Wednesday, January 20, 2016

Thought process redux

I have experimented with the master level exercises of ICT 2, lately. One thing that became clear, is that we need an "inner Robert Coble" chess module, that tells us "Qc4+!!" (ok, I will stop joking about that). Aox (luckily) never got tired to tell us, that we need a thought process. He is right. Just sitting back and focussing out works somewhat, at least it prevents tunnel vision, but it is not enough. We need to interrogate the position somehow.

To some extend, a thought process is highly personal. If you never miss a pin, there is no need to add it to your thought process. It is already internalized.
I have done my share with thought processes. I  trained the systems of other guys, and I invented my own brand. Usually, after a few weeks or months, I forgot the exercise and it went into oblivion. Yet I have a habit to sweep the board for targets and attackers, which stems from those periods of training a thought process. In general, those training sessions were quite experimental, since there was little proof that it actually worked. The purpose of the training was not quite clear. Or actually it was: experimenting.

But now I have come to the same conclusion for the umpteenth time, the goal of the exercise becomes much clearer. So it is easier to fine tune the exercise.

I'm not going to adopt a system that is invented by somebody else, since that leads easy to an overkill. Trying to adopt too many things at the same time, is a recipe for failure, is my experience. Less is more, and if it works, it is easy to add new elements later. In the first stage, I want to automate the following steps.

  • Step 1. Scan for attackers.
  • Step 2. Scan for targets
  • Step 3. Scan for attacking squares.

The first step is usually done by me before the computer moves, at CT.
The second step usually doesn't take much longer.
The third step needs some explanation. Take a look at the following diagram:

White to move

 Attackers: Q, R, R
Targets: Q, K
Attacking squares are squares where an attacker can attack the target:
d8 -1
f8 -2
e6 +0
e8 -1
c6 -1

The standard reasoning is:
  • What is the most promising square. In this case: e6
  • Can you add an attacker or?
  • Can you annihilate a defender? 
1.Rd8+ deflects the black queen from e6
1.... Qxd8
Now the pattern of the epaulette mate pops up. Mission of the TP accomplished.

Another example:

White to move
Attackers: Q,R,N,B
Target: K
Attacking squares:
f8 +0
g7 -1
g6 -1
h7 -1
e6 +0
f7 +0

The standard reasoning is:
  • What is the most promising square. In this case: f8,e6,f7
  • Can you add an attacker or?
  • Can you annihilate a defender?
 What pops up is that the knight is overworked. Mission TP accomplished. Now the moves suggest themselves:
1.Rxf8 removing a defender of e6 and h7
1... Qxf8

What pops up is that the rook is overworked. Mission TP accomplished.
2.Bxe6+ Rxe6
With the TP, everything that is happening is so much clearer. Taking the burden of the STM by simplifying the position.

Now a quite different animal. Has nothing to do with the previous. The following diagram provides an excellent exercise in visualisation (I). Can you visualize this beautiful mate in 8? If you do 3 of these before breakfast each day, you will become a visualisation (I) monster.

White to move


  1. If your thinking process is "good" then your rating will not decrease a lot while using it. Keep it for several weeks ( to make it a skill). Then stop using it conciously: your rating will raise, stay at the high level 1,2,3 weeks and then decline again. Seemingly this procedure can be repeated. That is at least my experience

    By the way the first one with the idea of a reduced mandra thinking system was Munich.. he did use : "to take is a mistake" with sucsess
    The idea that a thinking system create skills is seemingly my idea at least i did find it nowhere else buuut maybe munich was mentioning even this before me ;)

    There is a implicit step of "chosing a target" in your thinking system.

    At the first puzzle you have a typo:
    1.Rd8+ Qxd8 not 1.Re8 Qxe8

    1. If you need to stop using the TP consciously, you already know that you haven't internalized the skill.

  2. Thx for the correction.

    Already in june 2005 I was romping around with TP's. Inspired by the Knights Errant, as I read in my old post.

  3. Very inspiring post! :). Now I feel you have make a breakthrough - at least in clarity of thoughts and simplicity of ideas (merged into one). I LOVE it!

