Saturday, May 06, 2017

Harvesting subtleties

I'm slowly working my way through the 40 positions I have been investigating since Dec 3rd. Now I have a Tree of Scenario's, I have something to give the more subtle knowledge a place. It is remarkable how commonly usable this knowledge is.

Did you for instance know: A target is protected by two defenders. When you trade off one defender, the other defender becomes immobile.

When this kind of knowledge is discovered, it feels as being subtle knowledge. But when you think about it, it is actually rather surprising that you have never realized it before. It is no rocket science, and very obvious. And remarkably commonly usable.


  1. A target is protected by two defenders. When you trade off one defender, the other defender becomes immobile.
    In fact it is one of the reasons for overprotection. Overprotection keeps your defenders mobile.

  2. Sounds like a real complex thinking system. You replace "number crunching" calculation through an expert system logical program where you have to do a functional analysis of the pieces extra. A piece is not say a Queen at c1 but a defender of the rook, attacked from the opponents Queen and defended by a knight and so on. The hope is that this new method is of lower complexity.

    I think this method is more human like but you will have to automatize it to make it really faster:

    1. run many puzzles "by hand" through that process and to tag/classify these puzzles according to your new system of pattern

    2. run many puzzles of each tag ( several times ) separate.

    By the way: the functional classification of pieces is board vision: is_Defended, is_Defending, is_attacking .. that have been boardvision exercises ;)
    We already know : Masters can do that ( usually ) quicker.

  3. The idea came in sight while studying a position. The reason I saw it is because it is so obvious. When you take away one defender, the pressure on the other defender increases. Since the plf-system is about guiding your attention to the points of pressure, this kind of knowledge is the bread and butter. Standalone, the idea might look complex, but within the framework, you can see it in a glance since it is very logical. When you know where to look for.

    The board vision exercises are way too superficial and incoherent. Besides that, we don't need to become faster in that area. We are more than fast enough already. Once our attention arrives at the scene where it is needed.

    The tree of scenario's is a framework to hang new knowledge on. So you can find your knowledge back when you need it. You can only learn by feedback from the past, when you have a storage system for your knowledge. Free hanging knowledge with no index will be lost after some time. You need an index. That is what the tree of scenario's is meant to be. From there you can develop a thinking system, but that is essentially something quite different.

    Right know, the tree of scenario's has about 40 branches. I think that is about the ideal amount. Once I think it is complete, I will scrutinize it for practical use (= develop a thought process). Only then we can talk about a thinking system. All 40 scenario's will be perfectly accessible with about 3-4 knots. So you need only to answer 3-4 questions to guide your attention to the heart of the position. From there, a little calculation will do the rest.

  4. I think this is the part where we cue Robert to say something about Beginners mind . 8)

  5. @Jim: I shall try not to disappoint!

    “Be humble: In the beginner’s mind there are many possibilities, but in the expert’s mind there are few.”

    “The goal of practice is always to keep our beginner’s mind.”

    ― Shunryu Suzuki

    The connection of the formulation you discovered (regarding attackers and defenders, and the resultant immobility if one defender is captured) to overprotection is a vital component of the PLF system. It will not be the last such connection found.

    As has been noted previously, chess is 99% tactics (or 100%, depending on whether you prefer Teichmann's or Petrosian's estimate). Since 89.37% of all such statistics are made up on the spur of the moment (including this one), we can "agree to disagree" on the relative percentage. However, to divide chess into two different spheres (tactics and strategy) is to miss the forest for the trees. In actuality, there is no clear demarcation line between them. Separating the two spheres MAY be useful (I don't think so) for pedagogical purposes, but to maintain that dualistic mindset is to NOT "see" from the vulture's eye view - and to limit our potential improvement. Compartmentalized thinking does not lead to improvement overall.

    Quoting Nimzovich, My System:

    "I was led to advance the STRATEGICAL proposition that one must over-protect his own STRATEGICALLY important points, that is, PROVIDE DEFENSE IN EXCESS OF ATTACK: LAYING UP A STORE OF DEFENSE."

    [This is the justification for what Temposchlucker described above, in different terms!]

