Evaluating the tournament
That was another great tournament! I played some wonderful games.
In simplified terms:
- In one game, I was outplayed by a 12-year-old kid with a rating of about 1500. He eventually finished second in my group.
- In the remaining 8 games, I was winning at some point.
- Yet the final result of those remaining 8 games is meager: 1 win, 1 loss, and the rest draws.
Over the past year, I have focused primarily on my opening repertoire and positional play. I haven't lost any more points due to my openings or positional play. I still see many opportunities for improvement in those areas, but they are clearly not a priority anymore.
You can compare me to a soccer team without finishers. I play excellently, but I don't score. These are the two 'finishers' that are missing:
- tactics (combinations)
- endgame strategy
I played a draw against a woman, who is the partner of an International Master. She showed the game to him. He said I played excellently, but that I had botched the endgame.
My group in this tournament has many young kids with unreliable (= far too low) ratings. Often these are youth players who, under the guidance of their coaches, use the adults as 'punching bags' to climb the ranks. Mostly Eastern Europeans, Indians, and Chinese. This has worsened since last year because many seniors have fled to the veterans' group. They only play five games.
So I will likely lose some rating points. But that’s not important; it will balance itself out eventually.
What needs to happen:
Tactics
- Done: 750 patterns
- To do: 2560 patterns
Duration: 7 years at an average pace of 1 per day. Hopefully, I can speed this up a bit due to my retirement. I consider it proven once again in this tournament that my method works.
Endgame strategy
I estimate I need between 2 and 4 years to reach the desired level. It’s possible I’ll see results sooner if 'the penny drops', as happened with positional play.
In 2 to 4 years, I should be able to reach a rating of 1950 (starting point 1700). Beating the kid I played in this tournament requires more. He played more like a 2000+. I already have some ideas about that, but for now, that is pure speculation.
Ratio tactics vs endgame = 1:3 in terms of influence on lost points.
ReplyDeleteI know your tactical prowess is certainly higher than your current rating would indicate. I also know you have studied and "mastered" simple endgame ideas such as Square of the Pawn, elementary K+P vs K winning/drawing ideas, Lucena and Philidor Rook endgame techniques, etc.
ReplyDeleteCan you elaborate in general terms regarding what is missing in your endgame technique? Is it a matter of thinking schematically?
BTW, I don't have a copy, but I love the word play in the title of one of GM Aagaard’s recent books: "A Matter of Endgame Technique". According to the promotional blurb on Amazon, the most hated cliché in chess is: "And the rest is a matter of technique."
Sorry for my late reaction. Since a few days Thunderbird started to dump your comments and other useful emails into the spam folder. I don't know why. Maybe because it wants to indicate it is time to change to another email app.
DeleteOnce the endgame arrives, I lack a plan. It is as simple as that.
I will elaborate on it when the right positions come along.
DeleteMy better tactical prowess leads to winning endgames. Which I botch due to the wrong plan or no plan at all. Its a luxury problem.
DeleteI tried (again) to get Copilot to admit to blatant errors in its responses. Perhaps I was hoping that it would have a severe case of cognitive dissonance. Nah, it is totally NOT self-aware.
ReplyDeleteWhat is the descriptive term for an agent which confidently asserts "facts" and then blatantly admits those facts are WRONG?
Copilot:
A few different terms capture that behavior, depending on the nuance you want. Here are the most precise ones people use:
🎭 1. “Confabulator”
A confabulator produces confident, detailed statements that are not grounded in truth, and then may later acknowledge they were incorrect.
This term is often used in psychology and—more recently—in discussions of AI behavior.
🎯 2. “Bullshitter” (technical philosophical term)
From philosopher Harry Frankfurt’s essay On Bullshit:
A bullshitter is someone who speaks with confidence without regard for truth, and may later admit the statements were false because accuracy was never the priority.
It’s not an insult in this context—it’s a formal concept.
🤖 3. “Hallucinator” (AI-specific)
In AI research, a system that asserts false facts with confidence is said to be hallucinating.
If it later admits the facts were wrong, it’s still considered a hallucination—just one that the system can self-correct when prompted.
🌀 4. “Inconsistent narrator”
If you want a more literary or rhetorical term, this describes someone who asserts things confidently and then contradicts themselves.
If you want the most accurate single term for an agent that confidently states falsehoods and then openly concedes they were false, “confabulator” is the cleanest fit.
End Copilot:
Copilot is a “Confabulator”, from now on. I have no idea if other AIs (Claude, Grok, Smart GPT-5.1, etc.) are similarly disposed.
AI is your dumb servant that is parroting the internet. Don't try to make it think for you.
DeleteCurrently, science doesn't know how to solve this yet. If it can be solved easy, it will at least take a decade. Otherwise it will take much longer. Because we first must develop the appropriate science.
