Quote:
Originally Posted by Barm McLir
[You must be logged in to view images. Log in or Register.]
I read about a teacher in a statistics class that assigned her students the task of either A) flipping 100 coins and writing down the results or B) making up the results. Their choice. She could always spot the fake results because the students couldn't conceive of seven or eight heads in a row.
|
She'd also be able to spot the bad pseudo rng (or something else going on) on p99 when 1 in 1000 streaks happen several times more often than they should literally constantly.
I actually got as far as reading about goodness of fit tests anddownloading R the other day, but I didn't get around to actually figuring it out yet and running tests on p99 /random to see just how random it really is. I'll get there sooner or later though.