Saturday, January 02, 2016

It should start with vision

We have somewhat of a terminology problem here. I'm hesitant to use the terminology of others, like Lasker's. I once used the term "function" for a piece whose task it was to be captured, and Robert rapped me on the knuckles for that ;)
I'm using terms in my usual sloppy way and don't bother too much about the definition of other people. So please bear with me.

A typical solution of a problem by me goes this way:

  • Trial and error
  • Think
  • POP 1 the popping up of motif 1
  • Trial and error
  • Think
  • POP 2 the popping up of motif2
  • Trial and error
  • Think
  • POP 3 the popping up of motif3
  • Thought process about stitching the motifs together to a combination

What it should be, though:
  • POP 1 the popping up of motif 1
  • POP 2 the popping up of motif2
  • POP 3 the popping up of motif3
  • Thought process about stitching the motifs together to a combination

 The trial and error/think stage before every POP isn't necessary, and cost an incredible amount of time and energy. The POPs are not the result of the trial and error/think stage. They happen despite of it.

In stead, you should scan the board while looking for the function of the pieces. That is the goal to have in mind while scanning. When yo try that, you will find that our vision is too weak. The visual cues who trigger the retrieval of the patterns are too weak to fire easy. That is what our training should be about: reinforcement of our visual cues. We must perform cue-building, so to speak.

We talked about visual cues and vision before, but I don't think another example will be too much. Look at the following diagram:

White to move

 This problem ended up in my fail list. I was troubled by the amount of possibilities. I hope you are too, otherwise it is difficult to understand what I am talking about.

First I analysed the position until I had a clear understanding of it. Only then, cue building can begin. It seems that high rated chess players use their brain lobes that are dedicated to motor skills (look for motor skills in my blog). An important technique for training motor skills is visualisation. And that is exactly what I try to do when training. The most important cue in this position is the function of the black knight. Every time ANKI pops up this diagram, I visualize the function of the black knight. This morning, the function popped up all by itself. And it made the position look so simple, that I even can't imagine that I had ever problems with it. Once the popping up becomes automatic, it is very easy to infer the winning move without any calculation at all.


  1. Okay, I'll go first (as a guinea pig).

    The FIRST thing that I look for is King safety. White is to move, so I look to "see" how safe Black's King is. It is IMMOBILE — this is a MAJOR cue for an all-out assault, throwing everything available against the Black King!

    Is there a possibility of attacking that immobile King? Yes, there is a White Queen and White Knight threatening checkmate on h7.

    What Black piece(s) prevent the checkmate? The Black Knight on f6, combined with the Black King on h8.

    Is there another White piece that can be used for a direct attack on the Black King? No.

    Is there a White piece that can change the function of the Black Knight? YES! The White Bishop can pin it against the King, removing its defensive function.

    1. Be5 comes immediately to mind.

    Is there ANY Black piece(s) which can restore the defensive function of the Black Knight by (1) protecting h7; or (2) relieving the pin on the Black Knight? No.

    Problem solved: the Black King will be checkmated next move.

    I actually did not "see" the White Rook hanging on d2 until after I was satisfied that Black is going to be checkmated.

    If a checkmate is possible, it does not matter how much "wood" is thrown on the fire. Anything and everything other than checkmate can be ignored.

    The "solution" came to me in 10-15 seconds. Not instantaneously, but still fairly rapidly.

    I note (without any "knuckle rapping") that you VISUALIZED the function of the Black Knight, and that made the solution process look simple. This is the very essence of my insight obtained from re-examining Dr. Lasker's ideas about motifs as cues. I do NOT take any credit for what has been there for so long, right in front of my face (so to speak) and I did NOT "SEE" it.

    Our views on the importance of VISION seem to be converging.

    1. @Robert

      Your posts are VERY inspiring to me! It is clearly visible you have studied quite much material from different sources! (just looking at you as you quote is amazing).

      You missed just ONE small detail. However it is quite important one, because you probably omit it due to "simplicity". Actually it may be a tiny gap in your knowledge or understanding.

      Where is the missing part (an error) located? Look closely about the Black's knight function.

      BTW. You have just triggered my internal "pop up" idea! :). Thank you very much!

    2. @Tomasz: I confess - I am "blind" to the missing part; I do not "see" any additional function of the Black Knight that is important. Please help me to understand, and thereby, (hopefully!) to learn and improve.

