Saturday, January 16, 2016

Three types of vision

Yesterday I listened to a video of Dan Heisman about vision.
He discriminates between three types of vision.

  • Board vision
  • Tactical vision
  • Visualization
Board vision.
This encompasses the roles and tasks of the pieces, and the squares they are covering. For simplicity I will call the latter the "aura" of the pieces. Investigating this at CT lead to the conclusion that I'm explicit weak at visualizing the aura of the pieces and I tend to overlook overworked pieces.

Tactical vision.
My tactical vision is well developed after a zillion+1 exercises. The problem is that somehow the cues that should trigger the recognition do not fire immediately, but only after some period of trial and error. When triggered, the "aha" feeling is immanent. The habit of  uncontrolled trial and error should be erased.

Heisman uses this term for seeing how the position looks like in the future, when pieces have moved. The main problem here is that information of the pieces on the board contradicts with the information of their future position, causing interference. Without a good board vision, visualization of future positions is impossible.

Embedded chess knowledge.
What is missing here is the embedding of chess knowledge. When you see a pinned piece, you should start looking for how to attack the pinned piece. If not there, you should look if you can target the defenders of the pinned piece in stead. In the mean time you have to look out for the standard methods of unpinning. Most of this tactical knowledge is quite simple and well known. Yet it takes ages before it rises up in the thinking mind.  Somehow these reactions must become automatic. It's no rocket science, there is no need to invent the same wheel over and over again.

Isolated training.
The examples that Dan Heisman gave in his video were isolated exercises, specific designed for targeting isolated skills, part of the different types of vision. I don't think that is a very good idea, although it helps to understand what he is talking about. You become good in what you train, no doubt about that. But there is a whole bunch of skills that you should learn this way. We might underestimate the magic of the unconscious. When it can add the sound of the motor , the speed of the car and the distance to the next bend in the road to the art of shifting gears, without our thinking telling it how to do it, it might need more space to organize itself. I trained Troyis as an isolated exercise, and now I can move the knights around in my head like a grandmaster. But if you look at the contribution of this isolated skill to the ability of winning an OTB game or solving a tactical puzzle, its effect is marginal.

Renko's ICT 2.
I have solved about 200 puzzles of the levels "intermediate" and "advanced". I'm sure they are good to train your board vision. But in stead of doing all 1400 or so problems (maybe later), I changed to the "master level" exercises. That is way out of my comfort zone, which is good for my chess karma, I have been told, but it gives a good idea how the different types of vision work together.
The following diagram is a beautiful example. Try to visualize the whole tree of analysis. It is a forced line, meaning that each move of yours is a check. You will find out that that knowledge isn't of much help. The end result is mate in all lines.

White to move

r1bq3r/pppp3p/6kn/3NQ1b1/4PR2/8/PPP3PP/5RK1 w - - 0 1


  1. The puzzle is the same "situation" as the puzzle here:
    And the method related to the situation then is appliable again.

    Situation: its almost mate but...
    The number of defender is ~3 and the number of attacker is ~3

    Method: reduce the number of defender and/or increase the number of
    attacker (in both puzzles). With Rf6+ you interfere with the Qd8 and you increase the attacking powers of the Rf1.
    After Rf6 the number of attackers is almost 4 and the number of defenders is almost 2.

    For a sucessful attack you usuually need +1

    This is an example where "positional thinking" helps a lot in tactics. I did see Rf6 instantly and im shure Munich would see it even a little faster. I would play Rf6 even without lots of calculation, its obvious.

  2. The move Rf6+ is pretty evident of course. But that is not what the position is about. Essentially you are gambling if you play it without calculation. The question is: what happens in your mind when you calculate all possible lines of the tree of analysis until quiescence? What does it tell you about the 3 types of vision?

    1. I have finally solved this puzzle with the use of chessboard (at computer screen to save all the variations).

      It looks like - it is VERY difficult puzzle to visualize to me! I was sure it is mate no later then in 4-5 moves. And after I realised how many ways of escape the Black has... I was astonished!

      The only cue to me is the possible mate patterns (the configuration of piece that may give mate). If I have had about 30 seconds to calculate till mate - I would fail without any doubts!

      I would never think Black King may be mated at move 6! This type of position is mate in 4-5 moves (no more than that!).

