Sunday, March 17, 2013

Seeing everything as new

Say, there is a big tree in my garden. It casts a lot of shadow, so it makes my house ïnside very dark. At one day, I decide to chop it down because I got so annoyed to have to put on the lights by day, even in the summer. After a good night of sleep, I have forgotten that I did chop the tree and why I did it. When I wake up, I look in the garden and I start to rave: "which idiot has chopped my tree while I was asleep? When I now sit in the garden it will be too hot for me since there is no shadow!"

This happens when you see everything always as "new". When you don't see the relation between your actions and their consequenses. This may look highly unreasonable and highly improbable, yet this is exactly what happens in a chess game. Today my opponent made an unexpected knight move. I was in a bad position, and I was looking for eternal check. I looked at the consequences how his knightmove and I concluded that I could still put my rooks where I wanted them. After a few moves I gave a check with my rook. In stead of moving his king away, he simply put his knight in between. I had totally missed that move. I had seen the unexpected knight move. I had seen some consequences, but not all. You may argue that in this case, it was not me but my opponent who made the move. It was not you, but it was your neighbour who chopped the tree. But that doesn't make a difference. Somehow, we are blind for the consequences of actions, when it comes to chess. We see the action, but don't see the consequences.

This, and only this, is the main reason why we are so bad in chess. We see the actions, but are consequence blind.We see everything "as new". As if the consequence is there without a cause. No matter if we were the cause or our opponent.

This is what a brainscan of the amateur reveals. We see always everything as a new position. Every move. No matter if it was your move or your opponents.


  1. This was a board vision problem, you where not aware of the possibilitys of that knight. Your attention was somewhere else and / or you did not want o see that move. (subconcious hope chess)
    One of our stronger player told me, that he is looking every time at every possible move of himself and that he is looking at every possible response of his opponent to his planed move too.
    That is a blundercheck and a 0board vision exercise and could be done at every standard problem
    I recognised at the Fritz Attack training, that i had problems to see some special moves too like bishops backwards..

  2. It indicates a fundamental flaw in the way we perceive the world. We separate the object from it's functions.
    Say your daddy is a teacher. It is as sometimes you see a guy and forget it is your father, and at other times we forget that our daddy is a teacher. If we don't see the guy at all, we call it a blunder.

    If you look at your opponents f7,g7,h7 during the openingsfase of the game, there never is a moment you are inclined to consider to put a bishop at g6. We can't see the pawn without its function. That is how close the object must be related to its functions.

    By making diagrams from positions, I describe the functions of the pieces by means of arrows and colored squares. It helps. But there is still a long way to go. In a way it is a low level vision problem, probably.

    I don't know what the threshold is of the brain for connecting an object to its function. I mean, when you learned to drive a car, a car that comes from the right has a different function than a car that approaches you from the left. You perceive them differently, and it is no problem to ignore cars from the left while you are aware of cars from the right. You learned this without much effort. You learned it conscious, but it has become automatic with very little exercise.

    If I see a new position, at first I see the objects and the most familiar functions. Like the pawnfront standing side by side. Then it takes time to see more and more functions of the pieces. Some functions I don't see at all.

    If the conditions are right, we assimilate functions from objects easy, with little effort. As with learning how to drive a car. But in other cases, the mind seems very reluctant to add new functions to objects, even after years of exercise. I wonder what makes the difference.

  3. Fundamental flaws are uninteresting, they cant be changed so why thinking about it? My grand grand grand onkel said: "The philosophers have only interpreted the world, in various ways. The point, however, is to change it."

    We have a process of analysis and of couse we need to see the pieces first, they are the carrier of the function.

    To me the "whole problem" is to see "what is going on", "what is the position about","what is the idea, the essence" ( and that "~ quick" [ so we have easy access to that skill at the end of our calculations to enable a good evaluation] ). That is ~ : to "create" your diagrams (if they "realy" represent the "essence" of the position) "quick".

    We programmer know: There are 2 general methods to solve a problem ( metaphysics :) : top down and bottom up.

    In chess Top down is this:

    and bottom up is this:
    The idea of "contacts" is common, for example in Averbakh's Books.

    I try to combine both methods
    Top Down: 1) Scan Material balance, 2) King safty ... X) Not protected pieces(= a special weakness)...
    Every Scan gives "general information" but these scans are in a way bottom up too because they are a training to see/recognise (a) function(s) of each piece.
    The big art is to filter these informations and to extract the (an) idea.

    But back to your concrete problem: you where not aware of a piece and its "function".

    In a way (not 100%) Cor van Wijgerden does adress that in his exercise: . You see a position for a few seconds, the pieces vanish and you have to find the shortest save! path to give check. So you have to remember the pieces AND! their conrolled squares.

    Our STM is to small to keep "everything" in it, we need to chunk. A! chunk is "the piece at its square" -> controlling other squares ( and be able to protect against check ). Most of the Board Vision Exercises are trying to create just this type of "Piece-Square-Function chunks": Fritz Attack, Defence and Check Training, and ( weaker ) my Mobility Trainer.

    So: wellcome to one of my ( minor important ) construction areas :)

  4. I've read through all of your posts from the beginning. Wonderful food for thought! I certainly hope you have not gone "dark."

    Crazy Bob

  5. Hi tempo,

    this is the latest entry of your blog, so I write here, even though not to this specific post here.