    3. @Robert

      There is probably a blind spot around you ;) :). No problem my chess friend - I will help you.

      The Black Knight fullfills the additional function you COMPLETELY missed. However it is a normal case as you do not know (or has not learnt yet) the magical function (I will write about it soon).

      The missing part? Here you go!

      "Is there ANY Black piece(s) which can restore the defensive function of the Black Knight by (1) protecting h7; or (2) relieving the pin on the Black Knight? No."

      [my answer]
      Black Knight prevents white from checkmate in 1 (Qf7#). However we have an additional function hidden (!). It is the function of Black Rook and White Knight. There is a defence against checkmate in one (at h7 square) - Black Rook can go to f7 square and DEFEND this key square! However now we can see tha White Knight additional function* - it can take the Rook and give checkmate!

      BTW. The 'additional function' of Black Knight* was a deep hidden trap! I wanted you to force your mind to change the colour to White Knight. Anyway I hope you had some fun and now you can be more concious about additional defence... and additional function of attacking piece(s) - especially around squeezed King!

    4. @Tomasz: I think I understand my problem. I did not approach it as a "White to move and mate in 2 moves" problem. All I saw was "White to move" under the problem. My attention was focused on the "idea" to be "seen" and the subsequent first move. I did "see" that the Black Rook could also "defend" from f7, but it was a useless defense of h7, since 2. Nxf7 is also mate.

      Sloppiness on my part in writing everything out, I'm afraid. In the current problem, it does not change anything (Black still loses), but it is a very valid point to make: EVERYTHING must be "seen!" Thank you for pointing that option out explicitly!

      As part of my visualization of opportunities, I use a "box" around the King. Think of it as a 3x3 square box, with the King position always sitting on the center square. I use this box regardless of the actual board squares. For example, if the Black King (as in this problem) is sitting on h8, FIVE of the "box" squares are off the board! I do this so that I can "look" to see what constraints exist on the King's movement. The individual "squares" in the box can be occupied by pieces of the same color as the King, can be "attacked" by the opponent, are can be inaccessible because they are off the board. The idea is to "see" the restrictions on the King's movement, regardless. For instance, the "back rank" mate uses (usually) three squares that are off the board, combined with three squares on the board that are either occupied by Pawns or pieces of the same color as the King, or (more difficult) are controlled by the opponent's Pawns or pieces. I guess you could call it an abstraction of the limitations on the King's movements. I learned this "box" idea from learning how to visualize all of the notes on a 5-string bass guitar in a particular key of music. If you apply the "box" pattern, you can play any key on just 3 strings.

      Yeah, I know: I ain't "right!"

    5. Great insight! You probably know much more than you can express (at least verbally). Of course your approach is correct - I have no doubts about it. Anyway if Ng5 could NOT make any moves (i.e. if it could have been pinned) then Rf7 allows to prolong the defence.

      The box idea of 3x3 related to the King is a SUPERB one. I want to work on this idea with my own "magical" idea ;). The better you understand the idea of "cutting off" (restricting) the King, the more efficient you can check if there is a reasonable defence.

  2. We agree on the importance indeed. I don't know whether visualization of the function of the pieces is the best way to train for reinforced vision, but it looks logical, it seems to work, and I can't think of anything better just now.

  3. PART I:

    One of the things that I think is also important for "seeing" is the geometrical motif. I know this sounds a little weird, but I seriously have to force myself to look along each "leg" of a line-moving piece all the way to the edge of the board, ignoring any obstacles (same-side and opponent's pieces) in order to "see" what is happening with that type of piece. I do the same thing with the Knights, ignoring any obstacles that might occupy or control potential Knight squares one, two or even three moves in the future.

    Since I began forcing myself to "see" FIRST, rather than to "think" first, I seem to have a much better "vision" of the position, and the possible ideas POP quite suddenly. It does focus the attention on the important factors in a position in a way that previous attempts to look directly for tactical thematic patterns (pattern recognition?) did not.

    A most important "cue" is the opponent's last move. A lot of times, I will try to find the game from which a tactical problem was derived. If I can find it, I try to set up the position by backing up one move from the tactical blow that begins the "solution" of the problem. If I can't find the game, then I spend some time, trying to imagine what the immediately preceding move must have been. That approach seems to yield good insights into the problem of NOT "seeing."