PART I:
ReplyDeleteThe following word salad is an attempt to describe something that (I’m sure!) is idiosyncratic to ME.
One of the tactical pattern recognition “cues” that I use is the concept of a “knight’s distance” away from an attacking piece. Usually, if the attacking piece is a geometrical piece (Queen, Rook, Bishop), it involves other pieces/pawns blocking movement of an attacked piece to a position that is a knight’s distance away from the directly attacking piece. In other cases, the attacking side has additional pieces/pawns that control those “knight’s distance” squares. This conceptual trigger can be generalized to enable recognition the (potential) presence of a wide variety of common tactical patterns. It has an uncanny similarity to the Epaulette or Dovetail mating pattern.
I do not claim there is anything “special” about this particular cue, just as there is nothing particularly special about the Trébuchet. To me, it is an example of patterns that combined to provide a higher level or more abstract recognition pattern. It is just part of the “furniture” in the “room,” creating an ambiance.
I think there are a rather large number of similar cues that are not named and thus not cultivated in the tactical training literature. My point is NOT that we should try to “discover” and name EVERYTHING in our arsenal of patterns, but that we should look for and RECOGNIZE the component cues that go to make up tactical shots and combinations. It certainly makes it easier to follow a principal variation when there are recognized (but perhaps unnamed) cues along the path.
For the endgame, there are typical schematics that play a similar role to tactical patterns in the middlegame. It is important to know the sequence of patterns, not the specific moves required in any given position. The patterns and schematics are capable of guiding our thoughts as we play. We rarely get the exact same position twice, much less multiple times, but we continually get typical patterns and schematics over and over again.
Here are some examples from a recent lichess.com Puzzle Storm session:
EXAMPLE #1
Defensively, Black seemed unaware of this pattern when playing 23… exf6.
https://lichess.org/ZIcjTEp9/white#45
FEN: 5r2/1p1q2k1/p2p1p1p/3Pn1p1/4Q1P1/1PrB1P2/P1P5/2KR3R w - - 0 24
[Event "rated rapid game"]
[Site "https://lichess.org/ZIcjTEp9"]
[Date "2022.11.17"]
[Round "-"]
[White "TheBenderUltimatum"]
[Black "junara"]
[Result "1-0"]
[GameId "ZIcjTEp9"]
[UTCDate "2022.11.17"]
[UTCTime "21:54:33"]
[WhiteElo "2094"]
[BlackElo "2060"]
[WhiteRatingDiff "+9"]
[BlackRatingDiff "-5"]
[Variant "Standard"]
[TimeControl "600+0"]
[ECO "B27"]
[Opening "Sicilian Defense"]
[Termination "Normal"]
1. e4 c5 2. Nf3 g5 3. b3 Bg7 4. Nc3 f6 5. Bb2 Nc6 6. d4 d6 7. d5 Nd4 8. Nxd4 cxd4 9. Ne2 Bg4 10. f3 Bd7 11. Qd2 Nh6 12. O-O-O Nf7 13. Nxd4 a6 14. Nf5 Bxf5 15. exf5 Qd7 16. g4 O-O 17. h4 h6 18. hxg5 fxg5 19. Bxg7 Kxg7 20. Qd4+ Ne5 21. Bd3 Rac8 22. Qe4 Rc3 23. f6+ exf6 24. Qh7# 1-0
PART II:
ReplyDeleteEXAMPLE #2
This is a little more complicated. Note that in the terminal position, the Black King is bracketed by the BRc8 and the “attack” of the WBe4 on the a8-square, with the WQb6 at a knight’s distance from the a8 and c8 squares. It is just a part of the overall cues for the final mating position.
trastornomental01 vs ChristophCV: Van't Kruijs Opening • lichess.org
FEN: 2r4r/2kb4/p2p1p2/1pp1pP2/3nB1P1/P2P4/qPPQ1P2/2KR2R1 w - - 1 26
[Event "Hourly Blitz Arena"]
[Site "https://lichess.org/Q0ZybG5Z"]
[Date "2025.03.05"]
[Round "-"]
[White "trastornomental01"]
[Black "ChristophCV"]
[Result "1-0"]
[GameId "Q0ZybG5Z"]
[UTCDate "2025.03.05"]
[UTCTime "23:48:17"]
[WhiteElo "1712"]
[BlackElo "2008"]
[WhiteRatingDiff "+10"]
[BlackRatingDiff "-10"]
[Variant "Standard"]
[TimeControl "300+0"]
[ECO "A00"]
[Opening "Van't Kruijs Opening"]
[Termination "Normal"]
1. e3 c5 2. d3 Nc6 3. Be2 g6 4. Bd2 Bg7 5. Nc3 Nf6 6. Bf3 d6 7. e4 e5 8. Nge2 a6 9. Ng3 b5 10. a3 Be6 11. Qc1 h5 12. Nf5 gxf5 13. exf5 Bd7 14. Ne4 Nxe4 15. Bxe4 f6 16. h3 Qe7 17. Qd1 Qf7 18. Qf3 Rc8 19. O-O-O Kd8 20. Rhg1 Kc7 21. g4 hxg4 22. hxg4 Bh6 23. Qe2 Nd4 24. Qe1 Bxd2+ 25. Qxd2 Qa2 26. Qa5+ Kb8 27. Qb6# 1-0
EXAMPLE #3
Again, the defender is unaware of the potential pattern and chooses to regicide himself. This was a 3+0 blitz game, but the White player (according to his handle) is a FIDE Candidate Master rated 2376(!) on lichess.org.