    Just to be sure if I think the same way as you:
    1) TP stands for Thinking Process, is it right?
    2) What are the values after the listed moves (-1, 0, etc.)
    3) Diag 1: what is the "c6" stands for? Do you mean b3 - as a check from the Queen?
    4) "Attacking squares are squares where an attacker can attack the target" - do you mean ALL the pieces that can give check to the opponent's King?
    5) Diag 2: The last description: "f7 +0" - is it a mistake or any piece can give check from this square? (at least I do not see which one)

    The last puzzle was AMAZING! I have solved it within 2 minutes!!! I am really proud of myself as I noticed the theme (motif) immediately! After that I thought it is mate in 3-4 moves, but the King ran away quite a long way. Anyway I checkmated it at move 8! It was great pawn move - 8.g6 mate final! :)

  4. 1. TP = Thinking Process, indeed. There is a list with abbreviations in my sidebar. It is pretty up to date.
    2. #attackers minus #defenders. With >= +1, you dominate the square.
    3. Rxc6 attacks the Queen. For the sake of completeness. I hope that the relevant squares pop up after the training. c6 might then not be among them, I assume.
    4.In these specific cases it is about the checks, most of the time. But if the theme is a double attack, then the square from which the double attack is given is the attacking square.
    5.Well, yes and no. According to my definition, f7 is not an attacking square. But from there, you cover the attacking square g7. So you can call it an attacking square of the second order. Since two of the white pieces converge at f7, it might be an important square to reckon with.

    You have to personalize the TP, of course. It is about what works for you.

    The third diagram took me a minute or 20. I didn't quite see the correct auras op the pieces when the king was in the middle, so I thought he would escape. I did a few of these type of problems today, and I noticed an increase in ease of the visualisation(I) process.

  5. I merely note as an aside that in problem 2, the two phrases "the knight is overworked" and "the rook is overworked" are indicative of the "function motif." I think (perhaps) that this motif is incorporated indirectly into your first two steps: 1. Scan for attackers; and 2. Scan for targets. I thought of h7 as the "ultimate" target, with f8 and e6 as intermediate targets, to allow for mate on h7. But then, I try to look for functions and geometrical relationships first.

    I believe there are multiple ways to "look" at a position (different thinking processes). The most important issue is: does it work for YOU consistently, regardless of the position? If so, then it is "correct" for YOU! Someone else may (and most probably DOES) use a somewhat different thinking process/sequence to arrive at the solution.

    1. The overworked pieces are guarding attacking squares. Hence they reveal themselves once you start to interrogate the attacking squares. I often overlooked the overworked pieces exact for that, I looked only at the attackers and the targets, not at the squares.

      Indeed it is quite personal. If all sorts of information pops up when you say to yourself "look for the motifs", then go for it. I have not build such network of associations around "motif".

      I have done a lot of thinking about the duplo attacks and their associated targets, attackers and attacking squares, and the method I'm proposing in my post perfectly fits in. For me.

  6. Here's a "funny" experience: I saw the obvious first three moves, and then got stuck on how to checkmate the Black King in the center of the board. Finally, I loaded it into Fritz and misplaced the White Pawn on g5 to f5! I played around with THAT (altered) position for a while, not seeing how to checkmate. Finally, I turned on Stockfish - and found that Black was winning in every variation!! (Remember, I loaded the position incorrectly.) Finally, I stopped the engine and went back to the original diagram, and slowly discovered MY error. At that point, I just turned Stockfish back on, and saw what I had not "seen" previously. It is a beautiful mating sequence! Who would expect that the Black King would travel across the entire board to be checkmated on g7?!?

    1. I was afraid that adding the FEN would deprive you from the euphoria when you finally reach your goal.

  7. I am going to comments the second puzzle - as it contains MORE info than we (or at least - me) assume. I will do it as a format you could copy and paste into your chess program.

    I am not sure, but the STEP by STEP analysis is quite good to me - it helps me to open my (mind) eyes.

  8. You can copy the text below into your chess program.

    [FEN "3q1nk1/pp2r2p/2b1p1pQ/6N1/3P4/1BP5/P5P1/5RK1 w - - 0 1"]

    1st move has two goals: destroying the guard of h7 and deflecting the black queen

    1. Rxf8+

    double attack (check) K+Q. Now Black has only ONE legal move...

