mookiemcgee wrote:Thanks Duku, that type of post is helpful and highly relevant to the crux of my post.
Perhaps this question will go unanswered because you just aren't in a position to confirm or deny... but is it possible we are still using a 50k file, just one that was updated to get to an average dice roll of 3.50? I realize you are acting more as a historian and examining the timeline, rather than disseminating information for BigWham. I'd just be curious to know if that possibility was disproven or looked at by others.
What you posit is exactly what degaston asserts now, but he hasn't collected any new data since his big magnum opus in 2013. He believes that a new 50K file is being used in place of the old one, just one that doesn't have the flaws of the previous one. IcePack says that's not the case, and he's kind of the special apostle whom our webmaster trusts more than the rest of us, so his information is probably good.
I can't say with any certainty, and certainly we have nobody now willing to put in the effort that degaston used in 2013. I will say this, however: if a similar method is being used, there's actually nothing wrong with that. In the good old days before computers, when statisticians needed a set of random numbers, one common method was simply to fan open the phone book to a random page and take the last digit of everybody's phone number on that page. Then along came computers and computer-generated "random" number algorithms, which were quicker to use. The interesting part, however, is that the algorithm-produced random numbers were actually an inferior product. IOW, they produced numbers which did not stand the tests of randomness to the same degree as the phone book did.
Now, depending on where you live, there's probably more than 50K numbers in your phone book, but other than size the phone book basically is a fixed file with a large but finite number of random numbers. It really is the same methodology. So if the 50K file has been replaced with something similar like a 500K file or whatever, that's actually a pretty good method. The only reason the 50K file was ever exposed was because it was an extraordinarily bad 50K file with the now-infamous shortage of 1s. Lackattack just had extraordinarily bad luck in the file he chose.