    "Weak points, still more strong points, in short everything that we can include in the collective conception of STRATEGICALLY IMPORTANT POINTS, ought to be over-protected. If the pieces are so engaged, they get their reward in the fact that they will then find themselves well posted in every respect."

    While modelling in the past as part of the design of large, complex software systems, I came to a realization similar to Temposchlucker's in that there are some things that do not appear "obvious" and seem complex, until you examine the implications. Suddenly, what seemed complex and obscure becomes simple and obvious. BUT, YOU DON'T REACH THAT STATE UNTIL YOU LOOK FOR A DIFFERENT MENTAL REPRESENTATION THAN YOUR CURRENT ONE!

    The value of a model depends crucially on what is left out of it, rather than just by what is included in it.

    How do you determine the limits (edges) of any modelling system? When you are forced to use general prose descriptions and NOT the language of the model, you have found the limits of the model. This is the most fertile area for gaining new (and better) mental representations of the subject being modeled.

    I did NOT discover the following observation but it is apropos:

    "When faced with a difference between reality and the model, BELIEVE AND ACCEPT REALITY OVER THE MODEL!"

    If it works, keep it. If it doesn't work, discard it.

  6. I feel if I did a similar deep study my epiphanies would be around decoys to knight forks.

  7. off topics
    some months ago i had an idea how to make a chess-engine "explain" a move. Now i found an astonishing tool where you can investigate each part of Stockfish static evaluation function :

    put your position on the little board on the top right and step through the elements of the evaluation and see how they are calculated

  8. @ Aox:

    Thanks for the reference to the Stockfish Evaluation Guide. It's really fascinating to use!

    There are no real instructions for how to use it. There is a box in the upper right which allows FEN or PGN to be entered.

    I noticed the forward/backward arrows, and tried to enter the moves from a game I recently played. Just click on a piece and drag/drop it on the new square. After you have entered the entire game, you can use the backward arrow to go back to the beginning, and then see the evaluations move-by-move and the significant factors that changed.

    After returning to the beginning, select (one of the tabs on the upper left) either Table or Graph.

    Table provides a comparative (relative) table of values for White (Middlegame/Endgame), Black (Middlegame/Endgame) and Total (Middlegame/Endgame) for Imbalance, Initiative, King, Material, Mobility, Passed Pawns, Pawns, Pieces, Space and Threats.

    Graph provides a comparative (relative) bar graph for the 43 different factors included in the main static evaluation function.

    I found the Table view to be most useful because it is in chess terms more easily understood by human players.

    One of the interesting features is the capability to look at different individual moves in a given position, and see how the evaluation changes and the factors that caused the change(s). Simply make your move, look at the evaluation, then use the backward arrow to return to the original position. Select a new move, and examine the change in evaluation. Unfortunately, if you do this, any subsequent moves that were previously entered are lost and will have to be re-entered. There is no "stepping stone" mechanism that allows the original move sequence to be saved.

    This could be a very useful tool for analysis! I concur with your assessment of it as "astonishing!"

  9. Tempo, can I ask you when are you going to write (publish) your next article? I miss your new ideas and great discoveries (or refutations) very much... :(

  10. I was watching IM John Bartholomew's YouTube Tactics Training #3 video, first problem on Chess Tempo, and thought it is an excellent example of the PoPLoAFun approach (without being described as such).

    Link: Tactics Training #3, 58 seconds from the start

    What I found highly interesting is that he suggested Bxg7 multiple times, apparently without ever realizing that there is a cross-pin in the position, absolutely pinning the White Bishop to the White King and preventing its movement. (It may be that he considered it subconsciously and just didn't note it out loud; [SHRUG} stuff happens.) Here are some of his observations.

    (1) [1:14] "There are multiple pieces hanging. . . . There is also a Black Bishop at g7 that is undefended. So Bxg7 is a factor to be considered. However, Bxg7 would not be on our high priority list because it would lose the White Queen to bxc3."

    (2) [3:35] "If Bxg7, Black just takes the White Queen and I don't see any effective followup."

    (3) [(3:51] "Let's look at NON-forcing moves." [I personally consider a mate threat combined with a threat to capture the opponent's Queen (a duplo attack) to be VERY FORCING!]

