Chess engines play like engines not humans so when talking about positional sense, you are comparing them to other engines...just like all engine rating lists which compare results only with games played against other engines. Engines still have a long way to go in their understanding of strategy which is why they are pretty much useless when it comes to trying to understand a game using only engine evaluation output.
Today engines don’t play blindly aggressive chess like they used to. All engines will sacrifice material for an immediate material gain and that’s often one of the problems when comparing their evaluations to that of a GM. Engines place the emphasis on material, GM’s realize there may be positional merits that outweigh the material. This especially is true in positions in which the material is unbalanced. Personally I will always rely on the GM’s intuition in these situations.
In recent years more chess knowledge has been programmed into engines. Another big change that’s been made in recent years in Fritz is that it’s being programmed to play against humans rather than engines. It’s becoming harder and harder to play “anti-computer” chess. You know, playing for closed positions where the engines couldn’t find any tactical shots, so just started shifting cyber-wood until you could form up your forces into an irresistible attack that was beyond its horizon.
I am not familiar with Hiarcs but have read that it plays human-like which I take to mean it’s programmed to have more feeling for, and give more weight to, positional themes. I have used Shredder but did not install it on the new laptop simply because I much prefer the interface and printout of Fritz. Shredder seemed to be oriented more towards positional play and I think it's evaluations were usually less optomistic than Fritz'.
Still, if I understand things correctly, all this is somewhat moot because you can tweak things like the “contempt value” for these engines and thereby alter their “style.” Raise the contempt value and the engine will play more tactical chess. The selectivity setting is also important because it determines the depth of search and size of search tree and will result in altering an engine’s style. The aggressiveness setting will also alter how aggressively the engine will play. You can also adjust the values for things like P-structure, King safety and the value placed on the individual pieces. In short, you can mess around with programs like Fritz and alter its style of play. I don’t because of my uncertainty of what kind of monster I will be creating simply because I’m not that well versed in all this stuff. Besides, it seems like most engines play their best at default settings. From my observations Rybka seems to be more modest in its assessments than Fritz which is why I prefer to use Houdini which some people have accused of being a clone of Rybka. I like solid more than speculative! However the main reason is that Houdini seems to be the stronger of all the free engines. I just finished an engine 5 minute tournament using my installed engines with the following results:
1. Houdini 1.03a w32 2_CPU..- D 1 1 1 1 4.5
2. RobboLito 0.09 x64............D - D D 1 1 3.5
3. Stockfish 1.6JA.................0 D – D 1 1 3
4. FireBird 1.0x64..................0 D D – D D 2
5. Fritz 12.............................0 0 0 D - 1 1.5
6. Crafty 23.01.......................0 0 0 D 0 - 0.5