Monday, April 06, 2009

Chess for Zebras, Chapter 7 (Rowson's criteria): Part I

Here's the first half of my summary of Chapter 7 of Rowson's Chess for Zebras. The Chapter title is 'Something that works for me', and it is the second chapter in Part II (recall Part II is called 'A mental toolkit for the exponential jungle').

Introduction and a caveat
In this chapter Rowson gives us the criteria he uses to evaluate a chess position. Evaluating a position is simply judging the strengths and weaknesses of the position for both players. Such evaluations are typically word-based explanations that patzers like me find so helpful. Rowson, of course, has spent an entire book trying to steer us away from such wordly pursuits. Indeed, even at the beginning of this chapter he repeats the chorus:
[T]hinking of such models explicitly while your clock is ticking will generally do more harm than good. Anything other than the images of moves and variations is likely to be unhelpful noise in your head that will lead you to create narratives based on applying the model to the position. This awkward predicament leads you to try to fit the position to what is in your head, rather than allowing you to concentrate on the position and enjoy the experience of playing...[Y]ou will gain more from the material in this chapter if you allow it to 'seep in' subtly and quietly, rather than using it as some kind of checklist during your games.
So, given that extremely important caveat (after all, the above quote captures the central thesis of the entire book thus far), let's see how Rowson evaluates a position.

Rowson's criteria
Rowson uses four criteria to evaluate a position: Material, Opportunity, Time, and Quality. As we will see, since the final criterion actually includes four subcriteria, he really has seven criteria. Let's get to it. In this post I'll discuss Material and Opportunity. In the next post in this series, we'll see what he means by Time and Quality.

For chess mortals, this is the God Criterion, the criterion that even grandma will use to evaluate a position. Who has more material? Against the usual crass point-count system (where Queen is worth 9 pawns, the Rook 5, Knights/Bishops 3), Rowson emphasizes that a piece's strength must be determined based on the concrete features of the position. What a shock: the guy who advocates transcending the rules suggests we shouldn't mindlessly assign numerical values to the pieces.

Along this vein, Rowson says, "The relevant skill in assessing the material dimension is not counting on the basis of arbitrary material numbers but looking at the roles of the pieces and thinking carefully about whether a material advantage is latently relevant, already manifest as Quality, or irrelevant in that particular position, perhaps due to the Opportunity or Time situation."

Most of us are familiar with games in which a player sacrifices the exchange to gain an opportunity to attack the enemy King. If you have a mate in two, it doesn't matter if you are down three pieces. Hence, Rowson's description of exeptions to piece values shouldn't be much of a revelation.

I have wracked my brain to come up with a definition of what Rowson means by Opportunity and the following is my best shot:
An opportunity is the temporary ability to achieve a concrete result.
Examples of concrete results include a material gain, an attack on the King, or dominating an open file. Typically (but not always) such Opportunities are offered to the side with a lead in development and more active pieces. However, there also exist swindles and other opportunities that arise despite our best efforts. Rowson's Opportunity category naturally absorbs such happy accidents.

As I insinuated in the previous paragraph, Rowson never actually defines Opportunity, and annoyingly says "Examples of Opportunities taken or missed can be found throughout this book, so for now I focus on only one example." To prepare for this summary, I looked at his single measly example in some detail and wrote about it here.

In that example he claims that with 8 Nf3, white sacrifices pawn structure for Opportunity. It seems that White gains what most authors would call an advantage in piece activity or time (e.g., black would need extra moves to "catch up" with white's development). White's pieces are developed, on good squares, while black still needs time to get his army mobilized.

Given that his example was a clear instance of generating piece activity , why don't I define Opportunity using more familiar notions of piece activity, time, or development? Because according to Rowson, while these traditional notions are close relatives of Opportunity, they are not equivalent::
I felt it was a mistake to refer to (non-clock) 'time' as a dimension of chess. Firstly, because it is easy to conflate and confuse with clock time, and secondly because the significance of 'time' varies enormously. It seems to me that where it makes sense to speak of an advantage in time, it is an advantage in terms of available opportunities. A lead in development only matters if you have the opportunity to do something with that development, and 'the initiative' is only important insofar as you can do something with it.
In other words, there is no guarantee that you will be able to turn a temporary lead in development into a longer-term advantage in the position.

The example (8 Nf3) I mentioned above is a case in which activity is the concrete result gained. This doesn't mean all Opportunities will be activity gainers. If I have an opportunity to grab a Queen for free, I'll do it. Hence, my provisional definition in terms of 'temporary ability to achieve a concrete result' seems to be general enough to absorb any species that Rowson wants to include.

Rowson hasn't made a very good case that we should ditch development as an evaluation factor. By analogy, it isn't always possible to turn a material advantage into checkmate, but that doesn't mean we should scrap Material as a criterion for evaluating a position in chess! It would have been nice if Rowson had delineated the different species of Opportunity, just as he delineated the different species of 'Quality' (which I will get to in the next post about the book).

My opinion is that a lead in development is still a lead, and Rowson is sacrificing usefulness for correctness by not including it in his criteria. This is a common error in treatises on evaluation criteria in chess. I have discussed this phenomenon here, where I said "We aren't constructing an axiomatic system, where we need to worry about finding the smallest independent set of axioms that can prove all the theorems. We are playing chess, where we presumably want to evaluate the board using factors that are useful and help us make good decisions quickly."

