Friday, November 20, 2020

Shifting holes

 Robert pointed out, that the metaphor with the holes in the bucket isn't applicable in our situation. Maybe it has some merit for patterns you have never seen before, but that represents only about 2-5% of the patterns, maybe. For the remaining 95% he is absolutely right.

Look at a problem I solved.

  • In 2012 I solved it in 12 seconds
  • In 2013 I needed 3 min 46 seconds
  • In 2014 I failed it after 2 minutes 23 seconds
  • In 2015 I failed after 34 seconds
  • In 2016 I needed 28 seconds
  • In 2017 I needed 12 minutes 16 seconds to get the right solution
  • In 2020 I got it right after 51 seconds

Somehow there is no system in this. As if I have a bucket with totally different holes each time. Why does the past doesn't give clues for prediction of the future?

The pattern in the problem is very familiar. So where do the fluctuations in solving time and correctness stem from?

Presumably, the fluctuations are caused by the course that system 2 (thinking) is following during the solving time. And apparently, that course is not equal, every time. The course of system 2 seems to be governed by chance.

That must mean something for the method to acquire skills. But what?

In the past, I have tried to discipline system 2 by adopting a thought process. But there are two draw backs with that:

  • A thought process is executed by system 2, which is slow by default
  • A thought process needs a lot of redundant rules in order to be applicable for a wide range of problems. Which slows it down too much.
What is it what we need to learn then?

Take the following diagram:

Diagram 1. White to move

This problem took me quite some time. When I solved it, I asked myself what took me so long?
The problem was, that I saw the rook on g6 as a piece. I saw the piece, but not what it was doing. Only after a while, I saw that the rook was boxing in the black king. I saw the activity or the action of the rook. The aura, as we sometimes call it.

I think that that is what we need to learn. Since once I saw what the rook on g6 was doing, the solution dawned immediately.


  1. Several years ago i made an intense checkmate-puzzle-training over some months.
    Result after checking my general performance: i was getting quicker in checkmate puzzles but slower in other puzzles.. over all performance remained stable.
    The thinking method has to become system 1 too, like so many rules of chess are automated now.

  2. Aox's observation regarding the thinking method is definitely correct.

    Tempo: "I saw the piece, but not WHAT IT WAS DOING."

    It's easy to get distracted. There is a B.A.D. a6 square. It is "natural" to look for ways to add an attacker on that square. The other Rook can move to h6, apparently adding that vital attacker. Black can respond in one of two ways. 1. Move the a6 pawn to a5 = [2:1] disappears. 2. Check the White King with 1. ... Rh4+ and again one of the attackers goes away.

    This "clue" (going for material advantage is not the right idea) should immediate;y trigger rejection and immediate shift back up to the vulture's eye view. What is this position really about? It's not pawn promotion (at least not in the immediate future) and it's not about material gain (Black already has a material advantage). So it MUST be about checkmate. That should trigger a recogntion of "king in the box". A secondary recognition "cue" is that the two White Rook are operating on adjacent parallel lines. If you (mentally) remove everything off the board except the two Kings and the two White Rooks, I'd be willing to bet that the solution jumps into view instantly.

    One little complication is that in order to checkmate, the h3 Rook has to check first, then retreat a square (checking for the second time), and then checkmate with the other Rook. System 2 does not like to take two moves to get to a square that can be reached in one move.

    A narrow definition of Function is to limit the concept to only defensive operations. Actually, Function (at the elementary level) is nothing more or less than the "contacts" between pieces and squares: (Attacking, Blocking, Controling, and Defending. As easy as A-B-C-D (yeah, R-I-G-H-T!).

  3. I've never read the book "Tune your chess tactics antenna" by Neiman, but I did read the chapter on "the seven signals" (you can see this in the table of contents at Amazon). The first one, and really the most important one, is "king position". If you sense that the king is exposed and outgunned, then it doesn't take long to solve the problem. If you don't sense that, then you spend time on unproductive lines.

