Reply to thread

Message: <blockquote data-quote="Gorgon Zee" data-source="post: 7843621" data-attributes="member: 75787">So, a bit of statistical practice needed here. Statistical odds are calculated assuming that you declare an experiment, run that experiment, and report it. However when people report odd dice rolls, they do not do that, they are reporting an unusual chain after it has been done, so effectively they are picking the "most unusual" of a set of dice results.This is not the easiest sort of thing to work out, so I'm going to make the slightly inaccurate simplification of saying that each chain of rolls is independent (so the odds of rolling 6 20s for the first roll of the night is independent of that for the second). This is obviously wrong, but won't affect the end probabilities by more than 10%, so I'm going for it.d20s are rarely fair; so 3 columns in the table below for slightly bad and very bad dice. As Umbran notes, with fair dice the odds of either 6 20s or 61 in a row is about the same as winning the Texas lottery. Doing the math on rolls / decade, that would give < 0.1% chance of seeing it happen in a decade.However with a badly unfair dice, it'll happen every second decade, on average. [ATTACH=full]115328[/ATTACH]</blockquote>