It is absolutely correct that hard data is the only way to objectively prove any kind of flaw in random number generators. This is often stated in response to those who claim it is flawed or out to get them personally, but the problem is there is no framework supported way to extract a statistically significant number of anonymous replay logs.sourisdudesert wrote: ↑23 November 2020, 19:08 If you have the mathematical proof (ie: probability calculation vs important number of random event) that the BGA random generator is flawed, this is useless to vote "yes": please post this proof here instead.
I hope there is at some point, because I think it would be really valuable to study design and gameplay issues that aren't covered by existing statistics (and generally don't need to be), such as first player advantage or whether certain cards are overpowered.
Personally, I'm not interested in debunking claims about the random engine, because those who are convinced it is biased usually won't believe any proof to the contrary and will always move the goalposts (a few entertaining threads come to mind).
