Friday, April 07, 2017


On my quest to organize backwards thinking, I already expected a few hiccups while proceeding.

Choice between two ducks
The first hiccup is formulated by Tomasz as "we have to find (work out) a small difference between BIG ducks". From time to time, there is not one single immobile piece, but there are two. My impression is that that doesn't happen very often. But maybe, now my attention is triggered to look at it, there is a chance that I have to change my opinion, but we'll see about that.

The ducks are deemed by their apparent size, which translates to chess as the value of the piece. Well, that shouldn't be too hard. In order of value: K, Q, R, B, N, p

I look from the perspective of minimizing the brain load. When we have a king and a bishop that are both sitting ducks, we start with the king, and forget about the bishop. But that leads to the formulation of the second hiccup:

When to abandon thinking about a sitting duck?
There comes a moment when you should stop thinking about a combination once you see the sitting duck you are craving remains out of reach. I can think ad infinitum about it, so I definite need a breaking mechanism. I hope that working out the scenario's will help me out here. When I have investigated every standard scenario and none of them works, that should be the telltale sign to move on.

Counter attack
The hiccup I dread the most is the counter attack. Where do I place it in my tree of scenario's? The best way to overload my memory is to start thinking about counter attacks too early. How can I make that I start thinking about counter attacks exact just in time but not too early? What is the telltale alarm clock?

The last hiccup I'm confronted with is the defensive move from my opponent. When do I start looking at defensive moves?

As you might notice, I want to be in control of my thinking. I don't want to get sucked into a tunnel that leads to nowhere without me noticing it. Which is the normal state of affairs when thinking about a position.

The following position is going to be the subject to investigate the hiccups above. The diagram contains two sitting ducks and a counter attack. While I'm investigating the position, feel free to comment on the subject already. I will update this post when I have found something relevant to say about it.

Diagram 1. Black to move
3RQbk1/5p1p/6p1/2P3n1/4B1q1/4P1P1/Pr5P/5NK1 b - - 1 1

The first thing my eye fell upon was the duplo attack 1.... Ne6.
  • It defends against the counter attack 2.... Qxf8# (Notice that I was aware of a counter attack!)
  • It interrupts the defense of the bishop
  • I could find no defensive move that saves both the white bishop and rook (Notice that I checked for defensive moves from my opponent)
What I failed to notice though, is that white has another counter attack against the point of pressure f7. With 1. ... Ne6 2.Rd7 white saves both the rook and the bishop, since I must take care for f7 and have no time to take the bishop.

The fact that I had checked for a counter attack and for a defensive move, gave me a false sense of security.

Mister Lasker gave some terrible advice: "if you see a good move look for a better one". This advice is impossible to follow during OTB play, since I will loose due to time trouble time and again. But now I'm in the study room, and I followed his advice. So I started to look at the white king.

This is a typical case of two sitting ducks: the white bishop is immobile due to a duplo attack and the white king is immobile due to lack of space. As I pointed out, it is more logical to start with the biggest duck. If I had done that, I would not have needed to look after the move 1. ... Ne6. Automagical pruning prevents the necessity to look for a better move. There can't be one, since you are already slaughtering the biggest duck.

I was aware that only a check could dismiss me from looking at the counter attack. 1. ... Nf3+ 2. Kh1 Nf2+
What I failed to notice here, is that this is the wrong check. After 3.Kg1 I cannot take the bishop. I did though, since I had forgotten the pending counter attack.

So here you have it. I do most things very well, but time and again, little slips spoil the perfect score. A little lack of precision here, a bit forgetting there, some minor oversight yonder, starting at the wrong end elsewhere. Only a systematic and disciplined approach can help me out.
  • Start with the biggest duck
  • Not every check dismisses you from looking for counter attacks
  • There can be more than one counter attack
  • When you see two moves that seem to accomplish the same, don't be satisfied until you know which one is best and why.


