roadrunner said:
a) order does NOT matter which is why i don't understand it lol...if order mattered its 31!/20!
(whats B-R)?
B-R is Baskin-Robbins, home of the 31 flavors. I asked about order because I wasn't sure about the context of the problem. So order doesn't matter because all twelve are being made at once. So the answer should be 31C12, since it is essentially 31 choices taken from a bin without replacement ( 31·30·...·20 = 31!/(31-12)! = 31P12 ), divided by 12! possible arrangements of the cones that we are counting as only a single possibility.
b) your not impoying 31^12 are you? (its in the chapter on combinations with repettitions...and 31^12 woold be if order matters)
Now that I know that order doesn't matter, I guess this gets modified a little. What I was suggesting, though, is that if
any amount of repetition is allowed, it doesn't really matter what choices you make for the twelve. So we have 31 choices taken from the bin
with replacement. That would be 31^12, but, as you say, we don't want to count the different orders of removal as distinct, so we would again divide by 12! (Even if all twelve cones got the same flavor, we want to count it as one outcome, but there are 12! arrangements in which individual scoops could end up in cones. I'm still thinking about this, though...)
This is a much larger number than for part (a), which makes sense.
The last one's a bit more complicated, since we have to make sure we don't allow more than 11 duplicates. I think we can start with the result for part (b), as you proposed, and subtract outcomes. But I think there are only 31 cases to be removed: those where all 12 cones have the same flavor. You don't have to deal with the cases of 11 identical cones and what to do with the twelfth, because those already exist in the set we've counted in part (b). So I believe (at the moment, anyway) that this would be
[(31^12)/12!] - 31.
I'll think about this some more, but I don't think the nCr combinatorial function comes into the last two parts, since you have the equivalent of selection from an urn with replacement (have you looked at lottery problems?). But feel free to argue with this -- combinatorics is one of the harder parts of probability theory to evaluate clearly.