    FM Martin Weteschnik, in his marvelous book Understanding Chess Tactics, pg 180, gives the following position from Petrosian-Bronstein, Amsterdam 1956:

    1rb2r1k/1p1n2q1/p2Q2p1/P2Npn1p/2P1N2P/6P1/1R3P2/1R6 w - - 0 36

    Black's last move was 35. ... Nf5. The greatest defensive player in the history of chess played a horrendous 36th move. What did he NOT "see?"

    (Full game is here: Not SEEING The Tallest 'Tree' In The Forest Of Variations

  4. PART II:

    This position is given in Weteschnik's discussion of the "status examination," which corresponds (roughly) to Beim's "examination of the position." I excerpt a little of the idea:

    "After you have learned about the elements of tactics [by which he is referring to his previous explanation of various tactical themes/devices], you need a method that will enable you to analyze a position as a whole. Otherwise you might arrive at faulty conclusions... This method has to be easy to understand and it also has to be easily applicable. Above all, the method should not turn into a tyrant. If it required you to go through every step every time it was your move, you would not use it. If you tried, you would lose on time. On the other hand, it makes sense to have some kind of checklist, ...

    The status examination provides you with similar checklists.


    Status Examination

    The status examination does exactly what its name says: it takes a close look at the status of each piece on the board. Principally, you have to look at two things with each piece. First, you have to find out its current status: whether it is attacked, defended, hanging, pinned, etc. Then you have to SEE this piece as an element of a PICTURE, which is related to other elements. Ask yourself how the status of this piece changes the status of other pieces.
    (My emphasis added.]

    Again, I think this is very close to the notion of "vision," applied step-by-step. It is my opinion that once the set of tactical themes/devices and mating attacks have been incorporated into the subconscious (those thousands and thousands of problems solved, maybe?), then Temposchlucker is correct: you have to SEE in order to play at a higher level. No sight, no "NOW!" If you don't "SEE" it (whatever "it" is), then you are flying blind, and blind pilots who survive their first flight are extremely rare. Perhaps it is partly intuition; I don't know. I do know that SEEING FIRST is more important than merely thinking in logical terms.

  5. GM Jonathan Rowson, in his chapter on Thinking in The Seven Deadly Chess Sins, has a subsection titled "Vision." He opens it with a quote from Richard James, author of The Complete Chess Addict and Chess Teacher: "Every time I see a kid making a mistake and ask him why he played that move his reply starts with the WORDS 'Well, I thought...'. 'Don't THINK' I reply, 'LOOK'. [My emphasis added.]

    "Many trees have fallen to explain the basis of this idea; which is that what distinguishes a strong player from a weak one in chess is 'VISION' (or visualization) and that this 'VISION' is based upon chess experience. The following claim is particularly significant: "The master doesn't calculate more than the expert. Rather, he SEES more, especially the more important things. — De Groot.

    . . .

    "Whereas relative beginners can only take in a piece or two at a time and very slowly grasp the ways in which the pieces are interacting [in the context of chess], the stronger player can quickly assimilate lots of different constellations of pieces and their role in the position as a whole.

    . . .

    "So the way a chess position is perceived differs markedly among players of different strengths. I have consulted lots of people on the matter of chess perception, asking them about the board in their head and whether it is a static image, whether it moves, has shape, is limited in size, etc. I have also asked about this image during chess thought-processes and, although the answers are varied, there is a common thread. This has not been a closed and careful experiment but I found a very interesting pattern which I will ultimately use to explain the limitations of "thinking" in chess. It comes down to this general claim: the stronger the player, the more abstract the VISUAL image.

    . . .

    "There are problems here of course, and these musings are not philosophically air-tight. For instance, as soon as you stop to think what you are thinking [meta-cognition], you are no longer thinking in the same way. But I wonder whether this VISION thing can really help us make sense of how we think. Maybe it's just a blind alley, as it were. Speaking of which, if such VISION is entirely abstract, is it the VISION of a blind man? Are we 'seeing' without our sense of sight? And if we don't see a VISUAL image, what is it that we LOOK for when we think in chess? What is guiding the search?