    I'm trying to use POPLOAFUN as a training method for sensing these different signals. Seems to be helping. --mfardal

  4. Automated thinking prozess

    I got distracted by the quantity of weaknesses
    Nf6 attacked by Pe5
    Re8 by Re2 and De1
    Qc2 is hanging

    But thats wrong thinking

    1 Kingsafty
    2 im in check
    3 take, interfere, step away
    4 PXQ, RxQ , RxR
    5 won material
    6 Make move

  5. Is the problem in the diagram the same as the problem you gave the time it took to solve over the years? I assume that it is a different problem.

    Aox: "But thats wrong thinking"

    I couldn't agree with you more!

    I think this post illustrates some important training points. Memorizing the solution to a problem does NOT guarantee that anything useful has been learned, other than one MAY be able to regurgitate the canned solution on command ("monkey see; monkey do"); that does NOT indicate that SKILL has been acquired. It certainly does NOT necessarily confer any ability to "apply what has been learned" (SKILL) to a new problem (and perhaps not even to the original training problem). To me, memorization the solution to specific problems does nothing more than "load ones memory with trifles" (to quote mister Lasker).

    Knowledge is not neutral. What we "know" acts as a "trigger" as well as a filter through which we "see". All of the things we place into memory become the "database" from which we retrieve patterns, categories, abstractions, and impressions of what is important and what is not important.

    A couple of applicable quotes:

    "My mind is made up. Don’t confuse me with the facts."

    “What gets us into trouble is not what we don't know. It's what we know for sure that just ain't so.”

    We start our investigation with percepts that condition/limit what we can "see." Usually, we stubbornly cling to that "first impression" because we "KNOW" it's the right one - we have a feeling of "familiarity" that is comfortable. Unfortunately, it's entirely possible for System 1 to "suggest" a potential solution (based on the information stored in memory) with a very believable "story" to justify it. System 1 will NOT indicate that it looked at any other possibilities and rejected them because they did not "fit" the story. System 2, being lazy, takes a "look" at the proposed solution, and if there are no contradictory elements to raise suspicion, it will then endorse the solution as the correct one.

    "Who watches the watcher?"

    System 2 is supposed to be the watcher. It is supposed to apply Popper's falsification. Unfortunately, we are preprogramed for confirmation bias: we WANT the intuitively proposed solution to be true, and we persist (beyond any reasonable amount of time and effort) in trying to make the proposed solution work. This is why we cannot allow the overall system to work on autopilot. At some point, when we realize that we are hammering away (without any nails), we must force System 2 to jump outside of the system and reorient the overall system to the specifics of the actual problem.

    This is why the status examination (Weteschnik, Beim, Neimann) becomes so important. If we start with a status examination (or a look at the "auras" of the individual pieces, if you will), we have a much higher likelihood that we will not get fixated and stubbornly continue to follow a blind trail beyond all reasonable bounds.

    Just thinking out loud. . .

    1. Is the problem in the diagram the same as the problem you gave the time it took to solve over the years? I assume that it is a different problem.

      To be honest, the digits over the years I made up. I couldn't find the striking examples that struck me anymore. But you get the idea. The patterns are in store, but the retrieval mechanism is governed by chance.

  6. @ Temposchlucker: Thanks for the explanation. I'm not as certain as you that the retrieval mechanism is governed by chance. System 1 does the best that it can with whatever it has to work on, matching the specifics of the position to whatever has been stored in memory, and creating a "story" that tries to match as many components as possible. I suspect that System 2 is more the culprit. It wants certainty and System 1 doesn't report out certainty. In fact, it only reports what it "thinks" is the best explanation of what it "sees" and SILENTLY ELIMINATES all other possible explanations. System 2 is lazy, and prone to accept the provided story unless there is a little "itch" accompanying the story that says "Something is not quite right about this story; this feature doesn’t match." System 2 operates STRONGLY with confirmation bias, and only does falsification under duress.

