Quote:
Originally Posted by Breaken
[You must be logged in to view images. Log in or Register.]
So you are considering the king to be the variable. I took the drops to be the variable. The drops are not iid, and if you are calculating the probability of the drops, you should consider the drops as the variable. Either way though, the CLT specifically mentions, a 'large' random sample (large number of kings) of random variables (drops).
|
I don't quite understand what you mean by the drops vs the king as the choice of the random variable. The multinomial distribution seems very natural to me as it mimics how the loot table works. Are you suggesting that it is better to model it with three separate boolean variables, i.e. P(Tunic), P(Robe), and P(Staff)? Even here, I don't see why each variable would not be IID wrt other instances of itself, and I would think CLT would apply just fine. In other words, the average number of tunics per king should approach the mean and have a gaussian distribution even though the event itself is a discrete 0/1.
P.S. Jeremy, where's my 1M pp