    "I had hoped that conversing with blind chess-players would help me to clarify this issue, but it seems that blind players "think" in a very similar manner to sighted chess-players. ... However, what struck me most from these conversations was a comment from Graham Lilley (Britain's strongest blind player) that first and foremost he 'SEES' an idea, usually as a familiar pattern, and only then checks it by means of analysis. He is also guided by his sense of value. Moreover, he claims to have NO VISUAL IMAGE whatsoever in his thinking processes. When he says that he first and foremost 'SEES' an idea, he means it in the sense of recognition rather than sight. He moves from one position to another on the basis of where his thoughts go, and as far as there is a direction to these thoughts, he says that he heads towards, as far as possible, 'what works'.

  6. My thought process in this position was very similar to Robert's...the first thing I looked at was the black king. In this position white's queen and knight next to the soon-to-be-dead king was a clear visual clue, and I immediately started looking for ways the two white pieces could create checkmate.

    There are in fact TWO mate threats here -- Nf7# and Qh7#. Once I saw those patterns, I tried them out out and saw the black knight is defending h7 and the black rook is defending f7. I wish I had seen Be5 right away, but I didn't. Instead I looked for ways to deal with the rook defending f7, perhaps because I saw that check (Nf7) first? After failing to find ways to do something to the black rook, I then moved to look at the rest of the position...I did ultimately find Be5, but I certainly could have found it faster.

    I then scanned the rest of the board to see if anything else was happening -- like whether black had his own mate threats (which would indicate how much time white has to deliver mate), and how easily white's other pieces could join the attack against the king. At this point I saw the white rook is hanging and that black threatens a mate in two...QxR then Qc1#/d1#/e1#...which is slow. The white rook is blocked from joining the attack by his own pawns (I did notice dxe5 brings the rook and white pawn closer to the action), and the white bishop can jump on the same diagonal as the black king in one move (Be5).

    At that point I saw the pin, allowing mate: 1. Be5 Rf7 (only defensive move) 2. Nxf7#

    Again, I could have solved this faster if I had applied reciprocal thinking to BOTH of the mate patterns instead of just the first one I saw.

  7. I find it interesting that two different people look at the position, quickly come to the same conclusion regarding the importance of the immobilized Black King, and then "SEE" the two different continuations, yet not in the same order. Is this a result of prior studies or a difference in vision? I don't really know.

    I think I "saw" the potential mate on h7, and then worked backward as I tried to figure out how to remove the defensive function of the defending Black Knight. Dan "saw" the same thing, yet it was the defense of f7 by the Black Rook that he focused on first. In neither case were either of us "wrong," we just followed a different order.

    I didn't even consider the potential back rank mate for Black (taking the White Rook with the Black Queen, followed by mate) until after I was satisfied that White was winning in both lines. Why not? Because the assault motif suggests that during an attack, the opponent is given no time to follow his own schemes. As long as White moves first (VERY IMPORTANT!) and maintains the initiative with forcing moves, he has no need to worry about a potential counter-attack. This is an obvious generalization, which is not necessarily applicable to any specific position, but it GENERALLY holds true. Like so many chess "principles," there is always a ceteris paribus (everything else being equal) caveat to consider. It is important to remember that in a specific position things are never totally equal.

    Anyone care to speculate as to WHY Dan and I followed two different lines of investigation? What did we each "see" that caused the difference in approach?

    1. The only reason I looked for other things in the position, like black's threats, was because I didn't immediately see the checkmate. Of course if I had found a forced sequence, there would be no need to look at anything else in the position other than black's defensive resources related to the mate. Also since Be5 is slow, you'd need to look for black's checks which could enable a counterattacking maneuver to defend.

      I think the only reason I followed the path I did was that I saw Nf7+ before I saw Qh7+. I used reciprocal thinking to deal with the defender of Nf7, and when I found nothing I should have then looked at ways to deal with the defender of Qh7...that was my mistake!

      I think it was complete chance, or perhaps a result of my training, that I noticed Nf7 before seeing Qh7. I've been doing a lot of work with my son on knights recently, so perhaps that was why. But you and I both saw the mate patterns and used the same technique of reciprocal thinking to solve the problem of the defender.

      As an aside, I believe there are only three ways for queen and knight to mate a lone king in the corner (two with queen delivering mate, one with knight delivering mate).

  8. Today I added looking for the geometrical motif to the training for the first time. It helped me to broaden my view. Hat-tip to Robert.