For this, I’d recommend something like Toga or even something like Crafty would do.

Computers are tactically very strong, and now-a-days even in endgames due to usage of endgame tablebases. But they can be strategically weak. Even though engines like Rybka can make sound decisions and still improving.

Many of us play against chess engines, under constrained strength, to get some practice. Of course, engines don’t have the human factor, but they can punish you severely if you make tactical mistakes, especially at higher strengths. But if you’ve capped the engine’s strength to say below 2000-2400 ELO, it’ll sometimes blunder tactically too. At lower levels, it can blunder a piece, at higher, maybe a pawn. I don’t know the exact details, but the ‘complexity’ of the blunder can also vary from a simple hanging piece to hard combinations.

Let’s say you can beat the current level to which you’ve set your engine to, not very easily, but with decent thinking. Apart from raising the level of the engine, what you should try to achieve is to take advantage of every tactical and possible non-intentional strategical blunder than your resourceful opponent makes. At a very low level, engines make tactical mistakes easily visible to an intermediate player, but maybe not to a beginner.

Since you know the engine won’t pick the best move every time, if you’ve constrained its strength, your job should be to try to find them during the game. This way, you can improve your tactical vision against an opponent who is willing to make deep tactical mistakes to train you.

Once a game is played, you can analyze it to see if you missed any tactics. Apart from that, engines might sometime not understand the significance of a strategic move. Depending on your skill level, you may realize this and ignore such lines. But be sure to check out the line completely, for your own satisfaction. Seeing the rate at which engines are improving, you won’t come across such problems that easily as you would have a few years ago.

Thus, apart from increasing of engine strength, you could even play match against opponents you’re comfortable with (not opponents which are just too easy and make stupid mistakes. The definition of stupid move is different for different levels of players) and see if you can catch its mistakes.

I personally own Chessmaster 10, and it has a lot of tutorials for beginning players etc. I’m not an excellent player, so this engine strength should suffice for now. But I’m eying engines like Fritz and especially Rybka, as I progress.