Wednesday, April 09, 2008

Read for Yourself

On reading the excellent dk-transformation's post at Chess Improvement titled Raming GM-Ram I was directed to a series of articles by GM-Ram author Rashid Ziyadinov over at Jeremy Silman's site. Apart from the merits of GM-Ram as outlined by dk, I recommend you too peruse these articles; they have some different and unorthodox views on chess and chess improvement compared to most, and whether you agree with any or all of the author's contentions I think you'll be stimulated and look at things in a new light.

Here's one statement that really struck me (see Part 3, "Concrete Chess):

Students sometimes lament that they cannot apply their knowledge during a game. They cannot apply their "knowledge" because they really don't have any knowledge! What they have instead are shortcuts to chess language, what I call "chess magic spelling" (like "open the position if you are more developed", "the two bishop advantage", "don't move the queen out too early", etc.). These shortcuts are useless generalities. Chess can only be expressed with concrete variations. This often-ignored concept is so crucial to mastering chess that it bears repeating with emphasis: chess can only be expressed with concrete variations!

To sum up, I note that my game annotations here often have some pleasing verbal description, but only a modest amount of "concrete variations," while the stronger the player, it seems, the more the annotations consist mostly of variations. Which is cause, and which effect? Any opinions, Dear Reader?

11 comments:

transformation said...

to my esteemed co-blogger. i read and reread this post, and try to have {Jap: hopo-moku} 'eyes wide open' as Sensei called it in Shorinji Kempo.

i try to read and 'see' instead of judge.

i feel that you have two posts here, is that correct?

one at once drawing FURTHER attention to GM-Ram as worthy of a serious look (thank you also for the acknowledgment, it is somewhat but not decidedly cold and alone out here in ++ GM-Ram land),

and is also a post touching upon the methodology of analysis of ones own games or your games?

i see no link in these two and, if they exist, kindly ask for your elaboration. i am not one to insist that 'consistency is' NOT 'the hobgoblins of small minds' but, then again, if all we wanted was to present our ideas and experiences without critical input, we would all journil beyond the view of others...

warmly, dk

Blue Devil Knight said...

I really like his articles, which have strong and extreme positions on everything.

My two cents on his claim.

I think that in practice he seems to be wrong on this. In a position is relatively quiet it would be a bad use of time to do extensive variation crunching, time that should be saved for sharp positions with lots of forcing moves. In such quiet positions, in which extensive variation crunching would be a practical blunder (Soltis rightly says, spending too much time on a move can be itself a blunder), it is really helpful to use general principles to make a faster move, saving time for the tough positions that absolutely demand deeper calculation.

I haven't seen that GMRAM book but it sounds horrible. No annotations, just a bunch of positions? It makes me cringe. I do understand what he is trying to do with the book, but would it have hurt all that much to give his take on the positions in an appendix or something? Would that actually hurt the reader, encourage reader laziness? It sounds more like a classic case of author laziness, analagous to informant-style game dumps that are sold as opening repertoires.

That said, I know many people love the book, so perhaps he really has picked some wonderful and illuminating positions (which presumably provide insights into general principles that can be useful in real games--hey wait, didn't the author say that general principles are useless?).

Anonymous said...

Thanks for pointing Ziatdinov's articles out - they are both interesting and entertaining!

To your question about annotations: Yes, he is probably right that ultimately only concrete variations count. But he also writes in one of his articles (forgot which one, sorry) that there are only 4 possible evaluations: 1-0, 0-1, draw, and unclear. When I read through variations presented by a GM, I hardly see 1-0 and the like as their outcome, but most likely evaluations like "White has the clear advantage" (which is usually not clear to me). So interpreting Ziatdinov strictly, these variations are pointless, too, because at the end they too depend on "soft" evaluations like a better pawn structure. To me, the main difference between GMs and patzers like me is, how obvious an advantage has to be to come to a conclusion and to stop calculating a variation.

And after all, doesn't he recommend himself not to look much deeper than at 1-2 move combinations, at least consciously? I think he presents good advice, but that advice shouldn't be taken too literally I guess.