    1... Qxf8

    the Black King has three vacant squares, but the only one of these is possible to move into (h8). What would it be IF the Black Rook did not exist? You could have make a checkmate at h7 (Qxh7#)

    2. Bxe6+

    Now let's list ALL the legal moves (capture, escape and interposing): A) Rxe6 B) Kh8 C) Rf7 and Qf7 There are 4 moves in total

    2... Rxe6

    [2... Kh8 the Black King leaves the protection of his Queen 3. Qxf8#] [
    2... Rf7
    the Black Rook has been pinned - and it cannot control the key h7-square. That's why the mate next move is possible
    3. Qxh7#] [2... Qf7
    and the last move to analyse - this time Black Queen has been pinned - and it cannot control the key h7-square (neither Re7 - as an X-ray theme).
    3. Qxh7+
    now as the White Queen is protected - the ONLY move is to escape with the King...
    3... Kf8 4. Qh8+
    White Queen attacks via the 8th-rank - now the ONLY move again... (please notice the negative role of Re7 at this position!)
    4... Qg8 5. Qxg8#]

    3. Qxh7#

    it is checkmate as the rook has been deflected


    And now my comment - I list these at the specific order:
    1) Did you analyse ALL the variations till the VERY END?
    2) Did you visualise ALL the variations till the VERY END?
    3) Did you analyse ALL the final (mate) positions?
    4) Did you miss or stop the analysis after 1.Rxf8+ Qxf8 2.Bxe6 Qf7? If yes - why?
    5) Did you find the mate after the previous variation (see point 4). I mean - did you find the moves Qxh7+, Qh8+ and Qxg8#?
    6) Did you notice that the first move consists of THREE themes (destruction of the guard, double attack and deflection)?
    7) What move order you have analysed the Black's defence at this position? [capture, escape and interposing or in a different order?)
    8) Did you stop your analysis too soon in ANY of the variations shown above?

    I hope these questions may be any help for you. This post has given me another ideas to consider with. I am VERY grateful for your works (findings, sharings, comments, conclusions and encouragement to "dig the ideas") my dear friends!

    1. 1. yes.
      2. yes. Visualization I
      2. The king is dead. What's there to analyse? I took my time for visualization II though.
      4. No.
      5. yes.
      6. yes.
      7. I assume that you mean after 2.Bxe6+. First Rxe6, then Rf7, Qf7, Kh8
      8. No.

      I take considerable time for all aspects. Like Robert, I don't belief in crunching big numbers of puzzles. There are three spearpoints: TP, visualization I and visualization II.

      The TP phase tends to be very chaotic, alas. I use a foam pointer, to point out the attacking squares at my screen. That helps to focus. Once finished, a position like this is so simple, that little further calculation or visualisation I is required. That emphasizes that the TP is paramount.

    2. It means you are doing good work! The question no.3 is incorrect - I meant: "Did you check if the MATE position is a real mate position (there is NO defence)".

      I do not believe at crunching thousand of chess puzzles WITHOUT a good method/system of solving either!

      The another question that popped into my mind is: DO you check the variations at the board/screen AFTER you have solved the puzzle? I mean - if you compare the written (remembered) variations with the real ones (on the board or screen) when moving pieces.

      And the last question that makes me wonder: [Diagram 1] - DO you analyse the move Qb3+? From the words you have written - I did not notice it. And if not - why not? I want to emphasize it as sometimes there is a tendency to MISS such called "long moves" - the moves that fulfills (one or) two conditions: they require to move many square by the line-piece (long range piece) like B, R or Q. And the second condition is that such moves may be "short" (just 1, 2 or 3 moves "physically"), but their influence is long-range. What do I mean? You have two (hypothetical) cases (Black King stands at h8 and White Queen at b1): 1) You play Qb1-h7# and/or 2) You play Qb1-a1 (short move by the Queen, but attack at the long distance to the King).

      What do you think about that?

    3. 1. Yes, I did a post mortem on the corpse of the king.
      2. Yes, all variations are given by George Renko, so I can walk through them with the chessbase reader.
      3. I didn't miss Qb3+ while solving. I did miss it while writing the post though. It is not so simple to find all attacking squares, and indeed, long moves in unexpected directions tend to be overlooked.

      What I try to accomplish is to be aware of the aura of all my attackers at the same time. How they work together. But it is easy to get distracted by something interesting.

