Sunday, November 06, 2011

Odds and ends


Why do I think there is a relation between visualization and concepts? You can find the position that caused the revelation here.

Why do I think that you should start with the solution and not with the problem? The main reason is that you can find no new things in your own mind within a reasonable time. Since there are only old things in your mind by definition. New things have to come from outside.

Remember that I'm talking here about people who have quite some tactical exercises under the belt. Who have gained 250 points by tactical exercises and stall. Myself for instance.

Of course you try to find the solution yourself first in order to create some emotional bond with the solution which helps to make more impression on your brains. But you should not ransack your mind for new knowledge which isn't there. That's simply a waste of time.

Of course it is just an invitation to not spill your time by an "experience expert of wasting time". Feel free to decline.

Low level concepts.
The concepts I formulated in the  example position  in my post about concept-guided-visualization are of little use. They are not general enough. In order to be useful the concepts you are looking after should be usable in a whole bunch of positions. You can find an example position here.
I cite the concepts:
  • When I move the king it must attack something if possible
  • I must not interfere with the communications of the rooks
  • I must not put my king on an open file that my rooks may need later
  • I must not give a chance for a skewer.
These concepts are both low level and they apply to a whole heap of positions.

High level concepts.
Once I investigated bishop sacrifices at f7 in an uncastled king position. I formulated a scheme of concepts for this sacrifice.

White plays Bxf7+.
There are a few concepts:
  • Most of the time there are 3 pieces which play in tandem. The bishop, the knight on f3 and the queen. So the question is: can they follow up.
  • Without the queen you can't mate, usually.
  • Square e6 is only covered by f7, so after the sacrifice it is an unprotected invasion square for the knight.
  • There are the following ways for this sacrifice to be succesful:
  1. The king gets mated on the backrank,
  2. the king is driven out of his shelter towards g6 and mated out in the open. (Or black has to give material back to prevent it)
  3. The black Queen and King are forked with a white knight at e6
  4. The black queen gets lost due to lack of an escape route (via c7 to a5, for instance)
  5. The black queen can escape, but must give up the protection of c7, so the rook on a8 is lost (Bxf7+, Ng5+, Ne6, Nxc7, Nxa8)
  • if the black queen escapes via the a5-d8 diagonal can it be trapped there?
  • if the black queen escapes via a5 has it counterplay there?
  • if the black queen escapes via e8, g6 has it counterplay against the white king via Qxg2?
  • if you can win the rook at a8, can you free the knight?
What you see here is that a whole bunch of low level concepts form one high level concept. In order to judge the sacrifice you have to check the methods of winning as described. Will the concepts work in this specific position? Little changes to the position make huge differences. For instance: what if there is a pawn on c2? What if there is a black pawn on d6? Etc.. Btw, in this position the sac doesn't work since the king can escape to g6 where it can't be mated.

So John Watson is right, there are no rules in chess. We call them concepts and you have to check if they actually work in the position at hand indeed.

Every opening has its grand schemes. Usually they are more positionally than tactically decisive, but specific tactics play a role, of course. I imagine a dragonplayer to be a person who studies 6 exchange sacrifices on c3 before breakfast.

The preconditions of Vukovic for a kingside attack are another example of a useful grand scheme.

Low level concepts versus grand schemes of concepts.
There is no real conflict here. We summarize facts in generic low level concepts and we summarize low level concepts into generic grand schemes.
Can we say something about the frequency of occurence?
The low level concept "put your pieces not in each others way" is applicable many times per game while the high level bishop sacrifice I described has to be considered only  a few times per year, depending on the opening you play. The preconditions of Vukovic you have to consider once per game, at most. Low level concepts do happen more often than a specific combination of low level concepts (= a high level concept). So I think we have to focus on low level concepts.

I think we should try to make use of the fact that concepts have the property that they can guide visualisation. The power of visualization is that it is necessary for calculation. Since calculation is only necessary when there are tactics and forcing moves around, it is logical to focus on the formulation of concepts in tactical positions. I even think this is the fundament under the supertrick I'm always talking about.
Which is not the same as saying that we should neglect positional concepts.

When we talk about chess knowledge and transfer we are usually talking about high level concepts. Since high level concepts have a low frequency of occurence, there is the problem that they do not pop up when we need them. I have no ready method to fix this problem, but I think that we avoid it when focussing on low level concepts with a high frequency of occurrence. (Pfeww, that was an easy way out!)

In the past time I have been busy with installing a new computer and some geological stuff, so I haven't exercised all that much. To make it up with you I will devote a few posts to tactical positions and conceptbuilding, so you and I will know what I'm talking about. Let's see if my figments can withstand publicity.

BTW,  since short there is an annoying underlining under text I write in blogger or gmail.
Probably due to a dutch spellingchecker checking english texts. Has anyone an idea how to stop unvoluntary invoked spellingcheckers?