      BTW. Do you know that the move 2.g4+ also wins? You can look at the variation below and its final position (FEN format). For me it is a really nice artistic pleasure of chess (giving mate as a beautiful pieces coordination).

      1... Kh5 2. g4+ 2 Nxg4 3. Nf4+ Kh4 4. Ng2+ Kh5 5. Qxg5+ Kxg5 6. R1f5#
      r1bq3r/pppp3p/5R2/5Rk1/4P1n1/8/PPP3NP/6K1 b - - 0 6

    2. It's a puzzle at master level, and we are given 5 minutes to solve it. That might be a bit reassuring.

      It brings to light any flaws in our visualization process. It took me ages to solve it, and I had a few terrible oversights. The strange thing is, that after my discussion with Aox about a focus guiding thought process, the position looked much simpler all of a sudden. So there is definitely a role for that.

      I don't know what conclusions I can derive from this experiment yet. So far this vision stuff seems a pretty complicated matter.

    3. I thought it is a medium difficulty puzzle as ALL the moves are forced (checks). It seems easy - unless you have a very good visualisation skills - and can make simple sacrifices (theme of destroyng the guard). I have studied such finishes when I was younger, but it was ages ago! I have a notebook with notes - when the position was the same type and I had to drive the enemy King about 6-8 moves away from its original position. The mate final lasts about 10-12 moves!

      I am not sure if this may be some guide (help) to you, but I find it VERY helpful to visualise the positions after I solve them at the chessboard (with the use of moving pieces). After that I just look at specific variations and I can "see" why some moves are illegal and/or why the last move is a checkmate. Have you ever thought (or experienced) about that?

    4. That is exactly what I'm trying to do. After solving the puzzle I try so visualize (II) every line, not as a series of moves into the future, but as a geometrical pattern in the here and now. Hoping that my unconscious chess module picks up the hint.

  3. ehhhhh???
    I dont see no substancial difference between tactical and board vision, board vision is just "simple tactics" vision. While it is board vision to see that a Bb2 attacks a Nf6 if the black main diagonal is open its tactical vision to see that a Nb5 may fork the Ra8 and the Ke8 at c7. Such visions replace/simplify thinking/calculation because such "easy" facts are not seen by beginners.

    Visualisation is a independend skill. While "vision" is the awarenes what pieces do ( to some degree ) Visulaisation is more or less the awareness where the pieces will be = more or less memory

    Vision and Visualisation are needed to have a good, quick and flexible inner board which we use to step through the djungle of variations.
    There are some further skills like memory of the already calculated variation = "a sense of orientation" needed too if we dont want to run in circles ..
    Without a systematic thinking process its hard to calculate all variations till mate.. you might miss lines

    By the way im reading Aagaards Calculation Book where he is talking about vision and candiadate moves and so on too.. i hope he will give some trainig tips.. So far he suggest to do easy tactics and special(??) puzzles(??) to improve in vision.

    1. I worked out a thought process for long king hunts, and it definitely helped to simplify the position.

    2. I am sure you can think of writing the post related to "King hunt" - with the use of visualisation skills. I think I can help to make some interesting remarks (comments). Anyway - the more ways to attack the enemy King - the harder to me to decide which one is correct (or at least - which one gives the shortest and simplest way to mate)!

    3. Maybe I will devote a post to that. A thought process is highly personal, since the things that you have already automated are superfluous to include. Everybody has different things automated.

  4. Although I managed to work out all lines to mate, I still overlooked an "empty" square (f5) that was controlled by the e4 Pawn. I think it was because the Black Knight at h6 contests that square in the initial position, so I set up a mental condition that was I did not want to put the White Queen there. For all intents and purposes, I became "blind" to the Pawn's eventual value for controlling the "box" around the Black King and protecting the White Queen on f5. I tuned it out of the mental picture, just as I tuned out the two Black Rooks and the Black Queen's Bishop. Although it only made a couple of lines longer to get to mate, that is DISASTROUS, IMHO. If I cannot "see" the shortest mate sequence, then that greatly increases the possibility of missing critical lines, which could (should?) result in a loss. Back to "looking" for the salt mines!

    1. Tuning out elements of the mental picture is an important feat to simplify positions.

      I shouldn't be too worried about the shortest line when your own line leads to mate for sure too.