    (4) [4:15] He finally gets around to Qg3, pinning the Black Bishop on g7 AND continuing to attack the Black Queen. He addresses that but not the fact that the White Bishop is pinned.

    At [5:10] he addresses his "textbook method" of analyzing and solving tactics problems. What is most important IMHO is that he considers Points of Pressure (PoP) and Lines of Attack (LoA) and Functions (Fun) [without using that terminology] but still overlooked the absolute pin of the White Bishop. Consider forcing moves (checks and captures) first. He finally "sees" that the Black Bishop can capture the White Bishop on d4 WITH CHECK, but does not address the previous pin on that d4 Bishop by the Black Queen. If no forcing moves are found, then consider non-forcing moves. These would be major threats like mate in 1.

    Please don't misunderstand: I think very highly of IM Bartholomew. I've watched several videos of his bullet matches with IM Daniel Rensch. They are VERY illuminating for a patzer like me. I've observed a few times (NOT VERY MANY!) when one or the other would overlook simple tactics, such as a mate in 1, in severe time trouble. Certainly they do not play like the proverbial flawless machine.

  11. I just went back to the video and read the comments, which I normally do not read at all. John, much to his credit, acknowledged that he missed the cross pin. One more thing to admire about him!

  12. IM John Bartholomew is talking and making a video while solving that puzzle that made him weaker
    He was
    1 analysing opponents last move
    2 analysing most obvious forcing moves
    3 counting material
    4 creating a list of candidate moves ( CC = check & captures )
    5 looking for non forcing moves meaning non takes and non checks but Threats

    During his CC he did gain insight in the function of the pieces and their relationships
    So he found the right move imediatly after the went to calculate threats

    In my terminology its a Tit for Tat puzzle, i take your queen then you take mine
    The standard method is to move away my weak piece (Queen) and threat something (Mate in 1) with it. The result is a double attack ( Queen Hanging & Mate in 1 )

  13. What interest me the MOST is the things none of these masters or grandmasters can (are able to) express...

    The most often reply to "why did you play this move" is - because it LOOKS good. And I am more than 100% sure they (IMs and GMs) use some advanced techniques and methods they cannot explain by themselves. It is especially visible when you can see these monsters playing bullets (1+0) and ultrabullets (30s) games. It is impossible to count most things they can recognize in such a lightning speed!

  14. I can certainly appreciate the difficulties of teaching by doing, and trying to make a useful video at the same time. The similar kind of thing happened to Bryan Castro in one of his videos. It still was very insightful to have his insights.

    I thought it was fascinating to get a glimpse into an IM's mind, and how he approaches solving problems/analyzing positions. I also thought it was interesting that he found the correct solution even though he missed the cross pin. In any event, that "miss" did not impact the solution. I think it goes to show that even IMs are human, and so there is some (faint?!?) hope for the rest of us to improve to some degree.

  15. I tend to do well with calculations, but if there is a "hole" in my patterns, I tend to miss those one or "two-movers". In that cross-pin problem, I simply missed the move Qg3, cross-pin the cross-pin is a typical theme and I missed it here.

    In problem #3 (that's as far as I've gotten) John misses why 1..Ne5 was the right solution, or to some extent he does, and then he doesn't read the Chesstempo comments either. I solved the problem on my own and then chuckled as he struggled, he was on the right track and would have ultimately solved it, but you can see that even these monster bullet players do not necessarily have or desire to have unlimited patience, even though he says at the beginning that he is more interested in solving them correctly!

    Problem #3 1...Ne5. If 2.f4 (yes then ...Qh5 is some sort of long-winded mate which I, John, and probably many others suspected, and then Stockfish confirmed for them), but instead of 2...Nf2+, 3.Kf2, a simple solution is to play 2...Qg6 (White is up a piece, and the queen trade will be enough), and if White avoids the queen trade with 3.Qc5, then 3...Nf3+ does win the exchange after all.

    He was also worried in this line 1...Ne5 about 2.Re3, but 2...Nc4 forks queen and rook. After 3.RxR+ RxR, 4.Qc5 Rc8! wins the knight on d5.

    I really liked how John admitted that the Chesstempo site is hard! In this problem, it goes beyond "tactics" training (i.e., pattern-recognition), and overlaps into solving a position (any position) OTB, which mainly consists of calculating lines deeply (this is where my tournament experience and study with chess books has helped me).