Robert Pearson said...

dk--yes, you read it right, I wanted to draw attention to your post, the book and the articles at J. Silman's site; the paragraph about concrete variations was an example of one (among many) that struck me during my reading, and seemed to apply strongly to me personally. Other parts for other people, I presume.

Honestly, this post reads as a bit of a mishmash, it is true. Not the clearest work I've ever done, but I hope it stimulates reading and thought.

Thank you for your perceptive and constructive comment.

transformation said...

Ok BDK: there goes your pronounced tendancy again to make others wrong and criticism as a habitual way of being.

my turn now. who told you that you were less so that you must so often put yourself there? i thought that you quit blogging or dedication to chess? what do you want now? do you know what it is?

you havent seen the book or the positions and how it is assembled and all this. oh my god. this makes no sense and is silly.

see how that one feels. :)

Blue Devil Knight said...

Watson and Aagaard have had it out over the issue of general rules versus concrete variations (Watson responds to Aagaard here in fairly strong words...it is quite entertaining).

As I said above, the most reasonable position I've read is not extreme (all rules versus all variation crunching), but that the right approach depends on the type of position. I got this from Soltis' book 'How to choose a chess move', but it seems to be agreed upon by many GMs (summarized here).

Speaking personally, my play improved a lot when I stopped spending a lot of time doing really long thinks in quiet positions. The ability to prioritize ones thinking time is really important especially in slow games. Heisman discusses this frequently, too. So I look at it fundamentally as a question of time management.

As for my comments on his GMRAM book, looking back over them, they still seem reasonable, but of course I am always open to revision in the face of a relevant intelligent criticism.

P.S. The irony of DK's message is exquisite.

Anonymous said...

The weaker the player, the more description is needed. It probably depends more on the target audience than anything. We're probably all familiar with seeing a variation, the writer giving a -+, and we're sitting there saying "Huh?" Then we realize that the commentary was (hopefully) not written for the class player.

Unrelated: The better you know the participants of a potential flame war, the more uncomfortability is generated when one starts to brew.

Blue Devil Knight said...

LEP: I think you are right about the target audience. Another example is in endgame books--you have the position dump style (learn these endgames), and the Silman-esque style that takes you by the hand a bit more.

I even met someone who got annoyed with Chess Tactics for Beginners, as they wanted an explanation of some of the tactics. Luckily there are tons of explanation-rich tactics books out there to start with.

The best method for finding quality work I've found is talking to people a couple of hundred points above me, finding out what helped them the most. Asking masters (who are not seasoned coaches) I have found particularly unhelpful. It is usually so long since they have read a book, and they are so out of touch with what it is like to be a truly shitty player, that they often end up recommending stuff way out of my league (e.g., 'Secrets of Pawnless Endings', a great book I've been told, but just way beyond what I am ready for).

Glenn Wilson said...

"Chess can only be expressed with concrete variations."

Chess is 100% tactics.

Now, how one arrives at the variations, the ideas behind them, etc can and does use general principals and other shortcuts.

Bot no one ever plays an idea. They play a move. A series of moves is a variation.

"These shortcuts are useless generalities." I disagree with this in that these shortcuts can be helpful in finding moves to consider.

Blue Devil Knight said...

I am surprised to agree with Glenn a bit here. I also use general principles to evaluate the end-node in a line I'm visualizing during analysis of a position. All else being equal, I'll go down the line with a rook on an open file.

That said one interesting thing I've noticed now that I've stuck to a couple of openings for longer than I used to, some principles have shown themselves to be simply critical in positions reached with the opening, others less so, some actually need to be violated.

But yes in practice, we all use "general principles and other shortcuts" (a nice turn of phrase).

But even the best computer programs use principles--will increase the evaluation function when it has a rook on an open file, that sort of thing, and they play much better when such rules of thumb are incorporated into their evaluation functions.

That said, since they tend to weight material above other factors, such factors will have a greater influence when the position is relatively quiet. Just like a human should do! And I think this is no accident--the evolution of chess evaluation functions in computer chess too a while to catch up to how the best humans actually evaluate positions.

I wish the companies would publish their evaluation functions. It would be so cool to see!

Anonymous said...
This comment has been removed by a blog administrator.