  1. Solved it after 1:12

    I was not concentrated and did start with looking for potentiel targets and i did need some willpower to get back to the routine, so i did restart at something like 0:30

    Material: White is up 2 Pawns
    Theoretical tactical weaknesses: Kg1,Be4,Pc5,Pa2, kg8 rb2, qg4
    What is the puzzle about : Checkmate ( for every move of black: either black is giving a check, proteckting f8 by playing qh6 or gg7 , or getting checkmated ),

    So only one move Nh3+, Kh1 and now i recognised the fork at f2, winning the bishop by Nh3-h2+ and Nxe4 is a selfmate via Qxf8+ but i recognised imediatly that Qxe4+ forces White to take at e4 and then the fork at f2 winns back the queen.
    Now the material balance is not big enough for CT but i saw that the position is won for black , the white a and c pawns are lost.

  2. Tim Brennan just sent this in his Tactics Time Newsletter #289, 6 APR 2017.

    [FEN "3q4/k1p2br1/1pQ5/p2pR3/P2P4/1P1B4/K1P5/8 w - - 0 39"]

    Think INSIDE the box!

    The position below is a good illustration of a mini-process:

    (1) "See" the "box" around the enemy King.

    (2) Visualize a potential checkmate.

    (3) Determine which of the opponent's piece(s) prevent that checkmate.

    (4) "Move" the opponent's piece(s) so that the checkmate CANNOT be prevented.

    This is the "creative" (tactical) step.

    (5) Checkmate the King.

    From this game:

    [Event "Dortmund Sparkassen Chess Meeting"]
    [Site "Dortmund GER"]
    [Date "2014.07.13"]
    [EventDate "2014.07.12"]
    [Round "2"]
    [Result "1-0"]
    [White "Fabiano Caruana"]
    [Black "Ruslan Ponomariov"]
    [ECO "C42"]
    [WhiteElo "?"]
    [BlackElo "?"]
    [PlyCount "81"]

    1.e4 e5 2.Nf3 Nf6 3.Nxe5 d6 4.Nf3 Nxe4 5.Nc3 Nxc3 6.dxc3 Be7 7.Be3 Nc6 8.Qd2 Be6 9.O-O-O Qd7
    10.Kb1 Bf6 11.h3 h6 12.b3 a6 13.g4 O-O-O 14.Bg2 g5 15.Nd4 Nxd4 16.cxd4 d5 17.f4 gxf4
    18.Bxf4 h5 19.g5 Bg7 20.Rde1 h4 21.Be5 Rdg8 22.Qf4 Qd8 23.Bf1 Kb8 24.Bd3 Bc8 25.Kb2 Bxe5
    26.Rxe5 Rg7 27.a4 a5 28.Ka2 Ka7 29.Qd2 Kb8 30.Qf4 Ka7 31.Rhe1 Bxh3 32.Rh1 Bc8 33.Rxh4 Rxh4
    34.Qxh4 b6 35.Qh6 Rg8 36.Qc6 Be6 37.g6 Rg7 38.gxf7 Bxf7 39.Re7 Qxe7 40.Ba6 Kxa6 41.Qa8# 1-0

  3. 1946.2 - it took less than 45 seconds to "see" the solution.

    PART I:

    When I posted previously, I didn't bother to look at the puzzle above. That's why this post comes later in time.

    Black is down in material by two passed Pawns, so he must either restore the material balance or force checkmate.

    The White King "might" be close to being in the box - or might NOT. So, the position might be about checkmate (3-piece rule) or possibly (re-)gaining material. (I ignored any promotion possibility because of the distance of the White Pawn from the queening square [3 squares], but that MIGHT have been important; see below.) The White Bishop is already attacked twice and defended once, so that's a ready target for tactical themes/devices. But, it is obvious (I think) that it is immune from capture - Black cannot just capture the White Bishop because White threatens checkmate with 2. Qxf8#. AHA! Here is the counterattacking threat!

    So, Black must work with forcing moves (checks would seem to be most appropriate). The most obvious move, 1. ... Nh3+ seems to be the most forcing move; it forces the White King to go to the corner with 2. Kh1 (due to the Black Rook at b2). The White King IS now in the box and has no moves. So, we "look" for a check, any check, no matter how "ridiculous" it appears at first glance.