    It's fascinating to watch an IM solve tactics problems, but at the same time I also feel less insignificant than I thought I was at chess before watching this. It tells me what I suspected, but am still surprised to know, that blitz games and solving actual tough, sharp, non-technical positions, are not one and the same. That said, Masters and GMs do have loads of patterns locked away on blitz recall that Class players truly struggle to find. Also, Masters recognize the threats on the board so much more quickly, and usually someone whose rating shoots up is getting better at this type of board-recognition, particularly with kids. I think adults do well with deep calculation though, but of course chess is largely about "never scr*wing up". ;-)

  16. The player who really impressed upon me looking for long lines was Kortchnoi. Kortchnoi would look 5-10 moves deep in these really boring, bland positions where you would think no calculation was required, rather than merely comparing and selecting the best idea. These GMs have their patterns down, and they do find and select the ideas, but they are seemingly always calculating still even after this. You would think they would find the idea and then have that smug look, but no way. Watch Karjakin's face, like no expression as he finds the ideas because he is still caught up in the calculations. To John's credit, he is also concerned about getting the calculations correct, not just finding the best idea and then plopping it out there, hand held out, waiting for the chess merit badge (which is a very tempting thing to do, but which the seasoned Master does not do).

  17. In Problem #5, John wanted to play 1.Qa2, but after 1...b5 it allows time to get the knight back in the game, whether that b5 pawn is lost or not (it probably is). The correct idea after 1.Rf7 (He helped me with sensing that this was the move)...Qa8 (guarding his knight) is that White can then safely triple on the 7th rank without worrying about ...Qxf2+, and if the Black queen opposes this, then the knight will fall. Black can resign because mate is coming.


  18. Simon Williams Tactic Training #4 Rules for solving puzzles like a Grandmaster

  19. @ Aox:

    Thanks for the link. I was already working my way through IM Bartholomew's and GM Williams videos in order, and just got to that one.

    GM Williams' basic suggested "thinking process:"

    When should you search for a Tactic?

    1) When "Instinct" kicks in - Smell It!

    [Sounds like a vulture to me!]

    2) At critical stages of the game

    [How to identify one (or more) critical stages?]

    3) In complicated positions

    [How to define/determine that the current position is a "complicated" one?]

    4) When you have the initiative

    [How do you know that you have the true initiative, and not a false initiative?]

    5) There is a clear weakness in your opponent's position

    [How to identify "clearly" the appropriate weakness(es) in the current position?]

    My questions in brackets are not meant to disparage GM Williams' suggestions, but merely illustrate that a "simple" step-by-step "thinking process" is neither as "simple" nor as "clear" as it would seem, on first examination.

    When you think there may be a tactic, what should you do?

    1) Analyze 'Candidate' Moves

    [How do you determine the candidate moves?]

    2) Use the 'Process of Elimination'

    [How do you eliminate moves if neither visualization nor evaluation skills are adequate?]

    3) Double check and triple check

    [Definitely a "sin" if trying to follow the Kotovian thinking process!]

    4) If need be, look for a better move

    [Dr. Lasker, I presume?]

    GM Williams' advice is an excellent summary of what GMs know and do when calculating. He IS a GM, after all.

    The questions I have interspersed are based on an inference that everything may not be as "cut and dried" as it seems on the surface.

    There is a very funny hypothetical dialog between an amateur and GM Kotov regarding his thinking process. I can't find the exact reference at the moment, so I'll paraphrase it.

    Kotov: What were you thinking in this position?

    Amateur: The first thing I did was to make a list of candidate moves. I then selected the first candidate move and began analyzing the subsequent variations once and only once.

    Kotov: Excellent! What was the first candidate move you selected?

    Amateur: I chose 1. Rd1.

    Kotov: 1. Rd1 is NOT a candidate move.

    OOPS! Cart before the horse.

