After prowling forums dealing with chess engines, I have come up with a general consensus of the strengths and weaknesses of various engines which are summarized below.
As that great teacher C.J.S. Purdy used to emphasize, a position either contains a sound tactic or it doesn’t and contrary to what a lot of players seem to think, you can’t play (sound) “tactical” chess if the position doesn’t warrant it. And when it comes to tactics all engines will quickly spot them.
Another thing to keep in mind is analyzing games at different time limits will give different results. That’s why I don’t really trust online engine ratings too much because the rating lists are based on blitz games. Anyway, in regards to engine “styles” engines are so unlike human play that we often can’t understand their choices in some positions. Unless you are a titled correspondence player trying to find the absolute truth in a position what we are looking for is an evaluation of a given position at a reasonable depth and most all engines will accomplish that.
Please note that these are not my opinions, but those of others.
Stockfish: Quickly finds profound ideas but upon going deeper its ideas may turn out to be not so good. Poor endgame knowledge. As I have pointed out in previous posts, it’s evaluations are usually much more optimistic/pessimistic than Houdini’s.
Komodo: Described as boring and unimaginative. Reasonably solid in all phases of the game. One player commented that when going through Spielmann's Art of Sacrifice in Chess he was surprised how many times Komodo liked the sacrifice more than other engines did. I’m not sure how that translates to the opinion that it’s boring or unimaginative unless it will not choose a sacrifice and has to have it shown to it. I’ve heard this engine plays human-like.
Houdini: Poor understanding of initiative/tempo, strong in endgame and leading up to endgame. Tactically solid. Sometimes flops around when there is no clear cut “best” line.
Shredder: Very refined understanding of tempo especially in early opening, greater understanding of many types of positions especially endgames but somewhat slow. This can result in it blowing won games.
Fritz: Sometimes heads for sharp positions it should avoid. Objective about types of openings. Can be predictable.
Critter: Tactically strong, good in the openings. Likes making pawn moves in the opening, somewhat ignorant of hypermodern strategy, good at middlegame planning, reasonably good endgame. Good at seeing options.
Rybka: No comprehension of early pawn structures or tempo. Slow ply progression, great at middlegame planning, decent endgame. In some early positions with many pieces on the board it can hit the search horizon and not progress. This can result in it falling into tactical traps. Not as good in the ending as Houdini. It’s programmer said he specifically worked on improving its understanding of attacks on the King in later versions. Versions up to Rybka 2.3.2 were relatively weak tactically compared to other top programs.
Junior: Very good at handling the Q. Too bold and too sacrificial to end up with good endgames. Sometimes sacrifices for no good reason. Understands hypermodern openings better than most engines. Not too useful for analysis because it’s too aggressive in pruning in long games. Often gives the most troublesome line (for the opponent) with regards to OTB play.
Naum: Only engine to have a reasonable grasp of fortress positions.
Spike: Does not gain much strength with extra time.
Vitruvius: One player didn’t think it is aggressive enough and its play was conservative.
Chess Tiger: Organizes attacks on a king very aggressively but burns its bridges by creating positional weaknesses.
WChess: plays like a human grandmaster. Games are virtually indistinguishable.
I can’t be sure of the validity of this information because I don’t have a lot of experience with most of them, but, hopefully, the comments will serve as a guide in deciding which engine to use. For OTB players who are looking for human-like play, it appears that good choices would be WChess, Chess Tiger, Vitruvius, Naum, Junior and Critter.
If you are looking for an engine that plays most like Bobby Fischer who wasn’t too shabby a player, in my post
Which Engine Plays Most Like Bobby Fischer? I reference a link to a site where the author tested a lot of engines against the games in Fischer’s 60 Memorable Games and came up with Gandalf 6 which came in with a 69% matchup. Others were Chiron 1.5, IvanHoe 9.46h and Fritz 11. Shredder 12, Stockfish 2.3.1 and Rybka 3 came in 7th to 9th.