Over my time here in Wikipedia, as an editor and then an admin, I think I have finally crystallized my thoughts on RFAs and RFBs. In a nutshell, a request for adminship/bureaucratship is a request for the community to comment on the judgment and trustworthiness of the individual under discussion. Do we feel that the candidate will uphold wiki policy and guidelines where they exist, and will make reasoned decisions where policy and guideline is vague?
Wikipedia seems to have a tendency to regress to number counting and cold statistical analysis when faced with difficult decisions. See Wikipedia talk:Requests for adminship/Archive 79 for a prime example, although there are many others. Personally, I think that is an error. Anything that can be relegated to nothing more than bean-counting can, and should, be handed off to bots. We have admins and bureaucrats in Wikipedia because there are times we need a human being to make a decision. Usually, those times are when the decision will be unpopular. When we opine on RfAs, RfBs, and ArbCom elections, we are actually saying "we trust so-and-so's judgment, and are willing to allow him or her to make the difficult decisions."
Of course, egregious behavior needs to be addressed, which is why we have WP:ANI and WP:RfAR. Further, it is always a good idea for the admin or b'crat in question to discuss difficult issues beforehand, but that is not a requirement. Exercising judgment is what we ask our admins and b'crats to do, and if that judgment is exercised in a non-capricious, impartial, and reasoned fashion, we cannot fault them for that.
Ask yourself before you opine on a RFA/B/etc., "do I trust this person's judgment?" Then if you do, while you may disagree with it, and post vociferously why so when that occurs, remember that is why you placed your opinion on their nomination. And if you don't, then you should have acted on that during the nomination process.