    1. HORROR!!!!!

      I finally found the source of the paraphrased Kotov dialog I gave above: it was Michael de la Maza!

      rapid chess improvement: a study plan for adult players, Chapter Three: How To Think

      "Note that this approach [his proposed "simple" 8-step thinking process] is sharply different from Kotov's thinking technique, which is probably the best known move selection method. In Think Like A Grandmaster, Kotov writes: 'All candidate moves should be identified at once and listed in one's head.' This advice is, of course, simply ridiculous for the class player. Unless a specific algorithm is given for identifying the candidate moves this is equivalent to saying: 'If there is a five-move combination that wins, make the five-move combination.' I can imagine a post-mortem conversation between Kotov and a class player:

      Class player: Here I played Ra7.

      Kotov: You should not have played that move.

      Class player: Why?

      Kotov: It is not a candidate move.

      Show me a class player who can follow Kotov's thinking technique and achieve success and I'll show you a Master masquerading as a patzer.

      Well, at least one thing "stuck" in my brain from that book!?!

      Please allow the observation without rehashing all of the pros and cons of MdlM's Seven Circles [of Hell] approach, and the allegations of "cheating" by MdlM; that has been beaten to death by various Knights Errant - mostly heat, very little light.

      Totally irrelevant: I was doing a search on MdlM, and noticed that he now has his PhD in Computer Science. Maybe the $10,000 World Open prize helped pay off his student loans.

    2. you seem to have problems with the selection of candiadate moves. GM Smirnov did break it down:

      Based on the positional features like pawnstructure, imbalances... you have a plan / strategy / todo list
      Now you simply analyse very quick every possible move from Queen downto Pawn, and check if they are possibly good according to this plan, then you calculate these candidates
      That can be done in "notime" because the position is almost identical with your last move. You need to carefully update the last 2 halfmoves and re-adjust the plans if necessary

      MdlM had his PhD before he won the price, his Thesis was about chess programming.

  20. All your questions can be answered : by experience.
    He simply take the first move he "see" and "calculate" a main line deep. No list of candidate moves first. His check and triple check is nothing but the analysis: if his main line is really the! main line.
    While doing so he gains more insights (= learn! more about the position ) so if he has to drop his first move he can chose the next candidate already better.
    He does "see" moves because of the neural network LTM (Long Term Memory = experience = learned pattern = chunks ). Many times this is already the best move.
    I am now convinced that a key skill for chess improvement is chess related learning speed. As more positions you have learned as more good moves pop up ( or more precisely, as less bad moves pop up ). As quicker you can learn as less you have to repeat already calculated lines and as more freely you can step through he variations and evaluations.

  21. @ Aox:

    You are correct, of course, regarding remembered experience as the key factor.

    I am reminded of this quote, attributed to many different people:

    Good judgement is the result of experience and experience the result of bad judgement.

    ― Mark Twain

    I'm still starring on the "bad judgement" stage as a headliner!

  22. PART I:

    Another interesting position to analyze! This is a 2400+ rated puzzle, so I felt pretty good after solving it!

    Master Level (2400-2620) Chess Tactics and Explanations! [23:16] - Infinite Flash Chess

    After 1. ... Kh7?? 2. Rf7, I "saw" two specific lines of play very quickly.

    2. ... Qh3 (to exchange off the White Queen) 3. Rxg7+ Kxg7 4. Bc3+

    4. ... Kh7 5. Qf7#
    4. ... Kg8 5. Qg6+ Kf8 6. Qg7+ Ke8 7. Qe7#
    4. ... Kf8 Now what? 5. Qg6! and White threatens the Black Rook h8 and the previous mate sequence (eventually).

    2. ... Qe8 3. Ne7 Blocking the Black Queen from checking and threatening mate on g6. Black can sacrifice the Black Bishop on f5 to open a line for checks to the White King, but those checks come to a rapid end and Black is down a piece. Sacrificing the Black Queen for either the WRf7 or WNe7 seems hopeless.

    3. ... Bf5 4. Nxf5 Qe4+ 5. Kg1 Qb1+ 6. Nf1 and the checks come to an end long enough for White to kill the Black King.

    In the video starting position, Black played 1. ... Kh7. WHY???, I wondered.

    When I started looking at the position, I couldn't figure out a good reason for such a move! It seemed to me that there should be at least one alternative that would be better for Black than that move! After solving the problem, I went back to the original position and let GM Stockfish check it. It seems my intuition about that position was correct.

