Quote:
Originally Posted by whswhs
I prefer not to roll 84 dice; it changes the statistical pattern. If you roll 6d, the standard deviation is 4.18 (so two-thirds of the rolls are between 16.82 and 25.18). If you multiply by 14, the standard deviation is 58.12. But if you roll 84d, the standard deviation is only 15.65. So that central range is 235.88-352.12 in the first case, but 278.35-309.65 in the second. In other words, rolling more dice makes the outcome substantially less random.
|
As an aside, this is something that irks me a bit about GURPS, the fact that lower damages are more random while higher damages are less random (of course, this isn't really unique to GURPS). I would prefer some option of having it essentially always be the same amount of random, but most methods I come up with for that would be messy in play (with the exception of "replace
nd with 1dx
n," but the spread one sees out of 1d is arguably a bit
too random; something that always used a 3d roll would give the sort of spread I'd like, but then you've got to multiply or divide the result by something with fractions and deal with rounding, and most players - and GM's - would rather not deal with all that).