Do mini skills add up to the Trick?
What have I found so far? The Trick is the decisive element from novice to super grandmaster. It is the decisive element between two players of equal level. Players with equal rating differ in what they have automated (absorbed by System 1). What happens in the game determines which skill is pivotal. That is a matter of chance.
The human mind is not suitable to play chess. Because the mind is easily overwhelmed. There are too many parameters. So we need a trick to prevent the mind from cognitive overload.
Luckily, evolution has provided us with a mechanism to prevent cognitive overload. We can simplify reality before it happens. By logic and conceptualisation we can simplify the world so it becomes more manageable. We even have an inner LLM which can predict the future based on what has happened before. Of course such simplification results in mistakes caused by bias every now and then.
When you hear a rustle in the bushes you better be biased that it is a lion and run, rather than to save energy assuming it's nothing.
Hence oversimplification, bias and hallucinations (LLM!) are part of the game in order to prevent cognitive overload.
The atomic cognitive element is the logical question/answer process IF I do this THEN he does that. Typically 2 ply deep. Since the question has a 1:n relation to the answers, there can be many answers that need an immediate answer (ELSE IF, ELSE_IF, Elif).
The logical component is simplified to Yes or No. IF I do this THEN he does that, is that good Yes OR No?
These questions need an immediate answer by System 1. If you need System 2 to concoct an answer, you are toast. The function of System 2 is to judge whether whatever System 1 comes up with is correct.
This imposes strict requirements on what System 1 must store during study time. This is about the logical consequences of a move. IF I move this piece away THEN it no longer protects that invasion square.
We already observed that a good functioning System 1 leads to improved visualization without special training. We observed that it increases the speed too without separate training. The reason for this is the storage of the logical consequences of a move by System 1. That's a micro skill.
By studying an (almost) random set of puzzles, we can unearth all the necessary micro skills. The question that now arises is: "Do al micro skills add up to the Trick?"
Or is there some mysterious skill that only grandmasters know about?
Since I'm no grandmaster, I cannot answer that question. Yet. But I'm going to find out.
While working on this three-part comment, I did not notice you had already gotten ahead of me (again). So I'll post the comment, and hopefully it will illustrate your points in this post.
ReplyDeletePART I:
Here are examples of how to go about ingraining a particular Trick into LTM. Find a pattern (any pattern) that repeats in different puzzles. Note the essential aspects (the “essence”) of the pattern and various alternatives branches of the logical sequence. When you have finished gathering sufficient puzzles to cover all variants, note the branches in the solution and commit to memory the triggering pattern and the logical sequence (including the branches). Lather, rinse, repeat with other Tricks. Over time, you will absorb more and more Tricks, covering most positions you will encounter.
NOTE: The specific examples are unimportant. Do NOT memorize the “pieces-on-squares!” Instead, focus attention on the Trick pattern as an abstract MECHANISM (the interplay of attack/defense) in the logical sequence ending in advantage.
The process of recognizing Trick patterns can be exemplified in the following two (composed) positions.
Position 1:
FEN: r2r2k1/1p3ppp/p7/8/3q4/P7/1P1Q1PPP/R3R1K1 w - - 0 1
Position 2:
FEN: 2Rr2k1/1p1Prppp/p7/8/3R4/P7/1P3PPP/6K1 w - - 0 1
PART II:
ReplyDeleteHere are some actual game positions with these patterns.
Position 3:
FEN: 3r2k1/pp3p1p/3q1Qp1/8/8/1P5P/6P1/4R2K w - - 1 37
Black’s last move was 36.… Q(f8)d6. Black WON in 1 more move! What was White thinking??? Perhaps: “Oh goody, I can add another piece to the attack on the B.A.D. f7-square!” Whatever he thought, he did NOT consider this Trick. What were ALL the things that Black’s last move did? In this type position, Black has no escape hatch for getting out of check on the back rank, so he is forced to capture on e8, removing the defender of BQd6.