    New game - Starting position (before 1. ... Kh7??)
    r1b3kr/3q2b1/p1Np3p/BP1P1RpQ/8/8/1P5N/7K b - - 0 1

    Analysis by Stockfish DD 64 SSE4.2:

    1. -+ (-2.76): 1...Bb7 2.Qf3 axb5 3.Ng4 Qe8 4.Nf6+ Bxf6 5.Rxf6 Rh7 6.Re6 Qf8 7.Qe4 Qf1+ 8.Kh2 Qf4+ 9.Qxf4 gxf4 10.Rxd6 f3 11.Bc3 Bxc6 12.dxc6 f2 13.Rg6+ Kf7 14.Rf6+ Ke7 15.Kg2 Rf7 16.Rxf7+ Kxf7 17.c7 Rc8 18.Ba5 f1R 19.Kxf1 h5 20.Ke1 Ke6
    2. -+ (-1.57): 1...Qxf5 2.Ne7+ Kh7 3.Nxf5 Bxf5 4.Nf3 Kg8 5.b6 Rf8 6.Nd2 Rh7 7.Qe2 Be5 8.Bc3 Re7 9.Bxe5 Rxe5 10.Qxa6 Rxd5 11.Qc4 Be6 12.Kg1 Rb8 13.Qb4 Kf7 14.Ne4 Rd1+ 15.Kf2 Ke7
    3. =/+ (-0.62): 1...axb5 2.Qg6 h5 3.Bc3 Rh7 4.Qxg5 b4 5.Bxb4 Ra1+ 6.Kg2 Ra7 7.Rf3 Rb7 8.Bc3 h4 9.Kh1 Rh6 10.Re3 Qf5 11.Qd8+ Kh7 12.Ne7 Rxe7 13.Qxe7 Qxd5+ 14.Nf3 Rg6 15.Bxg7 Rxg7 16.Qxh4+ Kg6 17.Kg1 Qh5 18.Qxh5+ Kxh5+ 19.Kf2 Rf7 20.Rd3 Bb7 21.b4 Be4 22.Rxd6 Rxf3+ 23.Ke2 Rb3 24.Rd4 Bc6
    4. = (0.00): 1...Rh7 2.Qg6 h5 3.Qxg5 axb5 4.Rf1 Kh8 5.Ne7 Bb7 6.Ng6+ Kg8 7.Ne7+ Kh8
    5. = (0.00): 1...Bf8 2.Qg6+ Qg7 3.Rxf8+ Kxf8 4.Qxd6+ Kg8 5.Bc3 Qf7 6.Ne7+ Kh7 7.Bxh8 axb5 8.Bd4 Ra1+ 9.Kg2 Rd1 10.Be3 Re1 11.Qe5 Re2+ 12.Kg1 Re1+ 13.Kg2
    6. +/- (1.23): 1...Ra7 2.Rf2 Kh7 3.Nf3 Qf5 4.Nxa7 Bb7 5.Nc6 Qxd5 6.Rg2 Bxc6 7.bxc6 Rf8 8.Rxg5 Qxf3+ 9.Qxf3 Rxf3 10.Rd5 Bxb2 11.c7 Rf8 12.Bb4 Rc8 13.Bxd6 Kg6 14.Kg2 Ba3 15.Bg3 Be7 16.Ra5 h5 17.Kf3
    7. +- (3.39): 1...Rb8 2.Qg6 h5 3.Rf3 Rh6 4.Qxg5 Qe8 5.Re3 Qf7 6.Bc3 Bg4 7.Qxg7+ Qxg7 8.Bxg7 Kxg7 9.Nxb8 axb5 10.Nc6 Kf6 11.Nxg4+ hxg4+ 12.Kg2 Kg5 13.Rb3 Rh5 14.Nd8 Kf5 15.Rxb5 Rh7 16.Ra5 Ke4 17.Ne6 Rb7 18.Kg3 Rxb2 19.Kxg4 Rg2+ 20.Kh5 Rd2
    *** 8. +- (5.91): 1...Kh7 2.Rf7 Qh3 3.Rxg7+ Kxg7 4.Bc3+ Qxc3 5.bxc3 *** Bf5 6.b6 Rae8 7.Qd1 Rhf8 8.b7 a5 9.Kg1 a4 10.Qd4+ Kg8 11.Qa7 Bh3 12.b8Q Re1+ 13.Nf1 Rxb8 14.Qxb8+ Kg7 15.Qc7+ Kg6 16.Qxd6+ Kh5 17.Kf2 Rxf1+ 18.Ke3 Bf5 19.c4 Re1+ 20.Kd4 Re4+
    9. +- (5.95): 1...Qb7 2.Rf1 Kh7 3.Rf7 Qxf7 4.Qxf7 Rf8 5.Qc7 axb5 6.Bc3 Rg8 7.Ne7 Bxc3 8.Nxg8+ Kxg8 9.bxc3 Ra1+ 10.Kg2 Bf5 11.Qxd6 Kg7 12.Qe5+ Kg6 13.Nf3 Ra2+ 14.Kg3 Rc2 15.d6 h5 16.Qe8+ Kh6 17.Qh8+
    10. +- (6.64): 1...g4 2.Nxg4 Ra7 3.Nf6+ Bxf6 4.Rxf6 Qh3+ 5.Qxh3 Bxh3 6.Nxa7 axb5 7.Nxb5 Rh7 8.Nxd6 Rd7 9.Kh2 Bg4 10.Nf5 Bd1 11.Nxh6+ Kh7 12.d6 Bh5 13.Nf5 Bd1 14.Bc3 Ba4 15.Kg3 Bb3 16.Kf4 Kg8 17.Nd4 Bf7 18.Ke5 Bh5
    11. +- (7.07): 1...Qxc6 2.Qf7+ Kh7 3.bxc6 Rf8 4.Qxf8 Bxf8 5.Rxf8 Kg7 6.Rd8 Kf7 7.Bc7 Ke7 8.Rh8 h5 9.Bd8+ Kf7 10.Bxg5 a5 11.Bf4 Ke7 12.Rxh5 Ba6 13.Rh7+ Kf6 14.Bxd6 Re8 15.Ra7 Bd3 16.Ng4+ Kf5 17.Nf2 Bf1 18.Rxa5 Rg8 19.Be7 Bc4 20.d6+ Ke6