    4. "What I try to accomplish is to be aware of the aura of all my attackers at the same time"

      For further board vision exercises see the Board vision links at my blog

    5. I have done my share with Fritz board vision exercises. Besides RSI I didn't reap much results from it. What I mean is to see the cooperation of my pieces. How the queen, knight and bishop cooperate while chasing the king. Now I often focus at one piece at the time. Qc4+, mmmm, hé, I see that my bishop can check now. Seeing your own pieces as new every time in stead of being aware of their influence all the time. The difference from feeling that you are in the variation in stead of having an helicopter view. I usually act more as a soldier than as the general of my troops.

    6. The Fritz board exercises are centered around the pieces, not the squares, if I remember well. And one piece at the time.

    7. These board vision exercises supose to improve a subsubskill which supose to enable!! to improve the next higher subskill .. .. ..

      If you try to accomplish "to be aware of the aura of all my attackers at the same time" you need to be aware of
      1) the position of each of your attacker
      2) the position of all of your attacker
      3) What each piece attacks

      That would be "bottom up" = 1 step at a time

      Personally i am convinced that such exercises have to be embeddeb in a related TP to show any positive results

    8. What do you consider a reasonable amount of mate in one (easy) per minute?
      The same question for the Fritz board vision exercises?

    9. There are almost the same board vision exercises at named "attackers" and "defenders".
      You can see here and here how high the ratings in this diszipline are and compare it to other ratings of these players. There are my own ratings too, i am aoxomoxoa there. My tactical strength in CT-Blitz is ~1930 so i "should" be something between 2??? and 4???? times faster than you... if you and i would have the same perfect hand (mouse) eye coordination and so on. ( What i mean is there is some "non chess noise".)
      Tomasz had about ~~40 mates per minute me ~~20 so, ignoring the noise factors Tomasz should be 100-200 elopoints better than me in ( fast ) tactics. Seemingly Tomasz has a very bad "K" ( meaning he is good in easy tactics (vision,..) but bad in complex tactics (thinking,..))

      My Thesis: Board vision is improvable and mate in 1 ( tactical vision ) not.

      Explanation, Observation:
      At the beginning of such exercises you improve seemingly quick, you get used to the size of the board, graphic of the pieces, way of handling the task at this server..
      After 1000? (300?-4000?) attempts you dont need to think about: what you have to do and how, but just simply solve the task, you are used to the system you reached your "real" score.
      And now we are supose to ( slowly ) improve... ( gaining a speed factor 2,3 ( which would be an improvement by 100-200 elospoints ) and or even more? Which does work with board vision training but not with Mate in 1. Which did give me the impression that tactical vision would be not improvable ( at least not decisiv improvable )

      Salim Belhaddad did give me the inspiration to work on the M1 problem further by simplifying it further but there is a gap, at least for me

    10. When I did the mate in 1 from Polgars brick, I took my time for each problem, and I noticed a shift in vision from the pieces to the squares. I always had the impression that that shift in vision played a big role in my rating improvement, since that coincided with my biggest improvement ever. Although that might be a "coincidence". When I do your mate in 1, I don't feel that shift in vision. Probably due to the urge to move on to the next problem, learning nothing from my mistakes.

    11. Yes somewhere there might be the explanation. Mate in 1 is not improvable ( to master level just by solving many mate in 1 puzzles ) but other more simpler tasks are (for example Fritz boardvison check
      A mate in one is "complicated" because its about the controll of 9 squares "at the same time". There are posibilitys for "simplification" as you did..
      There might be other different simplifications:the mating pattern, thats something Tomasz used..

  9. The theory sounds not illogical. Especially not when it can be backed up by evidence based on brain scans. I see a few areas of concern, though.

    Concern 1.
    I dabbled 5 minutes in the chess gym with pointing the defenders. Avg: ~20 pieces/per minute. Already I feel that my low physical hand speed is a limiting factor. With higher speed the noise will soon be stronger than the genuine signal. Maybe long before my maximum mental speed is reached.

    Concern 2.
    More precision is needed to identify the most critical subtask. I look what is going in my mind when solving tactics. What I see there is that the visualisation-I of a variation is hampered by a very specific type of board vision. I already told you that I have difficulty to see the aura of the pieces. Not to see if a piece is under attack, nor if it is defended, but how it restricts the movement of an enemy piece. Concrete: during a king chase, I have to see the no go squares for the enemy king. I can see them, more or less, with considerable effort. But it is quite far off from automatic, without error, fast and without effort.

    Concern 3.
    What happened to the magic of the unconscious? If I learn how to shift gears, my unconscious mind adds elements that I didn't ask for like reacting to the sound of the motor etcetera. Must I learn every single subtask? That is not logical.