    Is there a "stock" mating pattern with R+N? No, it just doesn't look like it. How about Q+N or Q+R? Nope. The White King and White Knight seem to be adequately guarding the h2 and g3 squares. Nothing is being triggered by pattern recognition for checkmate.

    We CAN give another check with 2. ... Qxe4+. White has no choice; he MUST recapture with 3. Qxe4 because there is no other legal move. The recapture also restores (temporarily) the material advantage to White, which is totally irrelevant until White can get out of check and regain the initiative.

    The White checkmate threat has been removed. Now Black needs to find a way to regain the material advantage or checkmate. None of the stock checkmate patterns seem to hold, so how do we grab material? Is there are tactical pattern trigger? AHA! There is a stock Knight fork of King and Queen on f2!

    Black finishes up with 3. ... Nf2+ 4. Kg1 (not 4. Kg2 Nxe4+, allowing Black to immediately capture the potentially dangerous White c-Pawn with check) Nxe4, having won a piece for two Pawns. After 5. c6 Rc2 Black will stop the possible promotion of the c-Pawn.

    I did not consider the possibility of promotion to be of interest when first looking at the position, because of the immediate checkmate threats. Black will be able to unpin the Black Bishop on his next move and the three Black pieces should be able to handle the White Pawn(s) without serious problems.

  4. PART II:

    Regarding one of your questions above:

    How can I make that I start thinking about counter attacks exact just in time but not too early? What is the telltale alarm clock?

    IMHO, the "alarm clock" is the moment that the initiative is lost. If you do not lose the initiative, you can safely "ignore" any and all counterattacks. In the example above, since Black can FORCE White to respond to his first two moves, there is nothing that White can do about it at least until the 3rd move. Black retains the initiative throughout.

    That does NOT preclude so-called "quiet moves." However, the critical issue when contemplating a "quiet move" is:

    Can the opponent actually DO anything critical with that tempo?

    If so, then you MUST consider what the opponent can do AT THE POINT WHERE THAT "QUIET MOVE" IS PLAYED. It is important NOT to get "sidetracked" unless or until there is a credible threat which can be executed by the opponent.

    If not, you can continue to ignore the counterattack.

    I am reminded of Mikhail Tal's quote, referring to his piece sacrifices with multiple pieces en prise:

    "They can only take them one at a time!"

    1. Time Score and Rating are related, we need more time to score better. We can ignore the opponents threats only at low rated puzzles.

      In my eyes there are 2 things to do
      1) attention to the last move of the opponent, it is often the key to the whole puzzle
      2) its necessary to look for the own tactical weaknesses too and not only for the weaknesses of the opponent

      At this concrete example the last move was Qe8 threatening Qxf8#

  5. I am not sure, but it looks I am the only one who noticed 1.Ne6!

    Why this move is good in my opinion?

    1. It stops mating threat Qxf8#
    2. It attacks Rd8 (without this piece mate threat is not possible anymore)
    3. It opens the attack against Be4.

    Probably the best reply is to play 2.Bc6 and after 2...NxR 3.QxR and the fight is on the line.

    Can anyone tell me if this variation is a good or bad? I want to know why no one pointed out this line. I waited a few days, but I could not wait any longer ;) :).

    1. @ Tomasz:


      Stockfish DD64 gives:

      1. ... Ne6 2. Rd2 Rxd2 3. Nxd2 Nxc5 4. Bf3 Qf5 5. Ba4 Qg4 = (0.0) - Draw by three-fold repetition (apparently)

      I know I was focused on "winning" for Black, using tactics (solving" the puzzle), and did NOT think of the position properly - as if I was actually playing a game. Consequently, I ignored White's perspective. This is BAD!

      Thank you for an invaluable lesson!

  6. @ Tomasz:

    On second thought. . .

    In the initial position: "Black to move." This implies that Black (somehow) must come out ahead of White when the dust settles. Black currently has a material deficit of two Pawns. If the best he can do is equalize, then your suggested line works to achieve that goal.

    On the other hand. . .

    Black is NOT forced to play your line!