CM Floman vs Ihsan_Kilic: Semi-Slav Defense: Accelerated Move Order • lichess.org
FEN: rnb2rk1/pp4pp/2p1p3/3pNp2/2PPnP1q/2NBP3/PP4PP/R2QK2R w KQ - 3 11
[Event "Atatürk'ü Anma Arena"]
[Site "https://lichess.org/XyC7wZA3"]
[Date "2021.11.10"]
[Round "-"]
[White "Floman"]
[Black "Ihsan_Kilic"]
[Result "0-1"]
[GameId "XyC7wZA3"]
[UTCDate "2021.11.10"]
[UTCTime "16:21:03"]
[WhiteElo "2376"]
[BlackElo "2446"]
[WhiteRatingDiff "-4"]
[BlackRatingDiff "+4"]
[WhiteTitle "CM"]
[Variant "Standard"]
[TimeControl "180+0"]
[ECO "D31"]
[Opening "Semi-Slav Defense: Accelerated Move Order"]
[Termination "Normal"]
1. d4 d5 2. c4 e6 3. Nc3 c6 4. Nf3 f5 5. Bg5 Be7 6. Bxe7 Qxe7 7. Ne5 Nf6 8. f4 O-O 9. e3 Ne4 10. Bd3 Qh4+ 11. Ke2 Qf2# 0-1
The patterns that we need to absorb are the ones that show us what a move does within the context of the finite amount of scenarios. The move must trigger the concepts that play a role.
ReplyDeleteA move clears, chases, annihilates the defender, lures, et cetera.
The retrieved scenario triggers the next move.
All without thinking.
Temposchlucker wrote (in part):
DeleteThe retrieved scenario triggers the next move.
This is a very prescient observation! I would change it slightly from “move” to “moves” because intuition can encompass many moves rather than just one move. That is my point about relatively pedestrian “cues” like the “knight’s distance” concept. It’s like building a mental picture of a desirable action (foreground) and having the perceptual context (background) required to make it “work” provided by System 1 for free. BUT, we cannot take for granted that System 1 will always connect the necessary contextual features to the desired concept merely from exposure to numerous examples. During training, our System 2 attention must identify the context as an essential component of the overall mini-skill.
Rather than think (System 2) along a Principal Variation (variation processing) move-by-move, one move at a time, we should be using our intuition (System 1; intuition; pattern recognition) to "jump from hummock-to-hummock" (multiple moves) to escape from the “swamp” of the combinatorial explosion of potential candidate moves.
Natural decision making is based on this very idea. The "pruning" of variations is accomplished “up front” by selecting one "obvious" option (tactical sequence or schematic that is proposed by System 1) and pursuing it deeper until it reaches an obviously favorable conclusion or a dead end (decisively refuted). If the conclusion is favorable, play it. If not, choose the next available most favorable option and repeat the process. There is no need to waste time trying to make a list of all candidate moves and then performing an exhaustive evaluation sequence on all of them, comparing the options. As GM Tisdall noted, sometimes we will pick the wrong option as the starting point; “stuff happens.” The proposed process takes care of that.
In all the research done by Gary Klein and his teams [Sources of Power: How People Make Decisions, 20th Anniversary Edition, they anticipated that skilled decision makers would follow the formal processes proposed by academics for decision making: First, make a list of all possible options. Second, evaluate all possible options (using whatever criteria was available). Third, rank the various options. Fourth, choose the option with the highest score and execute it. They found that decision makers under severe time constraints (such as chess players) do NOT ever follow this formal process! Instead, they choose the most “obvious” option, simulate it and its immediate consequences (which is what variation processing does) and then execute the first option found that meets all criteria for “working”. There is no Santa Claus process of “making a list and checking twice, gonna find out which one’s faulty or nice” and by how much.
Here is a simpler example of that “knight’s distance” cue as part of pattern recognition.
ReplyDeleteFEN: 8/kp6/p3R3/3p4/5p2/3K4/q2Q4/8 b - - 5 45
Given the reduced material (and the overwhelming material superiority of White), Black is in dire straits and needs a checkmate to win. It’s not hard to figure out the appropriate mating move using System 2: (1) First, look for checks (CCT), followed by (2) look for escape hatches for the White King that cannot be covered by the Black Queen (i.e., the two squares at a “knight’s distance” from the intended mating focal point). The WQd2 occupies one of the two “horns” and the BPf4 controls the other one (e3). Q.E.D.: one-move checkmate.