    Here's some more "word salad" (rather long-winded; sorry about that).

    PART I:

    I’ve added EMPHASIS for various words in the following comment, emphasizing what I believe to be the most important concepts.

    mfardal (in his comment above) referenced Emmanuel Neiman’s EXCELLENT book Tune Your Chess Tactics Antenna – Know when (and where!) to look for winning combinations. (I have this book by my bed, and reread sections of it fairly regularly.) Here’s the link to the Amazon “Look Inside” (I hope; Blogger doesn’t like secure URLs for some reason):

    Tune Your Chess Tactics Antenna

    Introduction (pg. 10):

    “I hope that after reading this book, you will go for the real stuff in this kind of position, and that BEFORE LOOKING FOR MOVES, YOU WILL LOOK FOR IDEAS.”

    “Here the idea is relative to the poor king’s position. Such an idea is classically called the MOTIF of the combination. The motif is the reason WHY there’s a combination or forced win (a combination normally implies a sacrifice, here we look for ‘tactics’—that is, a forced variation with or without a sacrifice).”

    “Generally, puzzle books are arranged according to THEME. This classification: double attack, pin, deflection, decoy, etc. is a useful tool for solving purposes. WHEN WE KNOW THAT THERE IS A COMBINATION, THE THEME [the book is in error: it reads “MOTIF”] PROVIDES A VALUABLE ANSWER TO THE “HOW” . In this book, we deal with the WHY; we want to discover if there is a chance of winning by force.”

    “We are looking for HINTS, and if we can find some, then we will look at the position with a solver’s eyes—a solver who has already done part of the job. The antenna has been erected, and it involves seven filters. In classical chess literature these filters are called MOTIFS. In this book, in accordance with our “antenna” theme, we call them SIGNALS.”

  7. PART II:

    What I found most interesting is that he considers “SIGNAL” and “MOTIF” (mister Lasker’s definition of that term, not to be confused with tactical devices/THEMES) to be synonymous. Perhaps there is more to this separation of MOTIF and THEME than just calling the same thing by a different name. IN THIS DISTINCTION IS SOMETHING VERY IMPORTANT CONCERNING HOW TO LOOK FOR “CUES” FROM THE VULTURE’S EYE VIEW!

    Per Introduction, A. What is a signal?:

    A signal is a weakness in the opponent’s position.

    “The signal (the MOTIF) is the reason WHY the combination exists. The THEME is the main mechanism (HOW) which allows us to make it work.”

    “If we don’t find the thematic double attack (THEME), we won’t win, of course. But the win exists only because of the signal (MOTIF), and here, once we are aware of the MOTIF, the realization is not difficult.”

    “Both (MOTIF and THEME) are complementaries; for example sometimes the signal is clear, yet the combination is difficult to work out, because we need to use complex calculation.”

  8. PART III:

    Having flogged that distinction between MOTIF and THEME, let’s switch our attention to the recommended “thinking process”. The Amazon link has the complete Introduction, B. The thinking process, (in case you don’t have the book), which I consider to be as important as the identification of the seven “signals” and what to do with them.

    Rather than copy this section word-for-word, I refer you to the link, and merely summarize the five main steps here.

    1. GLOBAL VISION For me, this is the “vulture’s eye view”. We are looking for a “general impression” of the overall position.

    2. ANALYSIS OF THE POSITION For me, this corresponds to Weteschnik’s “status examination” or (more generally) Beim’s “examination of the position.” Here is where PoPLoAFun comes into play.

    3. LOOKING FOR THE THEME This is where we begin to develop a “plan”, knowing what is “desirable”. If we have found SIGNALS (MOTIFS), then we can begin to look for THEMES to exploit those SIGNALS. There are specific THEMES related to each SIGNAL. This helps narrow our focus to the relevant issues.

