Wednesday, June 20, 2007

Improve your chess NOW

Yesterday I picked up a second hand copy of Jonathan Tisdall's book "Improve your chess NOW". Hattip to Blue Devil. These are my first impressions.

Alexander Kotov formulated for the first time a fundamental theory about chess improvement and especially about calculation improvement by means of the tree of analysis.
He hardly can be enough honoured for that. At the same time, while reading his work I realized that his idea's had to undergo some modification to make them applicable in practice. This doesn't detriment his original idea's, but adapts them to practical use.

While I was developing these thoughts, Blue Devil pointed on Tisdall's book. Tisdall has used the method of Kotov and developed it further by adapting it to practice. Two new idea's stand out in the first two chapters.

First Tisdall describes a method of "internal speech". During analysing a position you talk to yourself like "f8 is the only way the black king can escape mate, how can I prefend him from going there? Ah, Bc5 will prevent that!" etcetera. I already had realized the importance of internal speech while reading Kotov.

Second he advocates blindfold chess as a method to improve your calculation skills. He proposes (in my own words) to visualize the positions at the knots of the tree of analysis as stepping stone for the calculation of the following series of moves, thus hopping from knot to knot.

One thing his chapter about blindfold chess shows is that there is a lot of confusion about the used terminology. And indeed that needs a lot of refinement.
To give an example:
If blindfold chess leads to superior calculation, you would assume that blind people would all be of expert level at least. Since this is not the case, blind people do something different than what we use to call blindfold chess.
I have played a few times against a blind guy. He has a special board and feels the position with his fingers continuous. He has the same problems with projecting the future positions before his mind's eye as everybody.

While playing blind I can beat most non-clubplayers. But when I play a complex position then there are 3 elements that play a role. First the visualisation of the stepping stones. Second the calculation of the series of moves that leads to a new stepping stone. Third, the problem at hand. If you can't solve the problem, the other two are of no use.
While playing blind against a non-clubplayer, the third point doesn't play a big role. Since I can beat them anytime, the problems I have to solve are very simple in general.

All this needs to sink in and needs more testing. All my previous conclusions about visualisation are in need for an overhaul and need refined terminolgy and definitions. Exciting times lie ahead!


  1. I find it interesting that Tisdall criticizes the 'candidate moves' method of Kotov, where you analyze the full out bush for a given candidate move. Tisdall says this is unatural, that we naturally want to think through a particular variation to its end before we go off on sidelines. For Tisdall, instead of thinking ahead a move, listing all candidate moves, and then investigating each candidate, and so on, you instead pick your 'favorite' candidate and try to find fault with it (process of elimination).

    But note that Tisdall's suggestion is more than just avoiding Kotov's bush :) ; basically in the new method you climb Tisdall's tree and see if it breaks. If it breaks you need to go down and find a stronger branch to climb: basically once you find a reasonable move don't spend a ton of time trying to find a better move: this is an additional point beyond mere rejection of Kotov's bush. You could be Kotov like in that you explore many trees and branches, but Tisdall like in that you explore them quickly to their end-points rather than spending a ton of time down in the branches).

    Soltis in How to choose a chess move discusses when Kotov versus Tisdall versus two other strategies are called for. In more sharp, tactically rich positions, a more thorough consideration of alternate candidate moves is called for (more Kotov-like searching of move space (in the sense of thoroughness, not in the sense of stopping to consider each candidate move at each branch point)).

    OTOH, Soltis points out a danger with exploring a single line (Tisdall-like) in detail. While it is helpful in that it can often reveal aspects of the position that lead to even more interesting candidates, it can also blind you to tactics in the position because of what he calls the law of diminishing returns for tactics.

    Because of this, I always use a two-step procedure. I first just inspect the board, hoping my module will show me some tactics. If I see nothing, then I start a more Kotov-Tisdall like analysis (but I do very little calculation unless the position is sharp, as discussed here (incidentally, one of the most helpful bits of practical advice I ever got is there: don't do heavy calculation in quiet positions...probably obvious to you but it wasn't to me until recently)).

  2. Blue,
    there are 3 techniques to learn:

    Series of moves (the branch)
    The position (the knot)

    there are two techniques that can be used here. Inner speech (selbstgesprach) and what I called backward thinking in a previous post. There are positions that I can't solve even if I use an extra board and move the pieces by hand. If even a physical board is of no help, then a virtual board in my minds eye won't be either.
    The process of reasoning must define the candidate moves and their priority. Otherwise you get lost in the thickets.

    The bald branch is a series of moves (a film) what you can process in a serial way. This we seem to call "calculation".

    The knot (a picture)is the place where we can get lost because we forget which branch we already followed and which not (conscious parallel thinking is not possible). This we seem to call "visualisation".

  3. i read tisdahl, and thereby found many leads in my current study plan, not least of which, dvoretsky variously et al in his School of Excellence book series (less the perhaps unnessary openings volume, which, even now, is dated--it is said), gelfer's positional chess handbook.

    as i recall, the importance of tisdahl was pointed out by Aashard in his Excellinng at Chess, which despite enormous popularity and acclaim by all at the time, was suggested to me in ten years no one would still be reading it, surely not in forty, and now seems to be true. we never hear of this book now!

    from tisdahl and dvoretsky i am indebted the idea of fewer studies in attempted mastery of their content, but deeper study. to the point of memorization upon key positions.

    i do not see silman, pachman, tisdahl, aashagard, dvoretsky, art of the middlegame, soltis, of nimzovitsch as anyone being primal. i read them all, and to me they are one.

    i like Kotovs book Play Like a Grandmaster to be the best of all of these, and perhaps enjoyed tisdahl the most. of course, i read Kotovs big book first.

    the one that still awaits me, and according to Watson yesterday on ChessFm is that his Pawn Structure book is a very important book, which i plan to put along with Secrets of Pawn Endings, and Schereshevsky's Endgame Strategy, the cats meyow.

    warmly, dk


  4. Your comment about Tinsdall speaking about the position to himself reminds me of something I see in Bronstein Writing. he advocates talking internally to the pieces about where they should move. What I find striking about Bronstein is his imagination (he call fantasy) and his love of chess.

    in fact, I remember reading a game score that Bronstein dreamed he played in his sleep.


    The Dream game

  6. Tak,
    so it was Bronstein vs Bronstein, actually.

  7. Check out Kenilworthian for a recent discussion of the importance of the internal dialog to help integrate the concrete variation crunching with the abstract principled understanding (as Tisdall would put it).