[Event "rated rapid game"]
[Site "https://lichess.org/AaIU5hFe"]
[Date "2024.10.14"]
[Round "-"]
[White "gitoburung_jrc"]
[Black "Ajdar2-31-4"]
[Result "0-1"]
[GameId "AaIU5hFe"]
[UTCDate "2024.10.14"]
[UTCTime "06:38:02"]
[WhiteElo "2079"]
[BlackElo "2062"]
[WhiteRatingDiff "-6"]
[BlackRatingDiff "+6"]
[Variant "Standard"]
[TimeControl "600+0"]
[ECO "C02"]
[Opening "French Defense: Advance Variation, Paulsen Attack"]
[Termination "Normal"]
1. e4 e6 2. d4 d5 3. e5 c5 4. c3 Nc6 5. Nf3 Nge7 6. Bd3 cxd4 7. Nxd4 Nxe5 8. Bc2 N7c6 9. O-O Be7 10. Be3 O-O 11. Nd2 Nxd4 12. Bxd4 Nc6 13. Be3 Bg5 14. Nf3 Bxe3 15. fxe3 Qb6 16. Qc1 e5 17. b4 Bg4 18. Qd2 Bxf3 19. Rxf3 d4 20. exd4 exd4 21. Kh1 dxc3 22. Rxc3 Qxb4 23. Qd3 g6 24. Rb1 Qe7 25. Bb3 Rfd8 26. Qf1 Nd4 27. Re1 Qf8 28. Bd1 Nf5 29. Bb3 Nd4 30. Bd1 Nb5 31. Rf3 Nd4 32. Rf6 Rd6 33. h3 Rad8 34. Bb3 Nxb3 35. axb3 Rxf6 36. Qxf6 Qd6 37. Re7 Qxf6 0-1
Position 4:
FEN: r5k1/3b2p1/3p3p/2pPpp2/2P5/q7/3N1PPP/QR4K1 w - - 0 33
Black’s last move was 32.… Qxa3. I presume that Black thought it was “safe” to capture since the Black king is not confined to the back rank in case of a check by WRb1. An alternative line is that (after 33. Rb8+), Black simply moves his king 33.… Kf7 (or Kh7) and escapes any adverse consequences. Unfortunately, White has 34. Rxa8, capturing a rook for free. (This brings into play the “Reloader” theme/device in either direction: Black captures 34.… Qxa8 35. Qxa8 or 34.…. Qxa1+ 35. Rxa1.) This is often the primary alternative if there is no forced mate threat on the back rank. It is important to “SEE” this alternative as a integral part of the overall logical sequence of this Trick.
[Event "rated rapid game"]
[Site "https://lichess.org/c8hJMb2k"]
[Date "2022.10.03"]
[Round "-"]
[White "unbelievably"]
[Black "minibadger"]
[Result "1-0"]
[GameId "c8hJMb2k"]
[UTCDate "2022.10.03"]
[UTCTime "18:19:59"]
[WhiteElo "1902"]
[BlackElo "1906"]
[WhiteRatingDiff "+26"]
[BlackRatingDiff "-6"]
[Variant "Standard"]
[TimeControl "600+0"]
[ECO "B80"]
[Opening "Sicilian Defense: Scheveningen Variation"]
[Termination "Normal"]
1. e4 c5 2. Nf3 d6 3. Nc3 Nf6 4. d4 cxd4 5. Nxd4 e6 6. Bd3 Be7 7. O-O Nc6 8. Bb5 Bd7 9. Nf3 a6 10. Bxc6 bxc6 11. Bf4 e5 12. Be3 c5 13. Nd5 Nxd5 14. exd5 Qb6 15. b3 Rb8 16. c4 a5 17. Rb1 Bf5 18. Rb2 Bf6 19. Bg5 Bxg5 20. Nxg5 a4 21. Re1 axb3 22. Rxb3 Qc7 23. Rxb8+ Qxb8 24. Qa4+ Bd7 25. Qc2 h6 26. Rb1 Qa7 27. Ne4 Qc7 28. Qb3 O-O 29. Qc3 f5 30. Nd2 Ra8 31. a3 Qa7 32. Qa1 Qxa3 33. Rb8+ Kh7 34. Rxa8 Qxa1+ 35. Rxa1 1-0
PART III:
ReplyDeletePosition 5:
FEN: rr4k1/4bppp/pq2p3/1p1p4/1P1Q4/8/4NPPP/R1R3K1 w - - 0 27
Black’s last move was 26.… Qxb6. Looking at the back rank, there is no possibility of checkmate: BBe7 can retreat to f8, and both Black rooks defend it and the back rank. This illustrates another variation of the same logical sequence. After 27. Rc8+ (the all-important check as the highest level of forcing move), Black can simply play 27.… Bf8, and it appears that the BQb6 continues to be protected. IT IS – but the BRb8 now has TWO defensive functions: it must protect the BQb6 AND also the BRa1. After 28. Qxb6 Rxb6 29. Rxa1, White is up a rook for three pawns in material. The pawns will start dropping next, starting with BPa6. And now for the rest of the story: BLACK WON!! White apparently did NOT have this Trick up his sleeve!