    4. LOOKING FOR CANDIDATE MOVES Once we have a “feel for the overall position, have identified WHY some tactic(s) might exist, and what THEMES might be available, we are (FINALLY) at the starting point for GM Kotov’s prescription for “Think(ing) Like A Grandmaster”. There is no good reason to “think like a tree” and try to duplicate how a computer searches through the Tree of Analysis. Pick one reasonably strong possibility and then move ahead with it, looking “deep” before looking “wide”. This is GM Tisdall’s idea of “variation processing”.

    5. THE CALCULATION OF VARIATIONS If attacking, pick what you think is the strongest variation and analyze it first. If defending, pick the variation that you think is most critical (most likely that you will lose) and analyze it first. Pay close attention to any “surprises” (tactical shots) that occur down in the weeds; these may be useful if this particular variation doesn’t satisfy the “requirements of the position”. Having those “surprises” in reserve can be very helpful when you switch variations. A “tactical shot” that works in one variation may also work in another one; or, vice versa, a “kill shot” from the opponent in one variation may also be a “killer move” in another variation. This is a great time-saving device!

    Note that the “thinking process” is nothing more than moving from the “general” to the specific, in spirals of increasing detail. It is not the usual logical step-by-step process with which so many people become enamored because of its Logical appeal to System 2. Very important is what is NOT stated: allow System 1 to become familiar with the setting (the position) and all the characters (the pieces and their relationships to each other and to squares) before rushing to formulate a “story”. The richer the cast of characters and nuances, the more likely that System 1 will form an accurate “story”. This must occur BEFORE allowing (or directing) System 2 to begin its work of logical calculation of variations.

  9. My solving time is definitely influenced by what problems I have just done. If the solution is a pin and I just did another pin problem I am more likely to get it quickly. Also highly influenced by how awake I am. I've resolved to stop solving CT puzzles when I get two wrong or slow in a row.

    But it's also influenced by solving method. By taking time to look the state of the board (POPLOAFUN / "signals"), I don't solve any problems super-quickly, but I think I'm not wasting as much time going down blind alleys either. --mfardal

  10. PART I:

    Let’s approach this “question” from another direction. Paata Gaprindashvili has an excellent problem book, Imagination in Chess — How to Think Creatively and Avoid Foolish Mistakes. (I’m not so sure about that first part regarding thinking creatively, but the second part about avoiding foolish mistakes certainly rings a bell with me! Ding, ding, ding!)

    Chapter 7 —Imagination has some intriguing examples and some caveats regarding KNOWLEDGE. I’m not going to copy everything verbatim, but am going to use three of his examples to show a possible scenario where KNOWLEDGE can hinde rather than help us solve problems.


    Example 1: A. Selezniev, 1921 (conclusion of a study)

    White to play and win — in one move

    FEN: 3k4/8/8/8/8/8/8/R3K3 w - - 0 1

    We have shown the Example 1 position to many a junior chessplayer. The task — “White to play and win”, or again “White to win in one move” — left them bewildered. Hardly anyone could find the solution. WHY? Let us look into it. Two factors should be noted:

    (1) The pupils possessed the indispensible knowledge for solving the position; they knew about the right to castle.

    (2) They also knew that castling USuALLY takes place at the beginning of the game, and that if it doesn’t, something will happen to deprive the player of his castling rights. In other words, they had systematized and generalized their KNOWLEDGE. As a result, the children subconsciously INFERRED that castling must be impossible in an advanced stage of the endgame. It was this inference that prevented them from finding 1. O-O-O#.

    Most of the small number who did give the right solution had come across something similar before— see Example 2.

    Example 2: A constructed problem

    White to play and win — in one move

    FEN: 8/8/8/8/8/8/1R6/R3K2k w - - 0 1

    Discovery of the solution (to Example 2) made such a strong impression on them that the idea stuck in their memory. Thus they found the concluding move of the Selezniev study with the aid of an ANALOGY, that is, through KNOWLEDGE and recollection. What about the remainder of those who found the move? They had managed independently to rid themselves of false inferences restricting their imagination.