    2. I agree with both of you. Anyway I will show a small difference.

      If your line leads to mate FOR SURE - and it has not got ANY flaws (mistakes) it is equally good as the shortest way to mate (most often shown by the silicon monsters in a few seconds thinking).

      However if you DO NOT understand the shortest way to mate - it may show you the weaknesess area you should work on! That's because in some very sharp position - if you miss one of such called 'short moves' - your opponent may take the initiative - even at the cost of material. What happens next you probably know by your own games ;) :)

    3. Do you remember when CT pats you on the back with the words "Nxf5 is a good move, it wins the queen, but the computer has found a better move"? And after 6 tries and 6 pats on the back you finally have found the move that looses and you get penalized?

      For the sake of study, one should try to find the best move. But OTB the advice "if you have found a good move, find a better one" leads to time trouble.

  5. I repeat your statement above, just for proximity:

    Board vision.

    This encompasses the roles and tasks of the pieces, and the squares they are covering. For simplicity I will call the latter the "aura" of the pieces. Investigating this at CT lead to the conclusion that I'm explicit weak at visualizing the aura of the pieces and I tend to overlook overworked pieces.

    The nomenclature used is irrelevant for the purpose of improving our playing skills. However, it IS useful for discussions of how to go about improving. It seems to me that your "aura" seems to be a combination of the function motif and the geometrical motif. The more elemental "roles" are at a lower, simpler level from the motifs.

    It is my opinion that we do NOT have direct control of any conceptual "chunking" process that may (or may NOT) exist in our subconscious. I think that attempting to consciously "chunk" our expanding knowledge is a waste of time; the brain must do this automagically. With that caveat, I do think that "chunking" CAN occur, just not necessarily, as a result of accessible knowledge of specific relationships. I think this IMPLICIT "chunking" is an integral part of chess "intuition." In turn, I think that "intuition" is an important part of visualization. It allows for the capability to "sniff out" the line(s) of play that are most promising, given the current position and the potential hidden within it. Words are cheaper by the dozens, thereby being worth less than a picture.

    An example, extracted from Understanding Chess Tactics by Martin Weteschnick, pg 122:

    5rk1/1ppR2pp/2n1P3/p4pN1/5Qn1/q5P1/P3PPBP/6K1 w - - 0 1

    (Timman vs Short, Tilburg 1990)

    Does anyone believe that Nigel Short is unfamiliar with standard chess tactical themes/devices?!? Not ME!

    (An aside: I put this position into Fritz 11, running it through Stockfish and Komodo. Although the first move remained the same as played in the game, both engines came up with a different 2nd move for Black than what Short actually played. Since the engine evaluation is over 12+- for White, I think it safe to assume that Black is lost in all variations, no matter what he plays.)

    I think if you look at the overall picture, you will notice that there is a potential check by the White Queen on the c4 square. In conjunction with the White Knight on g5, what standard tactical theme/device comes to mind in conjunction with the Black King on g8? Use mentalization to (essentially) wipe off the rest of the White and Black pieces from the board. What do you "see"?

    I think of this as the “Aha!” conceptual “chunk” that triggers the thinking toward the right direction. On the other hand, if you don’t “see” that configuration of pieces as connected in a specific (typical) relationship, then I think that it will be very hard to play like Timman!

    My point is that if you “see” the relevant “trigger” pattern/concept/chunk/relationship/motif/tactical theme/whatever, then it comes with appropriate “built-in” moves. This simplifies the visualization required to “see” the lines of play that are required to “solve” the problem and play in accordance with the demands of the position.

    Or, it could be that I'm just "seeing" a mirage.

    1. Once you told me to pay attention to Qc4+, the position was no longer difficult, and it was easy to work out the moves. What's needed is an "inner Robert Coble" that tells us to look at Qc4+ first.

    2. Yeah, that's one of the greatest triggers for me! It is the concept of "magic pieces" - you should mentally make any of the "unnecessary" piece to dissapear and... voila! The hard thing is to know WHICH piece(s) should be removed (and why), but it may help a lot. At least it is VERY helpful to me.