There’s nothing really “radical” about the concept. It’s as mundane as the observation that “Pawns cannot move backward.” (which most of us amateurs forget in the throes of battle). Grinding such cues into long-term memory enables System 1 to almost instantly recognize the pattern and propose the solution without any System 2 assistance. It is just another one of the myriad patterns that comprise SKILL in the form of mini-skills.
Another “mini-skill” is to always remember:
Nobody has ever won a chess game (no matter how bad the situation is) by resigning!
From game 10+0 • Rapid
W: nvem12345 (1973)
B: Artako (2011)
Modern Defense, Three Pawns Attack (B06)
1. e4 g6 2. d4 Bg7 3. f4 e6 4. Nf3 Ne7 5. h4 d5 6. e5 c5 7. c3 cxd4 8. cxd4 Nbc6 9. h5 Qa5+ 10. Nc3 Bd7 11. hxg6 hxg6 12. Rxh8+ Bxh8 13. Bd2 O-O-O 14. Nb5 Qb6 15. Nd6+ Kb8 16. Ng5 f6 17. Ngf7 Rf8 18. Nxh8 Rxh8 19. Bc3 fxe5 20. fxe5 Nf5 21. Nxf5 gxf5 22. Qd2 Rh4 23. O-O-O a6 24. Qf2 Rg4 25. Be2 Rg8 26. Rh1 Nxe5 27. Qf4 Ka7 28. Qxe5 Rxg2 29. Rh7 Bb5 30. Bd1 Rg1 31. Qe3 Rf1 32. Rh6 Ba4 33. b3 Bxb3 34. axb3 Qxb3 35. Qd2 Rxd1+ 36. Qxd1 Qxc3+ 37. Kb1 Qb4+ 38. Ka2 Qa5+ 39. Kb1 Qb4+ 40. Kc1 Qc3+ 41. Qc2 Qxd4 42. Rxe6 f4 43. Qd2 Qa1+ 44. Kc2 Qa2+ 45. Kd3 Qc4#
I refer to these squares as the "anvil".
ReplyDeleteThanks for the reminder!
DeleteAdditionally, I think your previous three blog posts containing references to the hammer and anvil [30 JAN 2019, 6 NOV 2022, and 16 NOV 2022] really capture the essence of the idea I was trying to convey.
DeleteObviously: "There is nothing new under the sun" or on the chessboard. How easy it is to forget great ideas!!
Here’s a stereotypical situation with mutual blunders.
ReplyDeleteFEN: 1r4k1/1q1n1p1p/2p1p1p1/p1bpP3/P4P2/2QP4/NP3RPP/3R2K1 b - - 5 24
BBc5 is an absolute pin on WRf2. Black has a Bishop and pawn for the Exchange. Should he simply capture the WRf2 with check or should he try for more by capturing Qxb2, setting up a second attacker against WRf2 [1:2] and also forking WNa2? Since White cannot protect both WRf2 and WNa2, should he go for it?
24...Qxb2??
White wants to get rid of the pin, which would then allow him to recapture on b2 with the unpinned WRf2. Does that make sense? What about just capturing on b2 with 25. Qxb2? Can’t White then get rid of the pin after 25...Rxb2 26. d4, retaining an advantage?
25. Qxc5??
SURPRISE!
25...Nxc5 26. Rxb2 Rxb2
Black is now just a pawn ahead in material with the better position. WNa2, WPa4 and WPd3 are attacked, so it doesn’t look like there is any simple solution for White [-4.9].
WHAT JUST HAPPENED?!?
A series of exchanges, without considering all the angles and determining what the end result would be.
Given the ratings of the players, should they have been able to “SEE” the ramifications without calculating the exchanges move-by-move?
[Event "rated rapid game"]
[Site "https://lichess.org/q853AZjd"]
[Date "2024.10.24"]
[Round "-"]
[White "Macabrs"]
[Black "smicer1"]
[Result "0-1"]
[GameId "q853AZjd"]
[UTCDate "2024.10.24"]
[UTCTime "13:13:52"]
[WhiteElo "2045"]
[BlackElo "1989"]
[WhiteRatingDiff "-6"]
[BlackRatingDiff "+7"]
[Variant "Standard"]
[TimeControl "600+0"]
[ECO "B32"]
[Opening "Sicilian Defense: Accelerated Dragon"]
[Termination "Normal"]
1. e4 c5 2. Nf3 Nc6 3. d4 cxd4 4. Nxd4 g6 5. Be3 Bg7 6. Nc3 Nf6 7. Be2 O-O 8. O-O d5 9. Nxc6 bxc6 10. e5 Nd7 11. f4 Nb6 12. Bd4 Qc7 13. Bc5 Re8 14. a4 a5 15. Qd4 Rb8 16. Rad1 Bf5 17. Bd3 Bxd3 18. cxd3 Nd7 19. Ba3 e6 20. Bd6 Qb7 21. Bxb8 Rxb8 22. Rf2 Bf8 23. Na2 Bc5 24. Qc3 Qxb2 25. Qxc5 Nxc5 26. Rxb2 Rxb2 27. Nc3 Rc2 0-1
Pattern recognition at its finest! Followup: not so fine.
