pipsqueeek wrote:
Siniri wrote:
Jas O'Lantern wrote:
I wonder how much variety this will give.
The standard deviation for a 10-sided die is 2.872281323.
Variance=8.25.
Range=[1,10].
I only bothered because I still had the SD calculated from my earlier experiment.../procrastination
The distribution wouldn't be normal though(we hope), so standard deviation wouldn't really be relevant...
Unless you mean for multiple dice. But then it'd be different for different numbers of dice.
Right. I didn't use the normal distribution to calculate the SD.
Every distribution has a variance, and thus a SD (which is the square root of the variance). The variance does change with the number of sides on the die, of course. During my experiment, I relied on the Central Limit Theorem given my large number of trials (and thus I used z-tests to determine the p-values). But the SD's I used for each die were not based on the normal distribution. Admittedly, I looked the formula up on-line instead of referring to my old textbooks, but it gelled with the one in my memory -- and it gave the proper variance for a two-sided die, which is just a binomial distribution. FYI, for a discrete random variable, the variance is the sum of {(probability of outcome i) times [(<value of outcome i> minus mean) squared]}. So for a 2-sided die, it's 1/2 * [(1-1.5)^2 + (2-1.5)^2] (because the probability for each outcome is the same for a fair die, it can be pulled outside the summation) = 0.25.
Jas wondered about the variety that would be given by her roll -- and variance is a measure of how spread out a distribution is; I thought this might be what she meant by variety.