Whether or not there is a second wheel is immaterial in probabalistic terms. If it is assumed that the probability of getting a rare drop on the first wheel is p(rare) and the probability of getting a certain rare drop, e.g. d left half, on the second wheel is given as p(2ndWheel) then the overall probability of getting a d left half on any kill is a fixed value of: p(d left half) = p(rare)*p(2ndWheel) This probability can then be used in the binomial probability mass function to calculate the probability of achieving one drop in n kills. The observation that one d left is dropped per x kills is not unreasonable, the binomial distribution itself shows that the overall probability of a drop occurring in x kills increases as x increases, however the individual results will be independent. THat is to say the fact that the guy next to you got a d left does not in any way affect the likliehood of you getting one on your present or future kills. Thats just my opinion but there aren't that many ways of doing these things...