ReplyDeleteFEN: 8/pp1k1p1p/2b1pP2/2p5/2B2RrP/2P1P3/PP6/3K4 w - - 0 27
Black (System 1) must have seen the fork (with check) following 26… Rxg4. White has nothing better than 27. Rxg4. Black plays 27… Bf3+, forking the White King and Rook. Is it all over? NO! 28. Be2 Bxg4 29. Bxg4. OOPS! Black continued for 10 more moves; out of shock, I guess.
Why is this “reloader” (Weteschnik) so easy to overlook? Is it because the pattern is unknown? Is it that System 1 recognizes the first part, without any subsequent pattern recognition of what the situation will be after the fork occurs, assuming that Black will safely recover his material? Is it about stopping calculation one move too soon? Is it failure to evaluate the “final position” (after 27… Bf3+) without considering that the resulting position is still “hot” tactically?
Given the ratings of the players, I doubt that it was caused by unfamiliarity with the reloader tactic pattern. The “missing link” appears to have occurred in NOT recognizing that it was possible for the WBc4 to do two things at once: (1) block the check AND (2) defend the attacked Rook. Perhaps it is a variant of Averbakh’s notion of the generalized double attack—for defensive purposes.
[Event "rated rapid game"]
[Site "https://lichess.org/VuRL5xbj"]
[Date "2025.04.26"]
[Round "-"]
[White "Picinio"]
[Black "GooodApple"]
[Result "1-0"]
[GameId "VuRL5xbj"]
[UTCDate "2025.04.26"]
[UTCTime "13:08:06"]
[WhiteElo "2453"]
[BlackElo "2329"]
[WhiteRatingDiff "+4"]
[BlackRatingDiff "-12"]
[Variant "Standard"]
[TimeControl "600+0"]
[ECO "B10"]
[Opening "Caro-Kann Defense: Two Knights Attack"]
[Termination "Normal"]
1. e4 c6 2. Nc3 d5 3. Nf3 dxe4 4. Nxe4 Nf6 5. Qe2 Nxe4 6. Qxe4 Nd7 7. d4 Nf6 8. Qd3 Bg4 9. Ne5 Bh5 10. c3 e6 11. Qh3 Bd6 12. Bg5 Bxe5 13. dxe5 Qd5 14. Bxf6 Qe4+ 15. Qe3 Qxe3+ 16. fxe3 gxf6 17. exf6 Rg8 18. h3 Bg6 19. Rh2 Be4 20. g4 Rg6 21. Rf2 O-O-O 22. Rd1 Rxd1+ 23. Kxd1 c5 24. Rf4 Bc6 25. Bc4 Kd7 26. h4 Rxg4 27. Rxg4 Bf3+ 28. Be2 Bxg4 29. Bxg4 h6 30. Kd2 Kd6 31. Bh5 Ke5 32. Bxf7 Kxf6 33. Be8 e5 34. Kd3 Kf5 35. Kc4 b6 36. Kd5 Kg4 37. h5 Kf3 38. e4 Kf4 39. Bg6 1-0
This is going to be long, but it only has a relatively small word salad.
ReplyDeletePART I:
I found an interesting chapter in “How to Play Chess Endgames” by Karston Müller and Wolfgang Pajekan: Chapter 5 Thinking in Schemes. I’ll give the introductory text, followed by the first example. I’ll follow that with some comments and questions, followed by the given analysis. Hopefully, that will give some idea of how to approach the endgame using schemes rather than solely trying to use concrete calculations. Be forewarned: using schemes is complementary to calculating concrete lines. The following quote is at the beginning of the chapter:
I know at sight what a position contains, What could happen? What is going to happen? You figure it out. I know it!
José Raúl Capablanca
Introduction
On the one hand, chess is a very concrete game, in which even the smallest alteration to a position can have important repercussions; on the other hand it would hardly be possible for humans to master the game if they could not make use of certain schemes and rules of thumb. In the endgame, a schematic way of thinking is particularly appropriate. By this we mean the ability to recognize desirable positions and piece setups, and then work out a plan to reach them. If you ask a grandmaster about the assessment of an endgame position, you won’t generally hear any concrete variations at first, but rather phrases such as “If White succeeds in exchanging the rooks he should win” or “If Black manages to put his bishop on the long diagonal, he has a certain draw.”