    1. 1. Yes
      2. Yes!!!
      3. Give me the telephone number of your magican. Fact is.. we just dont get better ( in vision ) or did you? When the brain cant learn "it" as a whole, we may try to break it down. Babys dont start to learn ballet, they start with sitting, standing, walking, running, jumping, .. The implementation of a thinking process is just doing this: temporary the attention is reduced to a small(er) and more easy but useful subtask, you make every puzzle/position to several (board) vision training exercises. The hope is that if you improve enough relevant subtasks the brain will be able to see the light and learn the other tasks by magic (unconcious). My driving teacher did help me and gave me signs in the beginning, when to shift gears, even how to do it... he did not trust magic..

  10. The "magician" is embedded in the "inner Robert Coble" module. I have a spare one that I'm not using, if the price is right. ;-)

  11. @ Temposchlucker:

    I certainly understand (and agree with) your decision to NOT "adopt a system that is invented by somebody else." That understood, here is an excerpt regarding a generic thought process from NM Dan Heisman's intriguing book The Improving Chess Thinker:

    1. What are ALL the things my opponent's [last] move does?

    2. What are all the positive things I want to do?

    3. What are all the candidate moves which might accomplish one or more of those goals?

    4. Which of those initial candidates can I reject immediately because they are not safe?

    5. Of the final candidates, which one is the best I can find in a reasonable amount of time?

    There is an elaboration of those steps in the book.

    I will confess that I do NOT follow this process per se. Specifically, I do not think in terms of a process (the Kotov method) which lists the candidate moves FIRST and then tries to analyze each subsequent variation once as a "tree." My thinking is driven by looking at the interrelationships between the pieces and squares, the available (actual and potential) motifs, and the implications of those motifs in terms of known (to me) patterns of moves (clichés). I rarely think in terms of individual moves. I also follow the maxim, "Go broad before you go deep."

    NM Heisman presents a much more detailed process in his Chess Cafe ( Novice Nook article A Generic Thinking Process. I used to read his Novice Nook articles, but the articles are now behind the Chess Cafe pay wall, so they are not generally available for reference links.

    Early in the book, NM Heisman makes the following observation.

    The de Groot experiment allows a researcher to determine how chess players find their moves during competitive play. This is clearly different than how a player solves a puzzle. In a puzzle, the solution is guaranteed. Thus a player can adopt the attitude, "If this attempt does not solve the puzzle then I will try something else. The solution has to be there."

    In the de Groot exercise the players (or subjects of the experiment) are asked to find moves just as they would in a tournament game. During a game there is no guarantee that there is anything good to find, such as a mate or win of material. The position may contain no clear ideas or candidate moves that lead to winning or even drawn positions. In many practical positions there is no "best" move. There may be several almost equally good alternatives.

    Leaving aside the logical contradiction of the last two statements (if there are several moves, even although they are "almost equally good," there is at least one that is the "best" move), there is a mindset that I find helpful in competitive play, and so I adopted it. It is simply this: try (as much as possible) to apply the problem solving mindset to every position during a game. If something does not seem to work, then look for something different that will work!

    What does that mean? BELIEVE that a "solution" (to the problem of finding the "best" move) MUST be there, even if deeply hidden. If it is not obvious (and in most positions, it will NOT be "obvious"), then search as if you are solving a puzzle. Focus your attention on each and every position in the same way you would focus if your hair was on fire. Do NOT fall back on generalizations and basic chess "principles" as the justification for making a move. Make yourself analyze concrete variations.

    That requires considerable will power, and it will not always be possible on each and every move. However, I have found that when I focus that way and think concretely about the entire position, I play at a much higher level than when I play in accordance with general principles. I also try to monitor my thinking, and if I find myself falling back on generalizations, I force my thinking back into calculating variations.

  12. Since I'm investigating, I like the clean laboratory circumstances that tactics provide. When I have found what I want, it is early enough to look how to extrapolate my findings to the broader circumstances of actual play.

    As for the TP, Radovic mentioned perceptual learning. I haven't heard of that before,so I don't know what it exactly is,but if perception is needed for that, then visual perception seems to be a logical candidate. Hence I'm looking for a TP that can be translated into images. That might well be the only kind of TP that can be internalized. If a TP can't be internalized, I'm not interested in it. For the time being.