I am a biologist trying to quantify the likelihood of observed genetic phenomena… I’ve tried my best to formalize my question in accepted terms as follows:
Given a stack of n cards, each colored red (x cards) , blue (y cards), or green (z cards) where x + y + z = n, what is the expected number of draws, without replacement, before you draw both a red and a blue card? It doesn’t matter whether the red or blue card comes first. Preferably, a solution would be defined in general terms, but for demonstration’s sake, one could assume a stack of 10 cards, where x = 3, y = 5, z = 2.
Intuitively, I figured this could be solved using multivariate hypergeometric distributions. However, because we are looking for the expected minimal number of draws before two successes occur (both a red and a blue card), this approach appeared to require some tricky combinatorics.
I have also looked, without success, into adopting solutions from other problems. My sense is that between the following two posts there is a solution, but one I have not been able to intuit.
Expected number of card draws to get all 4 suits
Expected number of trials before I get one of each type
Any direction would be greatly appreciated -- thanks so much!
This may help. Clearly the maximum number of draws is $z+x+1$ assuming with no loss of generality that $y\leq x.$ Then apply the inclusion exclusion approach in the first answer you linked.