    I found new (old) ways to improve:
    study the best opening moves and try to understand why they are so strong, and what typical positions you achieve with these moves.
    I went this way recently, and it boosted my skill. Cant tell you how much, but I can tell you how I know:
    I usually was a bad blitz player. There is a guy in our chess club who is the opposite: an excellent blitz player. even though he is just above 1700 Fide elo, he used to beat me regualarly, with a positive score against me. I won maybe every 3rd or 4th game, and there were days I went home not having a single win against him.
    He is also scoring well against the other "big guys" with ratings above 2000.
    Now after my opening studies I started to be able to blitz faster. I bash out the moves and usually am doing well and fast in the early middle game. That is enough to score much better now. Against the strong guys and especially against this 1700-guy. Yesterday I beat him 6:3, and it was already my 4th or 5th week I got the better of him.
    He told me yesterday that he got the impression I improved a lot in blitz.

    What did I do? I studied openings.
    My assessment (to guide you for the start): the best opening for white might be 1. Nf3 and if black does not respond 1...d5 or 1...d6 (moving the d-pawn), the 2nd move is 2.c4 (otherwise it is 2.d4).
    I opt for a sicilian maroczy-bind or a kings-indian set-up (with h2-h3 to avoid Bg4) and the statistics for white are overwhelming, even though a computer program might not see anything "better" about these "best" moves - well, the thing is, that it seems very easy to play white while it is very difficult not to mess up the black position.
    It helped me, it made me stronger, no doubt about it.
    It is surprising because "to study openings" is not you would consider to be any good. But it is not "to study an opening" - but to find the statistically best way how to play the opening. Looking for the best moves and set-ups. It saves you a lot of time, it seem to work on your positional play, and that is why it makes you stronger.
    On the other hand - maybe my previous training show an effect with half a year delay. Last thing I did before the quest of best opening moves was: a lot of board vision training.
    Ah, by the way: I improved meanwhile approx. 300 Fide elo points within 2-3 years. (I started with ~1800 or below, and am now going to be at ~2100 or above, I can show you my rating (link) in a private email cause I like to keep my name anonymous).

  6. Somehow stuck, getting the feeling that there is no way to improve?

    I came back to "learning on autopilot".
    You said, learning on autopilot is not possible.
    However, I wonder why it is then possible to improve in like Fritz attack board vision?
    I mean after we have achieved a flow of approximately 40 attacks per minute, why are we then able to improve to 50 attacks per minute, and if we keep trying, we probably can beat my record of 51 attacks per minute?

    What is it, that drives us further in Fritz attack training?

    There where two issues (maybe even 3 issues) how to overcome a plateau:
    a) time pressure and haste
    b) not falling into the trap of the "o.k. plateau" (= being content and happy of what I can achieve)
    c) maybe (?) an "aim"/"target".
    World records get beaten just because we know level we need to achieve to a new world record.
    But this might be a sub-point of point b), the "o.k. plateau".

    I started training again, and my aim is to become faster.
    I improved most with my method of high volume/ easy puzzles.
    I created a Spaced Repetition (SR) set of "mate in 2" in the CT Blitz range 1200-1600.
    Puzzles I did slow or even failed will get served more often, puzzles I can already do very fast gets repeated far less. Target time for the SR set is set to 5 sec per puzzle.
    If I manage to get faster, I should be able to improve in tactics.

    Such a trainig was the first training I did, and it improved me most. I am hence returning to it.

  7. o.k. probably known anyway, but maybe forgotten:

    We do not improve by a lot of thinking.
    A hard difficult puzzle, solved within 20 minutes involved a lot of thinking. This thinking is worth nothing. It does not matter that we calculated a lot of side variations.
    Our calculation does not improve.
    If we found the right sequence, we might remember this sequence a bit better. If we fail we might remember it, too. It doesnt matter if we even got it right.
    What matters is to remember the piece of information that we can learn from that.

    What is this piece of information?
    A geometrical pattern? Possibly, but not neccessarily.
    Does tagging help? I guess so. It gives us the chance to understand the nature of the tactic better.
    But that is probably not the only thing needed. It might help, though.
    What else, if not our thinking well improve us?

    I cant really think of much else than memory. We need to fill our memory with lots of known tactics.
    When we spot them again, we need to be able to know the answer imediatly - without thinking.
    We need to do it on autopilot.
    The answer can be memorized, we probably dont even need to understand it besides that it helps our confidence that we understood the sequence once.

    This can be the only key: putting things into memory, and training to retrieve it from their fast.
    Every thought of calculation we did during our training is probably lost. It wont improve us. Thinking is useless, it is all about knowing.

    This is what improves us in the Fritz attack training. We know where pieces could be hidden. We apply the chess rules very fast. We really try to apply them fast: equal pieces can take each other:
    For instance bishop takes bishp, knight takes knight. We look out for that pretty fast. We dont have any understanding in the position. Nevertheless, after I did some boardvision training, I improved in "backward move required" tactics. In "rooks can move sideways" tactics. In "bishop hidden on a7 hits f2+" tactics.
    Boardvision improved us in hanging piece tactics.
    And I know now why: We memorize where else pieces could stand. Piece of information retrieved very quickly.

    Thinking and calculating a lot is not needed to make us better. Instead: Pure memorization of typical sequences help us next time to go through this typical sequence in our next real game. That is why it is so easy for us to follow patterns: we follow the typical squence and only check if it works. We have little difficulty in visualizing the sequence. With all this memorized stuff we are able to follow our "own footsteps in the snow" (our memorized move sequences).

    So what training is needed?
    I believe we need to be able to solve 2-movers very fast. In 3 seconds. Simply memorize the sequence. No need to understand it.

    We really need to retrieve it from memory that fast. Give the solution with no thinking. Just give it within 3 seconds if it is a 2 mover.