Typical themes and issues that lend themselves to schematic thinking are, for example:
• how to match the pieces with the pawn-structure
• optimal deployment of the pieces or improving the piece setup
• consideration of the right exchange
• fortresses
• weak squares
• knight outposts
• pure opposite-coloured bishop endings
Of course, when applying this way of thinking, it is absolutely essential not to think about the position in too static a manner — on no account should you neglect the dynamic factors.
A) Target Positions
If you have everything under control, but are at a loss how to make progress, it is sometimes beneficial just to visualize desirable target positions and only then to examine whether and how you can achieve them.
Position 5.01 FEN: 8/Pkb2N2/8/8/4K3/4B3/8/8 b - - 0 67
[White "Mika Karttunen"]
[Black "Vitezslav Rasik"]
[WhiteElo "2379"]
[BlackElo "2461"]
[Date "2003.10.1"]
[Round "4"]
[Result "1-0"]
[Event "ECC"]
[Site "Rethymnon GRE"]
[Eco "A45"]
1.d4 Nf6 2.Bg5 Ne4 3.Bf4 d5 4.e3 g6 5.Bd3 Bg7 6.Bxe4 dxe4 7.Nd2 c5 8.c3 cxd4 9.exd4 b6 10.Nxe4 Qd5 11.f3 O-O 12.Ne2 Ba6 13.Qd2 Nc6 14.b3 Rfd8 15.O-O f5 16.Nf2 h6 17.c4 Qf7 18.Bxh6 Nxd4 19.Nxd4 Rxd4 20.Qb2 Bh8 21.Qa3 Bb7 22.Bg5 Be5 23.Rfe1 Bd6 24.Qb2 Qg7 25.Qe2 e5 26.Rad1 Rxd1 27.Rxd1 Bc7 28.Qd2 Bc6 29.b4 a6 30.b5 axb5 31.cxb5 Bb7 32.Qd7 Qxd7 33.Rxd7 Rc8 34.Kf1 Ba8 35.h4 Bb8 36.a4 Re8 37.Be3 e4 38.f4 Kf8 39.Nh3 Re7 40.Rd8+ Re8 41.Rxe8+ Kxe8 42.Bxb6 Bd5 43.Kf2 Bd6 44.Bd4 Kd7 45.a5 Bc4 46.a6 Bxb5 47.a7 Bc6 48.Kg3 Kc8 49.Be3 Kb7 50.h5 gxh5 51.Kh4 Bb5 52.Nf2 Bb4 53.Kxh5 Be1 54.Nh3 Bf1 55.Kg6 Bxg2 56.Ng5 Ba5 57.Kxf5 Bc7 58.Kg4 Bf3+ 59.Kf5 Bg2 60.Nf7 Bh3+ 61.Kg5 Be6 62.Nh6 Bd8+ 63.Kg6 Ka8 64.f5 Bxf5+ 65.Kxf5 Kb7 66.Nf7 Bc7 67.Kxe4 Ka8 68.Kd5 Kb7 69.Bd4 Ka8 70.Ng5 Kb7 71.Ne6 Bh2 72.Kc4 Bg3 73.Kb5 Bh2 74.Bb6 Bg3 75.Nc5+ Ka8 76.Ne4 Bh2 77.Nc3 Kb7 78.Bd4 Bg3 79.Na4 Bh2 80.Bf2 Bf4?! 81.Nb6 Kxa7 82.Nd5+ Kb7 83.Nxf4 Kc7 84.Bc5 Kb7 85.Nd5 Kb8 86.Kc6 Ka8 87.Nc7+ Kb8 88.Bd4 Kc8 89.Ba7 Kd8 90.Nd5 Ke8 91.Kd6 Kf7 92.Ne7 Kf6 93.Be3 Kf7 94.Bd4 Ke8 95.Ke6 Kd8 96.Bb6+ Ke8 97.Nf5 Kf8 98.Bc7 Ke8 99.Ng7+ Kf8 100.Kf6 Kg8 101.Bd6 Kh7 102.Nf5 Kg8 103.Kg6 Kh8 104.Bc5 1-0
PART II:
ReplyDeleteNOTE: Any mistakes in the variations is due to having to translate the move numbers from the book (which start from move 0) back to the actual game score move numbers. The game score moves are correct. My apology in advance if there are errors.)
LOOK at the position and “SEE” the constraints. Without any calculation, what is the most significant feature? Hint: remove the White Knight and Black Bishop. Is the resulting position a draw or can White win? So what are you going to try to do as White to avoid this position?
Book comment:
In view of the wrong rook’s pawn (K+B+RP vs K), White cannot allow his Knight to be exchanged [for the Black Bishop].