This is the my least favorite chapter of the book thus far. The writing displays a certain intellectual laziness that was not present in earlier chapters. Rowson doesn't define 'opportunity' but then throws the term around as if it is a clear idea (compare to the term 'plan' as used but not defined in Chapter 6).

That is not to say that Opportunity was a silly criterion to use, but to treat its meaning as transparent was a mistake. He spent five pages discussing material, with which we are all familiar, and only two pages discussing opportunity. In other words, he spent the least amount of time on the factor we are most likely to be confused about, so Rowson screwed up his priorities as a communicator. The reader shouldn't have to put in four hours of work to figure out what Rowson means by Opportunity.

As a final related quibble, Rowson shows way too much respect for Hübner's ideas, and my hunch is that Hübner's influence polluted this chapter. Hübner is a top German GM, and clearly had a big influence on Rowson's thinking. It actually becomes funny in Footnote 12, where Rowson respectfully acknowledges a critique from Hübner of Rowson's use of the term 'dimensions' to describe the four evaluation factors. Hübner obviously doesn't understand how the word 'dimension' is used in mathematics, so his critique is simply wrong. Hübner may be a great chess player, but that doesn't translate into being a great ordinary-language philosopher of chess.


Blogger Michael L said...

BDK, thanks for this important post. I've read the book, but this was one of the chapters I was most confused by. He argues that we should stop using words and narratives, but when we use his four dimensions to evaluate a position, aren't we using words to understand the position? For example, if I were to take his criterion of material to evaluate a position I would be forced to use words: "Ah yes, my opponent's knight is passive so I shouldn't exchange my awesome bishop for such a lousy piece because the bishop is worth more than the knight in this position." At the point that I am stuck with using words in using his own criteria, I don't see how we are suppose to utilize his thinking model if it only devolves into a contradiction.

4/06/2009 02:16:00 AM  
Blogger Blue Devil Knight said...

Michael: good point--that's why I included that extensive caveat from Rowson that this is something you shouldn't use OTB.

OTOH, I think ultimately you are right. There has to be an end-node in the tree of analysis, so what are we to do with that final node? We have to evaluate it!

Also, I find an evaluation of the initial position a key way to find candidate moves. Perhaps Rowson's intutions are so well developed, the best 3 or 4 candidate moves just pop out at him as obvious, with no word-based evaluation needed. For me, realizing 'He has incredibly weak dark squares around his King' is useful for generating further plans/candidate moves.

4/06/2009 09:25:00 AM  
Blogger Loomis said...

I'm on your side regarding using word analysis to do final evaluation and candidate move generation. But I have the sense that this is not how the strongest players are doing it. I believe they are recognizing "don't trade bishop for knight" or similar without ever thinking it. If you show them that move they cringe before they can even formulate why it's bad. I believe they process not only variations, but also narratives subconciously.

4/07/2009 05:28:00 PM  
Blogger katar said...

Hey, i totally agree with Rowson's ideas about staying away from checklists and such (verbal "chatter") during a game. Much better to just form a gut feeling about a position (whether it is objectively right or wrong is not so important). And then only AFTER the game apply more objective analysis and checklists and whatnot (with coach or stronger player) to that gut feeling. I think you can tweak your intuition this way over time, but it takes humility to admit when you're just simply wrong, eg, like when i thought i was better all game but in fact it was simply equal, which my equally-rated buddies will let me know when we analyze together at a restaurant. Idnetifying and correcting those instances when your intuition fails you is where you can really drastically improve your game by leaps and bounds.

I think the goal is to see whole variations at a flash as an interaction of force vectors rather than discrete moves. I recently got Volokitin's Perfect Your Chess and after setting up each position i just look at it until an idea jumps out at me. Just try to "see" the inner workings of each position as an artist would view his subject. I think really absorbing master games (Alekhine, Bronstein are the best annotators) by contemplating each move and comparing to the notes is a great way to build intuition. At least, i say that as a person who read half of Zurich 1953 and about 2/3rds of Alekhine 1923-1937. Just my opinions. :) peace out.

4/07/2009 06:01:00 PM  
Blogger Blue Devil Knight said...

Loomis: that is an excellent point. I have certainly had that experience--go down a couple of game trees, and one branch just "feels" better and only later does it become clear why I felt that (and whether I was right).

Katar: excellent point that resonates with what Loomis says, so you are likely right.

My problem is that my intuitions are so off sometimes, I can't trust them.

Clearly good players don't go through a checklist in their head! My guess is much explanation after a game is post-hoc and ad-hoc. At the time they weren't thinking about any of this stuff!

Personally during a game I often think things like 'My bishop is about to become cut off from the game' or 'His light squares are weak', and these thoughts are, on average, probably helpful.

So while I agree that it would probably be counterproductive to do a checklist of sorts, thinking in terms of various criteria doesn't seem to be counterproductive (though perhaps one thing I should have learned in my planstorm session a few weeks ago is that indeed I do think too much about such positions, tend to have an overactive planning organ, spinning out narratives left and right about the position, and I don't do enough intuition-based or visualization-based thinking).

You know, like the hippies would do.

4/07/2009 11:54:00 PM  
Blogger wang said...

I find it odd that he would not define ths term. This is very disappointind coming from any author, much less one who's written a book such as this one. I am always upset when teachers/coaches are vague.

4/09/2009 11:05:00 AM  

Post a Comment

<< Home