Are computers bad at strategy?

When reading a chess book (such as Logical Chess Move by Move), I enter the moves on the computer so that I don't have to visualise it all in my head. I use Fritz 9. Sometimes I have the 'Infinite Analysis' window open at the time, to analyse each position in the game.

The move might be 'Qd3!', and the book will say that it's a fantastic strategical move, and yet Fritz will sometimes suggest that there is a better move, such as a tactical move which wins a pawn — even if I allow Fritz to analyse the position for a long time.

In this situation, is Fritz wrong, or is the book wrong? I'm hoping that Fritz is wrong, and that Fritz is a master tactician but not so hot on strategy. Or is it simply that computers are slowly taking over, and proving that tactics is greater than strategy?! (I hope not!)

Any thoughts?

Thanks

If there's a pawn that can be taken with no particular consequence, then it is certainly the book that is incorrect and not the engine. However, if capturing the pawn leads to difficulties... then perhaps the book is correct.

Turn the ****ing engine off.  Ivring chernev was a good player, and an amazing author, but not nearly world championship caliber, and made a few mistakes here and there.

The moves he gives a ! however may not be the best moves, but they are always good moves.  By looking at fritz and finding the BEST moves in a position, you are missing the point.

Think of it this way, whats better to know, knowing a move that puts you a pawn ahead but rarely comes up?  Or knowing a move that puts you a half a move ahead and comes up all the time?  Until you start reaching master class, the latter by a mile.

Again, fritz is a distraction, turn it off.  TURN IT OFF.  Knowing the best computer move isn't always helpful.

5pin, you seem a little angry! I was just asking a question.

I guess I was just worried, because I've seen this a few times in Fritz, and
every time I see it, I think, "Surely it's better to play the strategical move, and end up with a sound position, than to grab a free pawn and have to catch up in development." Then again, I'm not an expert (only ~1600), and maybe I'm missing something.

Unfortunately, I don't have any concrete examples to hand. If I stumble upon one, I'll be sure to post it.

Thanks guys

Dont trust Fritz, trust Rybka.

Leave the ****ing engine on!  Ivring chernev was a good player, and an amazing author, but not nearly world championship caliber, and made a few mistakes here and there.... the engine will show you alternatives he missed. (Apologies to 5pin for the re-mix).

I agree with the 5pin's other points.. like it or not, you're a human, and Chernev is showing you good moves that human's find and can make use of... sometimes the computer will show you useful and interesting sidebars, sometimes it will show you things that no human could or would ever play.

There are still some strategic positions that mess up computers -- but in general, computer's tactics _do_ trump human strategy. The Machines have won. Buy stock in Cyberdyne corporation if you know what's good for you ;)

"1. e fo-wa? Dats no gud... I will jeckmade you in 19 moofs"

.

I also often use a computer when studying so I can actually play a position several times instead of just pushing the pieces according to an example. I was going through Silman's endgame book, practicing endings, when I stumbled onto an example that just doesn't seem to work OTB the way it does in the book.

http://www.chess.com/forum/view/endgames/bishop-endgame-analysis

5pin has a good point though. Using an engine to find the absolute best move may keep you from appreciating the overall positional considerations of any given example.

I would say the you should neither turn the engine off or leave it on, but instead don't turn it on until you have read through the entire game and come up with your own conclusions. Then check  all of that against what the computer might say.  And then play out the examples the computer gives.  See if you like them or if you like the move in the book better.  That is what my coach makes me do.  I love Chernev by the way!!

The original question was: Are computers bad at strategy?

I would observe that the best computer chess softwares are nothing more than tree pruners with evaluation algorithms, biased by the choice of openings in their repertoires.

So, where does strategy come in? Depends on what you mean by strategy?

If by strategy, we mean the presence of a plan, chess computer software generally has no plan a priori, though a human might deduce a plan from the outcome of chosen moves. By this measure, computers have no strategy. (Although, one might argue - and rightly so - that practicing a given pruning strategy for deciding which paths to examine and evaluate constitutes a strategy. I agree, but this is not what we normally mean by chess strategy.)

If by strategy, we mean consideration of positional factors, chess computer software  - though starting out poorly many years ago - now incorporates a great deal of positional understanding into its evaluation algorithms, in addition to shear materialism. So, by this standard, computer software practices strategical - or longer range motivation - concepts by considering the structures resulting from a given move sequence in its evaluation.

If by strategy, we mean that we learn from bad execution, then computers can even do that now - looking at their own choices and outcomes and filtering future choices by those findings.

Again, we're mainly anthropomorphizing here. Computer software is merely executing instructions, though the results are fun to personalize of course.

I just found an example of the kind of thing I'm talking about. I entered the position into Fritz, and Fritz told me to start an attack on the opponent's king, while the book says that it's best to the develop the last undeveloped piece before starting the attack.

So I was about to post this here as an example of the kind of thing I mean, but then I left Fritz analysing it for 10 more minutes, and then Fritz finally decided it would be best develop that last piece before attacking.

So maybe I haven't been giving Fritz long enough! Still, if I find another example, I'll let it analyse it for ages and then post here if there's still an issue.

Fritz, running on this year's typical desktop computer, is way, way stronger than Irving Chernev was (NM-strength). But, if I had 5,000 hours of time with either, I'd prefer those hours with Mr. Chernev.

It's interesting, though, that it took Fritz a long, long time to realise that it was best to develop that last undeveloped piece before attacking, but to a human it was a very obvious move. It was one of those moves that almost yell at you, "PLAY ME!"

 I have to admit, this seems like a better idea what I've been doing.  So do this.

Although I'm personally going to stick to the chessboard mostly, I have a hard time working with engines.

Good books can teach you more than engines. However, top class is reading a book, finding mistakes (on your own!), then checking it with the engine to see if you're right (or wrong - that's also useful). That way you'll become a good chess analyst.