[Event "rated rapid game"]
[Site "https://lichess.org/wwtYbYmk"]
[Date "2021.08.26"]
[Round "-"]
[White "Alejandro_Yero"]
[Black "rubberorangefox"]
[Result "0-1"]
[GameId "wwtYbYmk"]
[UTCDate "2021.08.26"]
[UTCTime "16:03:24"]
[WhiteElo "1854"]
[BlackElo "1767"]
[WhiteRatingDiff "-8"]
[BlackRatingDiff "+8"]
[Variant "Standard"]
[TimeControl "600+0"]
[ECO "C02"]
[Opening "French Defense: Advance Variation, Milner-Barry Gambit, Main Line"]
[Termination "Normal"]
1. e4 e6 2. d4 d5 3. e5 c5 4. c3 Nc6 5. Nf3 Qb6 6. Bd3 Bd7 7. O-O cxd4 8. cxd4 Nxd4 9. Nxd4 Qxd4 10. Qb3 Qxe5 11. Bb5 Qc7 12. Bxd7+ Qxd7 13. Be3 Nf6 14. Bd4 Be7 15. Nc3 O-O 16. Qd1 b6 17. a4 Rfc8 18. Nb5 a6 19. Nc3 b5 20. a5 Ne8 21. Bb6 Nd6 22. b4 Nc4 23. Qd4 Nxb6 24. axb6 Qc6 25. Ne2 Rcb8 26. Rfc1 Qxb6 27. Qb2 Bf6 28. Qa2 Bxa1 29. Qxa1 Rc8 30. Rb1 h6 31. Ng3 g6 32. h4 Rc2 33. Rb2 Qxf2+ 34. Kh2 Qxg2# 0-1
Position 6:
FEN: 3r3k/2qP2pp/4Q1p1/1pp5/3p4/pP1P3P/P1P3B1/6K1 w - - 1 31
Black’s last move was 30.… Q(f4)c7.
[Event "rated blitz game"]
[Site "https://lichess.org/KWu9kr6B"]
[Date "2024.10.13"]
[Round "-"]
[White "Arcybishop"]
[Black "yeho1111"]
[Result "1-0"]
[GameId "KWu9kr6B"]
[UTCDate "2024.10.13"]
[UTCTime "11:38:22"]
[WhiteElo "1921"]
[BlackElo "1955"]
[WhiteRatingDiff "+6"]
[BlackRatingDiff "-7"]
[Variant "Standard"]
[TimeControl "180+0"]
[ECO "A00"]
[Opening "Mieses Opening"]
[Termination "Normal"]
1. d3 d5 2. Nc3 c6 3. g3 e5 4. Bg2 Nf6 5. f3 Bd6 6. e4 d4 7. Nce2 Qc7 8. h3 c5 9. f4 Nc6 10. b3 a5 11. Nf3 a4 12. Bd2 a3 13. O-O b5 14. Nh4 Be6 15. fxe5 Nxe5 16. Nf4 O-O 17. Nd5 Bxd5 18. exd5 Ng6 19. Nxg6 fxg6 20. Bf4 Nh5 21. Bxd6 Qxd6 22. Qg4 Qxg3 23. Rxf8+ Rxf8 24. Qe6+ Kh8 25. Rf1 Nf4 26. Rxf4 Qxf4 27. d6 Qc1+ 28. Kh2 Qf4+ 29. Kg1 Rd8 30. d7 Qc7 31. Qe8+ Rxe8 32. dxe8=Q# 1-0
More examples (“Drill, baby, drill!” but not in an oily sense):
ReplyDelete1. FEN: 5rk1/1p1Qbp2/p1q1p2p/4P2N/P3r3/2P3P1/1P4RP/3R3K b - - 13 32
Game Link: https://lichess.org/7i2BsdeL/black#63
Black saw the Trick and won.
