We each have a set of chess beliefs that directs how we play and study the game. The right set of beliefs can lead to good results and steady improvement. The wrong set, where we latch on to beliefs and hold them to be true when in fact they aren’t, can stagnate our play. Regardless of whether our chess beliefs are right or wrong, I think frequent re-examination of them is healthy, and is good for getting our creative chess juices flowing.

Let’s do just that, starting with a list of some popular beliefs.

Belief #1: Always play the best move, and assume your opponent will do the same.

Belief #2: Every position can and should be assessed accurately and objectively based on the characteristics of the position.

Belief #3: Opening theory is constantly evolving in the direction of best play.

We hear Belief #1 everywhere, and it’s certainly well motivated. You won’t get far in chess if you always play for crude tricks that you hope your opponents won’t see through.

Belief #2 seems plausible. Certainly publishers and chess programs seem to think so. Chess publishers have created „hieroglyphics”, sets of cryptic symbols, that allow an annotator to specify whether one side has a slight, moderate, large, or decisive advantage. Computer programs are even more refined. They get down into the smallest fractions of pawns to measure which move is better.

Not only is Belief #3 reasonable, it’s an implicit assumption in how many of us study openings. We get the latest opening books and then slant our repertoire towards those lines that promise a slight advantage rather than equality.

Together the beliefs seem to make a nice foundation for chess study don’t they? They’re simple, and they seem self evident.

Over the years though, I’ve noticed some things that just didn’t seem to fit the nice clean theory. Some observations in fact seemed to be real paradoxes.

Observation #1: Back in the days when computers could play humans in tournaments, one computer appeared close to victory against its human opponent. Suddenly, the computer sacrificed a rook for no compensation, landing in an immediately resignable position. A bug? No. The computer saw that it was losing because of a surprising and winning tactic for its opponent. None of the humans in the room, including the one playing the game, saw the tactical possibility. The computer played the best move according to the standard criterion. The paradox is, the best move lost instantly, any „weaker” move would have won easily.

Observation #2: Several years ago I watched a master trying in vain to win a drawn ending from a lower rated opponent. He had already tried all reasonable winning attempts so I was expecting a draw offer. Instead, the master deliberately played a move he knew would lose if his opponent responded correctly. His opponent didn’t find the best reply, and the master won. Afterwards the master explained that he was „pretty sure” his opponent wouldn’t play the winning move. An approach to chess that allows one to deliberately play losing moves was completely alien to me, as it probably is to most serious players. The paradox is, if that master guessed right over half the time he’d come out with more rating points than if he always chose the „best” move.

Observation #3: Some players criticized for not playing the best moves seem to have achieved a large part of their success *because* they didn’t play the best moves. Lasker tried to play moves that would make his opponents feel most uncomfortable. Tal often played sacrifices by feel, not worrying too much whether they were 100% correct. Probably neither player would have achieved such great success if they were constrained to always play the „correct” move.

Observation #4: Ask any Grandmaster or any strong chess program to assess a theoretically drawn K+P versus K ending, and they’ll say „equal”. But it *can’t* be exactly equal. We’d all rather have the extra pawn because of the possibility that our opponent might blow the ending. But if the ending isn’t exactly equal, how can we state objectively and accurately how much advantage the superior side has?

Observation #5: Speaking of objective and accurate assessments, if you assume best play for both sides, „slight advantage” or „.27 pawn advantage” become meaningless. The only possible best-play-to-completion assessments are win, loss, or draw.

Observation #6: I see no evidence that the popular openings today are any „better” than the openings played 10, 20, or 50 years ago. Lines go out of fashion, then come back in. Grandmasters continue to specialize in openings that supposedly aren’t as good as the most popular lines.

All these observations call into question how we should assess positions, and ultimately, how we should study and play. Playing the best move and assuming your opponent will do the same is not always a good approach. It may be impossible to assess a position so that the assessment is at the same time, objective, accurate, and useful. Assessments provided by the latest opening theory may not be useful guides.

So does that leave us without any foundations at all for our chess study?

No, there is a foundation for chess study but I think the rules are a little different than we’ve been raised to believe.

Here’s a viewpoint that not only addresses all the paradoxes I’ve raised, it also suggests some new ways to study chess.

Start by visualizing the tree of all possible moves for you and your opponent. It’s like a journey through potentially hazardous territory to your destination. Do you have a map to tell you which turns to make? Can you recognize landmarks along the way to guide you? Each move is like a fork in the road, and at each move you have to choose a path.

Next consider how you should assess each position you’ll face. It’s essentially a navigation problem. If it’s easier for you to choose the correct path than it is for your opponent, you have the advantage. How much easier determines the size of your advantage. Remember how you or Anand played this line before? Then you have a map to guide you. What if the position is new to you? Then remember how Kasparov played a similar position last year. That gives you a landmark to go by. You’re also in pretty good shape if you have only a couple of paths to choose from and they both look hazard free.

Note what this navigation analogy incorporates. Assessing a position depends on the navigation capabilities and resources of both you and your opponent. The view that encompasses Beliefs #1, #2, and #3 would suggest that there’s a single optimal path to your destination. The navigation analogy says the optimal path depends on the landmarks you and your opponent know and what maps you both have.

Treating chess assessment as a kind of navigation problem is scarier than a more traditional approach because it recognizes that the board isn’t enough. What your opponent knows is critical for accurately assessing a position but most of what your opponent knows will be hidden. The more you can study your opponent before the game, the less guesswork will go into your assessments during the game.

On the other hand, the navigation view gives you some new freedom and power in assessing chess positions. You no longer have to rely on an outside assessor, like an opening manual, to tell you how good a position is. Through study of your games and Grandmaster games, you can create your own maps and set of landmarks to help you navigate the opening. You may also find that the navigation view affects your choice of openings. Instead of considering only what you like, consider what your opponent doesn’t like. Drag him into hazardous territory where you think he has few landmarks and maps to guide him. I think this is happening in opening theory all the time. Opening theory is not evolving towards better lines, it’s flowing to the new and forgotten. That’s where you can nudge your opponent into unfamiliar territory and pose him difficult problems. And beyond the opening phase, try to make sure every fork in the road gives your opponent some chance to head towards a hazard. As long as every path he might take is smooth and hazard free, you might as well be playing a Grandmaster.

Forget about the hieroglyphics and fractions, even if they’re generated by a strong chess player. What you know about a position affects that position’s assessment, so, not only are you the best judge of a position, you can control the assessment through your study of chess navigation.

*****
This article appeared in the May 2004 issue of Northwest Chess.