I’m not an expert in how engines work but can make the following observations. I have the following engines: Rybka, Fritz, Shredder and Chess Tiger (all but Fritz I got free). The only one I use is Rybka but I prefer the way Fritz prints out the games and I run all the engines under Fritz. Fritz and Shredder often offer up the same move selection but vary slightly on the evaluation. Rybka often offers a different move, but all three usually do not vary much in their evaluation of the position. Chess Tiger on the other hand often suggests a completely different move and for some reason its evaluation of a given position is often much higher than the others. I do like the availability of Shredder’s online 6-piece endgame database though. (You do not need an engine to access it)
I have noticed while analyzing games of even very strong players I can usually make sense of their moves. Sometimes the engine’s suggested moves baffle me. One caveat with any engine…you have to let it run long enough to correctly evaluate the position. Analyzing a recently completed postal game (where I had a significant advantage) after my opponent’s move, I was shocked to see he had a mate in 5. It took Rybka about several seconds to realize there was no mate and my opponent was still lost. I once had a postal opponent make a piece sac that had me baffled as to why he’d play it. Later analyzing with Fritz I noticed it liked the sac…for about 15 seconds then found the obvious defense. Made me suspicious of what he was doing. When analyzing with a computer it will point out tactics, but won’t help you much with positional understanding…you have to work that out for yourself. I guess in the final analysis it doesn’t matter much which engine you use at least for us rating challenged players it doesn’t.