      BTW. I full agree - we should have installed "inner Robert Coble" module in our minds ;) :)

    3. May I have one of those "inner Robert Coble" modules too? ;-)

  6. Another example from Understanding Chess Tactics by Martin Weteschnick, pg 123. Quoting Master Weteschnik:

    "A game by Alekhine will be the next illustration of the pattern. In a simultaneous display in Breslau in 1933 he found a clever solution for the following position."

    r2br2k/pbqn2p1/1p2Qn1p/2p5/3P1P1N/P1PB4/1P1B2PP/3R1RK1 w - - 0 1

    I note that White has two Pawns for the missing White Knight, so White is slightly behind in static material count. The Black King is in a pickle: he has NO LEGAL MOVES! That immediately grabbed my attention! So, who cares about the material imbalance anyway?!? In every patzer's dream, "always check: it might be mate!" We are going to play like Alekhine!!

    Let's again conceptualize by removing "extraneous" pieces from the board (mentalization). Only three White pieces remain (Qe6, Be3, Nh4) and two Black pieces remain (Re8, Nf6), along with our selected "target," the Black King. There is a function motif between the Black Knight and the Black Rook; the Black Knight must protect the Black Rook. The white squares around the Black King look very inviting, don't they?! Notice that the White Bishop at e3 has the function of keeping the "box" closed around the Black King. So, the primary White pieces are the White Queen and White Knight, working together against the Black King. That information alone should begin to "suggest" a possible approach. In short, the contours of the position dictate the direction of our investigation.

    Can you rearrange the White Queen and White Knight along with one or two of the "useful" Black pieces so that Black is checkmated?

    To do this, just move the pieces into the proper position for the checkmate to exist. Don't think in terms of "moves" or "variations" of moves - JUST SHUFFLE THE FURNITURE! Does this now give some "visualization" of the concept required to checkmate Black? If so, does the concept now guide you with some "idea" of the sequence of moves required to accomplish this rearrangement of the furniture?

    This example did NOT cause the "Aha!' feeling to occur immediately (for me), even though I knew (from the context of the book) what was supposed to be "visible." WHY NOT, I wonder? Is it because the vast majority of examples/problems that I have seen on this mating theme utilize "stock" combination of pieces (Queen and Knight) in almost identical positions, with the result that I do not "see" the abstract relationship properly? Have I not hammered the abstract concept into my subconscious mind so that it can "POP" up the right "idea," no matter what actual position is directly in front of me? I must think about why I did not "see" this problem of NOT "seeing" immediately before now.

    1. Maybe there is no aha feeling since there is no mate? It is about gaining wood that black has to give in order to prevent mate.

    2. There is a possible mating sequence, but (as you point out) it can be disrupted into "merely" gaining a lot of material. I think this is one of the areas that I find so irritating about problems, especially composed problems. If I "see" that there is a mating sequence as one of the alternatives, and that the opponent can avoid the mate at the cost of a horrendous loss of material, I "assume" that being down in material is a lesser "evil" than being checkmated. "Where there is life, there is hope!" So, I opt for that alternative as the "solution" because I possess sufficient technique to kick butt whenever I'm way ahead in material and the opponent has no compensating attack on my King. NM Dan Heisman calls such positions "GTS": Go To Sleep positions. What happens? The training site scores "WRONG ANSWER!" with a frustrating drop in rating. (This is why I stopped paying too much attention to the rating when training for skill.) Perhaps I learned the wrong lesson many years ago. It may take longer, but it is absolutely certain that I will win if way ahead in material AND the opponent has no compensation. The lesson I learned (the hard way) is that the only thing that matters at the end of the game is the score: did you win, lose or draw? It doesn't matter what brilliant ideas you had during the game, what brilliant combinations you saw, or what the computer says you woulda/coulda/shoulda done at a particular point that would have given you the advantage.

  7. There are a few usable idea's in the comments. But I have difficulty to express myself (which happens very seldom) because I have the feeling that I am walking knee deep in a sticky goo of vague terminology. Time to formulate some definitions.

  8. This reminds me of an example in the first few pages of the Art of Attack. I will have to see if it is the same one. This will take a deep think but geometrically some thoughts. My rooks are doubled -reloader? . bishop in line- Pin? unlikely-Knight might be able to deliver finally checkmate as escape squares same color but as Robert mentions there needs to be some furniture to be reshuffled and new escapes squares may open.