    (Coble, Asheboro, NC 27.05.2017)

    1. The ratings at tactics trainer are a several hundred points higher than chesstempos. My TT ratings was >2500 while my CT-Blitz rating is ~1900

    2. @ Aox:

      Thanks for the information!

      I had no idea of the differences in rating between Tactics Trainer and Chess Tempo. The Chess Tempo brain twisters that Temposchlucker has used for examples are usually in the range 1900-2100. Those are some of the most difficult problems I have ever tried solving. Assuming that the problem above is equivalent to Chess Tempo 1900 (based on your description of the difference in rating systems), I'm still pleased that I could solve it. That feeling was related to solving it - PERIOD, not that solving it implied anything about MY rating/playing/skill level.

      I am under no illusion that I have suddenly gained ANY expertise at the FIDE/Elo 2400 level. I would be more than happy to be able to play OTB at my last official rating level of USCF 1810. I plan on returning to the local chess club for OTB competition starting next week. I want to find out if I have really "learned" the lessons available here.

    3. i think the best method to test your tactical skills is chesstempo either in blitz or in mixed mode.
      OTB games are influenced by so many factors, for example opening theorie.

    4. @ Aox:

      I decided to give your suggestion a try today. In 13 minutes, I raised my Chess Tempo blitz rating from 1435 to 1593. (It actually went above that but I haven't figured out how to exit Tactics mode without losing the last problem.) I only missed a few problems along the way. That was fun!


  23. PART II:

    After 1. ... Kh7?? 2. Rf7 Qe8 3. Ne7 I judged the position as totally winning for White. That also appears to be the correct assessment.

