Board Thread:Game Discussion/@comment-13368701-20130905202323/@comment-11490363-20130907135141

Andrewds1021 wrote: I suppose, let me see if I get this. So, what you are suggesting is not too look at total pulls per player but total aggregate pulls? For example, as a community, did we collectively get 50/50 or not?

As for the cyberdex thing, I think that could still be checked by individual totalls. If the community comes out to some probability but each individual was heavily one mode or the other, than that could tell us that, despite initial chances, what you get depends of your first pull. Exactly! Although looking at individual players' results would also be useful since it would allow us to do streak analysis and make sure there are not a lot of people with 1 in a million odds (according to our hypothesis of 50/50) since that would be very unlikely, but several people with 1 in 1000 is fine.

Also, a true random variable can be tested aggregately per person and use the cumulative formula of multiplying the odds of each component event, so a streak of 4 is 1 in 32. But we know that computers use pseudo random number generators and so that part of the analysis is not always as strong, but it might be strong enough, who knows.

As for the cyberdex thing, yes that would work, but the number of samples we would need to be relatively confident of the result is probably more than we would have. Whereas for just aggregate checking we probably have more than enough.