2. FEN: 2r3k1/1p4pp/p1q1pr2/8/3PQ3/P1P5/1P4PP/1R2R2K b - - 0 23
Game Link: https://lichess.org/6uiSpGmN/black#45
Black missed the Trick, and (eventually) lost the game.
3. FEN: 2r2k2/5ppp/3q4/8/P3Q3/7P/2p2PP1/2R4K b - - 8 32
Game Link: https://lichess.org/rE9yrQc0/black#63
Black missed the Trick, but still (eventually) won the game.
4. FEN: 3r2k1/p4pp1/6n1/6q1/8/2P1R3/PP1Q1PP1/6K1 w - - 2 35
Game Link: https://lichess.org/9Sqi6iGR/white#68
Black responded unexpectedly, and White did not take the best advantage of it, but did win eventually.
5. FEN: 4r2k/3q2pp/2p5/pp6/3Q4/PP6/1P4PP/3R2K1 b - - 0 32
Game Link: https://lichess.org/cuwUS2QV/black#63
White missed the Trick and lost immediately.
I just had a minor epiphany. (I have to take them whenever and however I can get them.)
ReplyDeleteWhile working my way through several sessions of lichess’s Puzzle Storm (looking for the examples I just gave), I realized the ultimate abstraction level for all kinds of weaknesses: it is the idea of constraint. My epiphany was triggered when I started thinking about the constrained defender, which is applicable to a lot more than just the examples I gave above. Synonyms for constraint could be restriction, limitation, restraint, inhibition, check (especially in the formal sense of chess) and so forth. Note that all these terms have been used to describe different (more concrete) aspects of chess!
I think (in a very abstract sense) that the Trick revolves around the idea of constraints as exploitable weaknesses. Constraints can take many forms.
In the formal rules of chess, check is the most forcing move; it constrains the player “in check” to do anything possible to remove the effect of that check. Failure to do that is the end of the game—checkmate. This is why chess engines will blithely toss any amount of “wood” on the bonfire to escape mate.
A physical barrier (for example, the pawns in front of a castled king) is both protective (a shield) and a weakness (the king’s mobility is impaired, which is described as a “back rank weakness”). We are given “advice” (in the form of generalities posing as “rules”) to “castle early” (in the opening) and also to “make luft” (in the endgame). The problem is that the important issue of deciding when to do either of those things has no such simplistic “rule.”
Lack of space (for example, a piece that has limited room to move for any number of reasons) is a constraint on the operations that can be conducted (whether offensive or defensive).
Lack of time (in the sense of tempi as well as on the time clock) is a constraint on the operations that can be conducted (whether offensive or defensive).
Lack of material (in the sense of having fewer pieces than the opponent) is a constraint on the operations that can be conducted (whether offensive or defensive).
Functions constrain the responses available (an absolute pin restrains the pinned piece to movement along the line of the attacker; for instance, a bishop pinned to a king by a queen along a diagonal can still move to capture the queen). A defender must have sufficient compensation to abandon the piece it is defending.
Take a look at any puzzle, and (if you squint your eyes) you may “SEE” some (or all) of the constraints in that concrete position. If you do, you will then KNOW the direction that play should go. Your choice of “candidate moves) will be constrained to those that address the constraints of the position.
Every weakness has a characteristic constraint (or multiple constraints) and the means to take advantage of that constraint are themselves restricted to (at best) a few possible moves/variations in a logical sequence. In short, the precursor to the Trick will ALWAYS be an awareness of concrete constraints.
