I think part of the problem is that the solutions are not always "complete" in some sense. The process of solving problems can be just as important as receiving an answer. I don't think that this is enough to warrant downvotes personally, though I can understand why some people might not like the answers too much. In general, I think that computational answers are often well suited to questions, but I think that they would be better received when presented in a "more theoretical" way.
For example, when you use mathematica to calculate the pdf from the mgf, you do provide a solution, but nothing is made clearer to the asker, apart from the fact that they know that mathematica could tell them how to do it. If a problem can only be solved numerically, then there should be no problem, but a key part of the answer should be how to solve the problem numerically, not just what the answer is. If, however, there is a theoretical solution, using mathematica to find the solution doesn't really help the asker if they want to understand the process. From what I can see, mathematica goes through some sort of "thoeretical" process to arrive at the exact answer. If this is the case, I'm sure that a few comments explaining what the program actually does would be much appreciated, even if it's just something along the lines of "generally, one would like to take this approach, in this case it is very involved so I'll just give a basic outline of what the program does and the answer it provide. All of this said, I don't think that the answer should have been downvoted.
With regards to the second question, here the answer is even less "complete" in that you only calculate the probability for a finite subset of the natural numbers. Everything you say is valid, but it doesn't give a total answer to the question asked. I think it is still a nice answer to have though, and it is a valid one. I think that the comment by Did comes across as unnecessarily harsh, and I disagree with the point made, approaching some questions computationally can be a valuable tool. You did respond a little flippantly (though were relatively restrained), and I can't see the "snide remarks" in the edit history of the post, perhaps there was something you said that I can't see that offended Did?