  11. PART II:

    Let’s take a practical episode from a game between young players.
    Kamsky-Tiviakov, Daugavpils, 1986

    Black to move

    FEN: 1rbq1rk1/3nppbp/2np2p1/3N4/1p1NP3/4B1PP/1PP2PB1/R2Q1RK1

    [FAIR WARNING: The solution to this position does NOT share anything with the previous two examples!]

    Of course when Kamsky made his last move, 1. Ne2-d4, he had considered what would happen if Black captured on that square. Yet he was only thinking about a knight exchange, which is characteristic of such positions. The unconventional move 1. … Bxd4! escaped his attention. After 2. Bxd4 e6 3. Ne3 e5, Black emerged with an extra piece.

    WHY did White overlook 1. … Bxd4? Kamsky (just like any other chessplayer) had probably seen a knight exchange dozens of times before in this type of position. Black’s surrender of the fianchettoed bishop, which in most cases runs counter to common sense, was something he had perhaps never once come across. HIS KNOWLEDGE HAD BECOME GENERALIZED; for him, the knight exchange had become the typical precept. Hence, 1. … Bxd4 was SUBCONSCIOUSLY excluded as a possibility.

    We may say that IMAGINATION IS ENDANGERED NOT BY KNOWLEDGE BUT BY GENERALIZATIONS FROM IT, FOR THEY ENTAIL VARIOUS SUBCONSCIOUS INHIBITIONS. And yet generalizations from knowledge are a human characteristic that is essential to life. How can this contradiction be resolved?

    Clearly, the ability to think in a non-standard manner needs to be developed in young players from the time of their very first steps in chess. This means the coach must not merely impart knowledge and teach them the principles and methods of combat, but also demonstrate exceptions to the rules, and, most importantly, stimulate their imaginations.


    What can we take away from this discussion? Knowledge is REQUIRED, but that same knowledge (after generalization and absorption into the subconscious) may end up causing problems in specific situations. This means that it is not necessarily lack of knowledge nor randomness in the operation of our systems that may be at the root of our inability to “see” the contours of a solution.

  12. As I understand it Tempo hopes to use POPLOAFUN annotations to absorb useful patterns. Maybe this works...but there are so many possible patterns to learn. My goal is somewhat different: to use it to train the skill of seeing what weaknesses are in opponent's position. Thought process training, basically, though not conscious thought. --mfardal

    Distracted, i found move 2 late. The situation is: I suffer from a double attack and i need to de-double attack: remove the piece from the attack to interfere the second attack

  14. An example of why we should pay attention to that “itch” that something is just not quite right about a position. . .

    FEN: q4rk1/1r1Qbppp/2p5/1p2p3/1P2P3/2P4P/3N2P1/2B2RK1 w - - 0 1
    This is from Test Your Chess IQ: Master Challenge, Test 38, Position 303.

    The instruction/hint for this problem are:

    By 1. … Qc8 Black allowed the capture of his bishop. What was the hidden point?

    Beats the heck out of me!! Given the Black Rook on b7, I would hightail it off the 7th rank, rather than exchange the White Queen for a Bishop.

    As soon as I looked at the problem, I felt something was “off”. So, I looked at the solution:

    Karaklajic – Boli, Sombor, 1957 (searched for the game but couldn’t find it).
    1. …Qc8! 2. Qxe7?? (2. Qd3 was correct [NO JOKE]) 2. … f6!! And after 3. … Rf7 White lost his queen and the game.

    I find that incomprehensible; White was rated almost 2500 Elo at that time.

    Then the rationalization kicks in. Maybe the piece on b7 is a Knight. That would deprive the White Queen an exit via d6 or c5 from e7 (after taking the Bishop). Alas! Without the game score, no way to determine if my conjecture is true.