I just started looking at puzzles this way, and was pleasantly surprised to find that motifs, themes/devices and so forth seem to magically “pop into existence in my awareness” even in positions with a set of constraints and a solution that I had never experienced before. Just trying to become aware of and focus on “constraints” inevitably leads toward the solution.
I think the generalization of abstract concepts (such as the concept of constraints) work to “fill in the gaps” in our set of mini skills involved in the Trick. I am quite sure that is what occurs for kids and grandmasters.
I’ve posted this notion before:
“Perceive that which cannot be seen with the eye.”
― Miyamoto Musashi, A Book of Five Rings: The Classic Guide to Strategy
"To know ten thousand things, know one well."
― Miyamoto Musashi, A Book of Five Rings: The Classic Guide to Strategy
It can't be otherwise. It's the only way a 12 year old can beat Magnus Carlsen. Now find an effective way to absorb it.
DeleteYou advised:
DeleteNow find an effective way to absorb it.
I hope you did not intend that to be merely rhetorical or sarcastic advice.
Absorption is a byproduct of repeatedly directing focused attention to salient aspects. System 1 automates the learning process and will store whatever System 2 considers important enough to focus on.
I think our training is fragmentary, composed of many isolated “silos” of knowledge and experience. We train on back rank mate, with emphasis on looking at the “fact” that the back rank is important. It is not; the weakness” of the back rank AND the availability of pieces to take advantage of the weakness ARE important.
We train on the precursor to the Greek Gift sacrifice (Bxh7+) without "connecting the dots" to the limited immobility of the Black king behind the pawn shield (that in itself is a weakness) and no protective knight on f6 (another weakness).
We train on forks, based on the mere physical representation of “piece A attacks two or more other pieces,” with no attempt to “SEE” the constraints that make that fork effective in addition to bringing it into existence.
We “know” that disparate “chunks” are part of the Trick, but we don’t focus attention on the constraints of those chunks as constraints. In short, we isolate small parts of our experience (usually) without trying to “SEE” the connectivity between all the myriad aspects of chess. Our training assumption is that we must analyze and absorb tiny little “chunks” (mini skills; micro skills) without attempting to connect them together into larger and larger constructs until we have a coherent “picture” of an entire position—regardless of the number of individual chunks contained in each specific position.
You stated the basic premise:
The trick is to SEE what a move does. You can find out what a move does by fiddling around with standard scenarios.
What a move “does” is constrained by what the moving piece is doing prior to and after moving—in short, its functions. We have to train our System 1 to gather that information in the training room. A first step is to be aware of constraints, somewhat similar to starting to analyze a position by first identifying “WHO IS TO MOVE?” Repeat the process on different positions to make sure that a System 1 automatic response eventually occurs; if System 2 “thinking” (using a step-by-step process) is required, then System 1 has not been trained sufficiently.
As you have noted, speed of recognition, visualization and calculation are all byproducts of the Trick, not elements to be trained separately and independently.
IMHO, if you have studied a specific position and focused attention on the constraints and consequent possibilities resulting from those constraints, then there is little to be gained by repeated “solving” of that exact same position. Gaining skill does require repetition, but it is NOT oriented toward “regurgitating on command” what has been seen repeatedly. Instead, find another position with similar (surface level) constraints and work through it, noting similarities and differences to previously encountered positions—what is actually there and what is missing. If you look for common patterns of constraints and corresponding moves in your working puzzle set by “fiddling around,” then the moves in the logical scenarios will begin to emerge on their own. System 1 can always be counted on to give an answer. It is System 2’s responsibility to vett that answer.
Chess is a game with perfect information. This means that you can only win by exploiting the weaknesses of the human mind.
ReplyDeleteThe chess moves are easy to see. But the restraints are not. There are basically 3 types of restraint: space (trap, mate), time (double attack) and function.
We are usually already pretty savvy in space and time restraints. But the restraints due to function are usually harder to see.
At the end of the day you will need an extra tempo in order to exploit the restraints.