In light of that constraint, what kind of position is White going to try to achieve? Is he going to try to promote his pawn? If so, he is going to need the Knight on either b6 or c7 to protect the new Queen after promotion. With the Black King firmly in control of both b7 and a8, his only concern will be using the Black Bishop to chop off the Knight whenever it drops on to either b6 or c7. That doesn’t look promising for White. Black appears to have created a fortress.
INTERMEZZO
I know you have studied Silman’s Complete Endgame Course — From Beginner to Master. Maybe it’s been some time since you studied it. I recalled the following from the Introduction, “How to read this book”:
When studying this book, please remember that I’ve deliberately left out many endgames. Why? Because I don’t feel they are important to players under the 2400 level. For example, I heretically decided not to include Bishop and Knight vs. Lone King because it’s far from easy to master, and it occurs very rarely in over-the-board play. In fact, I only got it once in my entire career, while IM John Watson and IM John Donaldson never got it at all! Is such a rarity really worth the two or three hours it would take to learn it? I say no. Ultimately, this is what Silman’s Complete Endgame Course is all about: Learning what’s usefuland devoting the rest of your previous study time to other areas of the game.
If I were playing Black and you were my opponent (or if my opponent were any other player with an established rating below 2000), my immediate move as Black in the example position would be 67… Bb8, allowing 68. axb8(Q)+ Kxb8, entering the “never studied” endgame excluded by Silman. Why? Because the odds are you will NOT have studied it (as a diligent student of Silman) and unless you have a considerable amount of time left on your clock, you will NOT figure it out. All I have to know is that I should try to stay in (or head back toward) a corner that is the opposite color of the White Bishop (a8 or h1), since it is impossible to force mate there. Given the game is already past the first (and possibly the second) time control, the odds of you getting it right within the 50-move rule are slim to none. Good luck!
END INTERMEZZO
PART III:
ReplyDeleteSo, what goal(s) have we arrived at? White is going to try to exchange his pawn for the Black bishop or exchange the two Bishops while protecting the pawn and retaining the Knight. That first idea seems a lot more feasible than the second one, so let’s roll with that as the idea/plan/scheme. (No variations yet!)
In the initial position, Black’s Bishop is somewhat constrained (if Black doesn’t take my suggested approach of immediately transposing into the most difficult endgame). White controls b8, b6, d8, d6, e5, f4 leaving Black only g3 and h2. Dominating the opponent’s piece(s) is always a good idea!
So how will we convince Black to take our pawn while we capture his Bishop? The only possible way is to set up the White pieces so that when White threatens to advance the pawn (protected by the Knight) that Black will have to capture the pawn, allowing a discovered check by the White Bishop with a simultaneous attack on the (assumed) unprotected Black Bishop. The resulting position will be a Bishop and Knight vs. Lone King. We continue the rest of this verbiage from the book.
But then how can he [White] win? If you try to find a win here by concrete calculations alone, you will immediately realize that this is by no means easy. Instead of wasting time on difficult computations, it is much more sensible here for White to think about the best possible arrangement for the pieces. Black’s bishop is already restricted as much as possible; the knight will need to support promotion threats, so the idea comes of sticking the king on b5 and the knight on d5. In fact, once this setup has been achieved, the win will be quite easy.
Throughout this maneuvering process, the White bishop will be responsible for hanging on to the a7-pawn, so it’s going to remain on the current diagonal. The White King and Knight must be maneuvered into position.
The authors give the following moves without comment:
67.Kxe4 Ka8 68.Kd5 Kb7 69.Bd4 Ka8 70.Ng5 Kb7 71.Ne6 Bh2 72.Kc4 Bg3 73.Kb5 Bh2 74.Bb6 Bg3 75.Nc5+ Ka8 76.Ne4 Bh2 77.Nc3 Kb7 78.Bd4 Bg3 79.Na4 Bh2 80.Bf2 Bf4?!
PART IV:
ReplyDeleteHere is the only suggested divergence from the game.
This makes it easy for White. 80...Bc7! 81. Nc3 (Diagram) and now the black bishop can twist and turn all it wants, but there is no escape, as Hecht showed in CBM:
FEN: 8/Pkb5/8/1K6/8/2N5/5B2/8 b - - 0 81
a) 81...Bh2 82. Nd5 Bd6 (82...Be5 83.Nb6 is similar) 83. Nb6 Kxa7 84. Nc8++ (double check) +-
b) 81...Be5 82. Nd5 Bh2 83. Ka5 (zugzwang) +-
c) 81...Bd8 82. Nd5 Bg5 (82...Ka8 83. Ka6) 83. Nc7 +-
81.Nb6 Kxa7 82.Nd5+ Kb7 83.Nxf4 Kc7 84.Bc5 Kb7 85.Nd5 Kb8 86.Kc6 Ka8 87.Nc7+ Kb8 (Diagram)
FEN: 1k6/2N5/2K5/2B5/8/8/8/8 w - - 0 88
Now we have achieved the standard position, in which the famous ‘W-manoeuvre’ of the white knight comes into effect (incidentally, this is also very schematic).