    New game
    r1b1q2r/4NRbk/p2p3p/BP1P2pQ/8/8/1P5N/7K b - - 0 1

    Analysis by Stockfish DD 64 SSE4.2:

    1. +- (23.87): 3...Bf5 4.Nxf5 Qe4+ 5.Kg1 Qxf5 6.Rxf5 Raf8 7.Rxg5 Bd4+ 8.Kg2 Rhg8 9.Ng4 Rg7 10.Qxh6+ Kg8 11.Qe6+ Kh8 12.Rh5+ Rh7 13.Rxh7+ Kxh7 14.Qe7+ Kg6 15.Qxf8 axb5 16.Qxd6+ Kg7 17.Qd7+ Kg8 18.Bc3
    2. +- (#10): 3...Qxe7 4.Rxe7 Bf5 5.Ng4 Bxg4 6.Bc3 Rhg8 7.Qxg4 Raf8 8.Qe4+ Kh8 9.Rxg7 Rxg7 10.Qg6 Rf1+ 11.Kg2 Rf2+ 12.Kxf2 axb5 13.Qxg7#
    3. +- (#5): 3...Qxf7 4.Qxf7 Rf8 5.Qg6+ Kh8 6.Bc3 Rf6 7.Bxf6 Bxf6 8.Qxh6#
    4. +- (#4): 3...Rg8 4.Qg6+ Kh8 5.Bc3 Bf5 6.Bxg7+ Rxg7 7.Qxg7#
    5. +- (#2): 3...Rf8 4.Qg6+ Kh8 5.Qxg7#
    6. +- (#1): 3...g4 4.Qg6#
    7. +- (#1): 3...axb5 4.Qg6#
    8. +- (#1): 3...Bh3 4.Qg6#
    9. +- (#1): 3...Bg4 4.Qg6#
    10. +- (#1): 3...Be6 4.Qg6#
    11. +- (#1): 3...Bb7 4.Qg6#

    (Coble, Asheboro, NC 27.05.2017)

  24. You need only one candidate: the best one. Only when that one fails, you need the next one.

    A lot of those thought processeshave in common that there is a lot of redundancy in the method. Just to give them the feel of completeness. But that redundancy renders every thought process as useless. We need to avoid every redundancy we can. That means that we must tailor our method towards our own mind. We already master a whole bunch of unconscious methods. We only need to trigger them at the right time, and ignore them when not needed. If you never fail to see a pin, you don't need a reminder for pins in your method. That means that the method becomes highly personal.

    A main problem is that our conscious thinking works in a serial way, while we often need to process items parallel. One way to process items parallel, is by the vultures view. Our thought process must combine the vultures view in an intelligent way with serial thinking. You must first decide in which room your car keys may lay, before you start searching.

    The vultures view is based on the fact that moves that are chained together over the serial coarse of time, form a geometrical pattern that can be seen at once.

  25. Blacks last move shows overprotection 3 times that failed.

    1. It is interesting to read the various comments, since this was a "problem of the day" (POTD). The commenters' focus appears to be primarily on the back rank mate as the leitmotif, with the various themes/devices referenced as contributing factors. However, if the PoPLoAFun approach is used, the solution (Black's "surprise" move) becomes "obvious" (in a certain sense): the b2 Pawn is pinned and cannot capture on a3, and neither the White Queen nor the White Rook can recapture on a3 without allowing a back rank mate. Two Black pieces are aimed directly at the back rank, with one direct (WRa1) and one indirect (WQd3) defensive piece protecting it. In addition to considering this to be an example of inadequate "overprotection," one could consider it to be a B.A.D. case (2:2) with the back rank being the "Barely Adequately Defended) entity. As James Carvill said, "When your opponent is drowning, throw the SOB an anvil." The "anvil" is 24. ... Rxa3!!

      After 23. ... Qe5 Black is threatening the b2 Pawn and setting up the back rank mate threat. The question becomes: how can White defend against both threats with a duplo defense? 24. Qd4 seems to solve this conundrum. If Black checks on e1, then the White Queen can return to g1 because the White f Pawn is missing. This also blocks the attack on the b2 Pawn. Black's doubled b Pawns aren't going away anytime soon, so there is no need to attack them immediately. An exchange of Queens leaves White in a slightly better or (at worst) an even position. (Thanks to GM Stockfish for different ideas!)

      It seems that the normal GM response to a duplo threat is to try for a counter threat (Rxb6) combined with a defense (Rook protects b2). In most cases, that seems to be a good approach; not in this specific case.