                 Especially for the Bright Knight: 


  1. me Munich:
    and now this position with the 3 pieces playing tandem - when I saw this puzzle my first question was "Where the heck is the pawn on h4?!" I saw it immediatly. Say it within few seconds. Then I continued reading your text.
    Anyway, the point is, that you need to know from memory, that you need a pawn on h4 that can give check in case the king gets to g6.
    Since the h4 pawn is not in place, I would think if it is worth "working" on this tactic. Shall I move h2-h4?
    This is what I mean, when I say, that all my tactic training improved also my positional play. I see the tactics BEFORE they are there.
    And: If you would not know about the "3 pieces playing tandem"-tactic, lets be honest: you would never even start thinking about playing h2-h4.
    Would I play h2-h4?
    No. It leaves an ugly "hole" on g4, where black can later place a knight or bishop. At the same time, h2-h4 is too slow. Black simply castles away.
    But the very point is: would you have thought about an ugly move such as h2-h4 if you would not know about a tactic where a pawn on h4 moves to h5 and gives check to a king on g6?

    If there is nothing in your memory, chances are high you would not consider h2-h4.

    I would never have the time to think about the concepts tempo gave here, but probably I know them more or less subconsciously. Here comes the example with reading. There is an analogy:

    In this tactic here, we have 3 main actors:
    - a bishop on the diagonal a2-g8
    - a knight on f3
    - a queen on the diagonal d1-h5
    We also have an "enviroment":
    - black did not castle.
    - typical actors around the black king.

    And now we compare it with reading words: the first and last letters are the 2 "main actors" (in the chess tactic we had 3 main actors).
    The environment needs to be typical. It is important that they are somewhere around, but there place is not so important, as long as they are there.
    If you know how the words shold approximately look like, you can read them. There are billions possibilities how to write the following text, as there are billions of possibilities in chess. But if the main actors in there typical environment is in place, you have no trouble in reading this:

    Aicdonrcg to rcaeersh at an Eginlsh uesrtiivny, the oerdr of lteerts in a wrod is not irmapnott for the txet to be rbaadlee, the olny rrieeemnqut would be taht the fisrt and lsat lteter of each word are kept in the right plcae. The rest can be a cpomtlee mess and sitll be raebdale. The reaeescrrhs bvileee this is bacseue we do not raed erevy ltteer iulnvialdidy but rahetr the word as a wolhe.

    --> how did you learn to read this? In your memory you have stored typical words. Tempo would say "concepts" of a words.
    You memorized these words, you have an idea about the word, and if you read this text, you compare this ideal idea (concept) with it. And suddenly you are able to spot tactics "immediatly". Same as you can read the scrambled text almost as fast as if it was written correctly. That is the smoking gun that points to the "supertrick".

  2. P.S.:
    Just in case you are not a native english reader (=english beeing a foreign language to you), here the original text:

    According to research at an English university, the order of letters in a word is not important for the text to be readable, the only requirement would be that the first and last letter of each word are kept in the right place. The rest can be a complete mess and still be readable. The researchers believe this is because we do not read every letter individually but rather the word as a whole.

    The more familiar the words are, the better you had been able to identify them. You might have had trouble with "researchers" or "according" since these are rather seldom words.
    Same in chess - the more typical tactics are broadly known, but the masters do know more concepts. (or as I would put it: have stored more "ideal" puzzles in there memory. The more you know, the more you can see/read).

  3. @Munich,

    h4 was not in my system otherwise I would have placed a pawn there (I concocted the position myself).

    Creating what is not in your memory OTB costs simply to much time.

    I like your formulation with the actors. Should be usable for more concepts.

  4. I solved the problem with the spelling checker. It was a new version of Firefox with the spelling checker on.

  5. @Munich,

    You seem to talk about a higher aspect of concepts: the essence. A concept is still a little narrative which takes time to read in your mind word by word. But the essence of a concept is the translation of the narrative into one token. Like a seed which contains the essence of a tree (new age music in the background). How do we create essences of concepts? Maybe that's the real transferproblem.

    Maybe the essence is the endpoint of a cue and when you start to pull it, it retrieves the memory.

  6. Me Munich: My son Joe (6 years old) reads words often like this: "d-o-g" --> "dog!"
    He had to spell each letter, before he found dog.
    We however, dont de-code each letter and set it together in one word. We "dcoede dfifeernt".
    Same in chess. I am sure.
    We once had to find these concepts you are writing about.
    Over time, after we repeated again and again similar positions with same or similar concepts, we were able to use these concepts very fast.
    With a speed that is comparable with the speed of readings words (=incredible fast).

    Speed readers train especially words that are quite frequent but cause them trouble (that slow them down).
    They especially train these high frequency words.
    I dont know how they find out which words they need to train. But in chess this is easy:
    you need to train easy puzzles, that are easy to solve, but not easy for you.
    It is more rewarding to eliminate your weak spots in the easy range. Because these tactics are more likely to come up. Also: complex puzzles often consist of 2 or 3 easy concepts.
    If you know the easy ones, you have better chances to decode the complex once.
    In reading we also have fractions of words that we learned "-ing" "-et" "-og" "-en" "-an" "-ark"... so we only need to decode parts of difficult words. Even a word such as "desoxyrybonuclein acid" contains "des-" and "-in".
    an experienced reader has less trouble to decode such a word than my son, who currently would not have a chance.
    Reading has a lot to do with memorizing. and so does chess. But it is not memorizing everything. If somebody would give you 1000 well assorted puzzles - probably these are sufficient to improve your tactics tremendously.
    There are puzzles worth to be memorized, while others are of less "quality".
    I dont have an assorted set. But I repeat and train especially puzzles that cause me trouble (=easz puzzles I did wrong, or which took me very long).
    And even if I had an assorted set - it would only make sence to train those I cant do very fast.