Fezz, serious question. As a great gambler but also an expert statistician, I was wondering at what point you would stop calling something variance and decide that what someone is doing is faulty? 20 trials, 50 trials, 100 trials or does it also depend on the result.
For example 30-70 after a 100 trials (30%) or 4-16 after 20 (20%) or something?