > happened, ever before, in the history of science). So in order for all > those 14,720 hits to be errors, it would be (1/3)^14720, which is the > figure that puts you off by 5000 orders of magnitude. > > > No, man. You're doing it wrong. The chance you're calculating is if they made exactly 14720 experiments, and all of them hit. If they made 3*1470 experiments, and the chance of a misread is 1/3, then you would *expect* something close to 1470 hits.