    Black has a forcing sequence that leaves him with a WINNING advantage, not just equality. All of the moves up to the last move are forcing and forced: 1. ... Nh3+ 2. Kh1 Qxe4+ 3. Qxe4 Nf2+ 4. Kg1 (not forced, but better than 4. Kg2 Nxe4+ 5. K moves Nxc5) Nxe4. Note the evaluation of line 1 below by Stockfish DD64: Black is WINNING (-5.09) BY FORCE from a position in which he was LOSING!

    New game
    3RQbk1/5p1p/6p1/2P3n1/4B1q1/4P1P1/Pr5P/5NK1 b - - 0 1

    Analysis by Stockfish DD 64 SSE4.2:

    1. -+ (-5.09): 1...Nh3+ 2.Kh1 Qxe4+ 3.Qxe4 Nf2+ 4.Kg1 Nxe4 5.c6 Rc2 6.Rd4 f5 7.g4 Ng5 8.Kh1 Ne6 9.Ra4 fxg4 10.Rxg4 Rxc6 11.Rg2 Rc1 12.Rf2 Kg7 13.Kg2 Bc5 14.Kf3 Ng5+ 15.Kg2 h5 16.Rb2 Kf6 17.h4 Ne4 18.Ng3 Nxg3 19.Kxg3 Bxe3
    2. = (0.00): 1...Ne6 2.Rd2 Rxd2 3.Nxd2 Nxc5 4.Bf3 Qf5 5.Be4 Qg4
    3. +- (19.71): 1...Rg2+ 2.Bxg2 Ne6 3.c6 Nxd8 4.c7 Ne6 5.c8Q Qf5 6.Bh3 Qxh3 7.Qxf8+ Nxf8 8.Qxh3 f6 9.a4 Kf7 10.a5 Ke7 11.a6 Nd7 12.Qxh7+ Kd6 13.a7 Nb6 14.Qxg6 Ke7 15.Qe4+ Kd6 16.h4 Kc7 17.a8Q Nxa8
    4. +- (97.18): 1...Qxg3+ 2.Nxg3 Ne6 3.Rd7 Ng5 4.c6 Rb3 5.Rxf7 Rb1+ 6.Nf1 Rb8 7.Rxf8+ Kg7 8.Rg8+ Kf6 9.Rxg6+ hxg6 10.Qxg6+ Ke5 11.Qxg5+ Kxe4 12.Qf4+ Kd5 13.Qxb8 Kxc6 14.Qc8+ Kd6 15.Qd8+ Kc6 16.Kf2 Kb5 17.Qd7+ Kc5 18.Ke2 Kb4
    5. +- (#8): 1...Nf3+ 2.Bxf3 Qxg3+ 3.hxg3 Rg2+ 4.Kxg2 Kg7 5.Qxf8+ Kf6 6.Rd6+ Kg5 7.Qxf7 Kh6 8.Qf8+ Kg5 9.Rd5#
    6. +- (#8): 1...Kg7 2.Qxf8+ Kf6 3.Qh8+ Ke7 4.Re8+ Kd7 5.c6+ Kc7 6.Re7+ Qd7 7.Rxd7+ Kb6 8.Rb7+ Ka6 9.Qa8#
    7. +- (#2): 1...h5 2.Qxf8+ Kh7 3.Qh8#
    8. +- (#2): 1...h6 2.Qxf8+ Kh7 3.Qg8#
    9. +- (#1): 1...f6 2.Qxf8#
    10. +- (#1): 1...f5 2.Qxf8#
    11. +- (#1): 1...Nxe4 2.Qxf8#

    (Coble, Asheboro, NC 09.04.2017)

    So, which line would you choose if given the BLACK side in the initial position? (That is the implication when the position is "Black to move".)


    I still think it was an outstanding observation AND that I also need to consider problems as just one step in a game, rather than just "solving" a problem.


  7. @Aox: the time score is different for different players, because they have a different "k-factor".

    It was interesting for me to find out that my "Ultra-Bullet" at is really poor. Much worse than one would expect for an expert player.
    I could imagine that Tomasz would score very well in "Ultra-Bullet" (= 15 sec time control for the whole game, no increment!).