88.Bd4 Kc8 89.Ba7 Kd8 90.Nd5
The first step on the W; the knight will end up on g7, via e7-f5, each tinme preventing the black king from heading back to the safe a8-corner. From e7 the knight controls c8 and from g7, c8.
90...Ke8 91.Kd6 Kf7 92.Ne7 Kf6 93.Be3 Kf7 94.Bd4 Ke8 95.Ke6 Kd8 96.Bb6+ Ke8 97.Nf5 Kf8 98.Bc7 Ke8 99.Ng7+ Kf8 100.Kf6 Kg8 101.Bd6 Kh7 102.Nf5 Kg8 103.Kg6 Kh8 104.Bc5 1-0
I have not run the game score through GM Stockfish looking for improvements. Apparently the game score is sufficiently accurate to not cause any problems.
PART V
ReplyDeleteI directed GM Stockfish to analyze the entire game. There were innumerable errors during the earlier part of the game. I include the following excerpt from the annotated game to illustrate that White and Black played fairly well in the endgame. White did "blunder" on his 85th move. Hey, nobody's perfect!
63.Kg6□*
Only move that keeps the advantage
63...Ka8 64.f5 Bxf5+ 65.Kxf5 Kb7 66.Nf7 Bc7 67.Kxe4 Ka8 68.Kd5 Kb7 69.Bd4 Ka8 70.Ng5 Kb7 71.Ne6 Bh2 72.Kc4 Bg3 73.Kb5 Bh2 74.Bb6 Bg3 75.Nc5+ Ka8 76.Ne4 Bh2 77.Nc3 Kb7 78.Bd4 Bg3 79.Na4 Bh2 80.Bf2 Bf4 81.Nb6 Kxa7 82.Nd5+□*
Only move that keeps the advantage
82...Kb7 83.Nxf4□*
Only move that keeps the advantage
83...Kc7 84.Bc5 Kb7 85.Nd5??*
Blunder +5.22 → +1.96
[ Avoiding the blunder 85.Bb6* Kc8 86.Kc6 Kb8 87.Bg1 Ka8 88.Nd5 Kb8 89.Nc7 Kc8 90.Bc5 Kb8 91.Bd4 Kc8 92.Ba7 Kd8 93.Nd5 Ke8 94.Kd6 Kf8 95.Ke6 Kg7 96.Bd4+ Kf8 97.Bf6 Kg8 98.Be7 Kh8 99.Kf7 Kh7 100.Bf8 ]
85...Kb8 86.Kc6 Ka8 87.Nc7+ Kb8 88.Bd4 Kc8 89.Ba7 Kd8 90.Nd5 Ke8 91.Kd6 Kf7 92.Ne7 Kf6 93.Be3 Kf7 94.Bd4 Ke8 95.Ke6 Kd8 96.Bb6+ Ke8 97.Nf5 Kf8 98.Bc7 Ke8 99.Ng7+ Kf8 100.Kf6 Kg8 101.Bd6 Kh7 102.Nf5 Kg8 103.Kg6 Kh8 104.Bc5
[ Game could have continued 104.Bc5* Kg8 105.Nh6+ Kh8 106.Bd4# ]
1-0
Somehow, I only seldom get a notification of your comments lately. Feel free to tip me whenever you think I missed one.
ReplyDeleteI got a notification your comment 2 days later
DeleteI think it is risky to gamble on the fact that people don't know how to play B+N vs king. I certainly used to know it, albeit that that was 20 years ago. I still know two crucial manoeuvers, so I figure that I can find the mate within a reasonable time.
ReplyDeleteYou are correct: it IS risky to gamble on someone else's potential ignorance. The Black player (rated 2461) did force the White player (rated 2379) to play out the B+N vs K endgame to within three moves of checkmate. I assume (without evidence) that the clock situation and the difference in rating may have been the crucial factors in Black's decision to not resign immediately after entering that endgame. I don't ordinarily gamble on anything in chess, but in that specific situation, I would—because, IMHO, the only alternative is to resign. If there is the slightest hope to escape a loss, then play it out to the bitter end!
DeleteMy biggest accomplishment the past year was my progress in positional play. That progress was due to the fact that I discovered the principle of invasion. When you don't know what to do, invade.
ReplyDeleteIn fact, it was an endgame that lead to that insight.
An invasion is an extra option in the middlegame. The endgame has its own extra targets: the promotion square and to path to promotion. I first must get that clear in pawn endings. It is too early to worry about pieces. As long as I don't know what to move, the king or a pawn, and when I don't know whether I must move, exchange or sac a pawn, my work is cut out for me: discover the schemes that give answers on these questions.
ReplyDelete