    1. k does tell us how many elopoints a tactician gains in standard mode if he thinks twice as long. This puzzle is interesting here: the rating of the puzzle is 1945 if its solved in 40sec and 1764 if its solved in 155sec.
      That is interesting because usually a 200 points stronger tactician should need halve as long to solve this puzzle, here he need just 25% of the time.
      I suspect k to be a calculation factor, players with good visualisation skills, good memory of the already calculated lines, which have a better organised thinking asf should have a higher k. This puzzle can be solved by a lot of logical reasoning, this does speed up the solving speed of the better tactician extremly.

    2. @Munich

      I am sorry, but Tomasz can barely play bullet games (1+0). Faster games are simply above my mind. I like playing 3+0 or similar time control (I can call it "fast blitz" or very slow bullet" game). However playing "Ultra-Bullet" at lichess requires extraordinary skills (and I do not have any of these). I have watched UB games by IMs and GMs at LC - they are very good internet connection, good computer (no freezing the screen or slowing down the system) and very high level of skills. Not to mention 75% of their moves are based at pre-moves.

      Recently I rather play longer time control games (like 30+30 or 45+45) and that's why it is extremally hard to play those 1+0 (not to mention faster) games at a decent level. Beside that I have no time to check any options: I am forced to move, move and move... all the time (no verifying if the moves are sound, etc.).

      My ultimate chess limit is 1 minute per player. It is the same as solving chess salt mines (mate in 1) with the speed of 50 MPS (mates per second). You would like to break this score, but it looks impossible.

  8. My final solution was Ke6 but only because I didnt calculate the knight fork after Q X Q. Saturdays puzzle has a nice tactical shot and two surprising moves in the combination. Wtm.

  9. @aox: but the average solver of the 1945 rated puzzle might have a lower k-factor than I have.
    He might even have the same k-factor (on average) like the solvers of the 1764 rated puzzle.

    Note that a "high" k-factor does not tell us anything about the level of rating, but only how much one can make out of more extra time.

    And usually: the higher the rating of a player, the less any extra time is helpful. Super GMs can still play reasonably well in a blitz game, and their long OTB games only gain a little bit of accuracy, but not that much anymore.

  10. I solved the problem fairly quickly (20 seconds). However, I think part of it was that I do a bunch of problems on, for which there is an inordinate amount of similar problems - e.g. sacrifice piece (temporarily) to attract/deflect another piece, then win that piece with a fork/skewer/etc.

    However, I don't always notice the motif right away, but I tried 1.Nh3+ as it was the only "safe" check, and then instantly saw the rest of the solution.

    Of course, this was kind of lucky to find right away.

  11. The takeaway for me was calculate to quiescence. Not to get lazy and take a shortcut.. also when the knight is involved always look for forks.

  12. I solved it in under 4 minutes.

    In hindsight, it looks like this should be a 30 second problem, but that's probably because it is very pattern-oriented, and I've got the pattern now. The thing that made it so solvable originally is that it's so forced, so process of elimination works well in this situation.

  13. Tempo said : "
    I was aware that only a check could dismiss me from looking at the counter attack. 1. ... Nf3+ 2. Kh1 Nf2+
    What I failed to notice here, is that this is the wrong check. After 3.Kg1 I cannot take the bishop. I did though, since I had forgotten the pending counter attack."

    To solve a problem quick we need: a parallel awareness of all important aspects of the position. But our working memory ( WM (or STM) ) is usually not big enough. So we lousy bad mortals use a thinking method to load different subsets of the important aspects sequential into the WM:

    Tempo said :
    " Only a systematic and disciplined approach can help me out.

    Start with the biggest duck
    Not every check dismisses you from looking for counter attacks
    There can be more than one counter attack
    When you see two moves that seem to accomplish the same, don't be satisfied until you know which one is best and why."

    This "splitting of the whole problem in a sequence of sub-problems and then to solve them" costs---->time!!

    We might be able to run such tasks automated after some training but more and bigger chunks would give us more speed.

  14. Off topic
    How do masters solve tactics?
    Here some videos of NM,CM,...GM solving tactics

    Simon Williams


    John Bartholomew


    Daniel Rensch



    Tal Baron

  15. Fascinating glimpse into the "master" mind! Thanks!