What is there to SEE?
To understand why 3 hours per move is not so unreasonable, you must know what to expect from the study. When you know what to learn, you will become more patient. Especially when you notice that the grandmaster missed the move in the game too because he didn't see it either.
 |
| White to move |
1r2r3/p5pp/2nk4/1R1pp3/P2P4/2P3B1/3K2PP/4R3 w - - 0 1
Warmerdam, Max vs. Elgersma, Simon
After the move 27.Rxe5, black has 3 captures. It takes time to work out the logical narratives for each capture. Then it takes time to absorb the three narratives so that you SEE them and don't need the verbal description of the narrative anymore. Once you SEE them, you can't UNSEE them.
Max played 27.Rxb8. And there is a narrative why this capture is not so good. It is still winning, but there is work left to be done. And this narrative must be absorbed too. So there are 4 narratives in total that are in need to be absorbed.
The main task of studying tactics is to unlearn the habit of trial and error, and to replace that with seeing the scenario of what you have to accomplish. Moves are only relevant for realizing that scenario.
ReplyDeleteWhen that is accomplished, I plan to save the position for future use. When I'm going to try to answer the question "how did this position came about?". So time that I did put in the position is not lost.
Usually, we never get to see what precedes a tactical puzzle, and we don’t get the challenge of actually winning a “won” game. I think that isolation can be detrimental to improving overall playing skill.
DeleteI've been experimenting with lichess.org. I play a session of Puzzle Storm. Afterwards, I go back over the "interesting" puzzles (especially those where the players are rated over 2000), resetting to the start of the game (“Toggle local evaluation” = ON, using StockFish 17 · 79MBNNUE) and playing over the moves until I find the blunder preceding the start of the puzzle; usually, this is when the evaluation function drastically shifts from one player to the other). I try to figure out WHY that blunder occurred; usually, it’s a simple oversight. I reset to the start of the puzzle and then switch on "Play with the computer" bringing up the analysis board. I then play through different variations against Stockfish until I am satisfied that I have seen everything of importance. I then reset and play the recommended main variation up to the end of the puzzle (if it is NOT checkmate), and then play as if it’s a regular game against Stockfish all the way to checkmate.
WHY? In order to improve overall skill and technique, not just to pick up isolated fragments of tactical patterns.
I plan to do more or less the same, but not at the same time. I have now a fixed problem set with tactical problems, and have absorbed most of them. I consider the set rich enough to use it many times for different purposes. It contains many endgames that I want to play out at some time. And I will study the build up of the combination, as said. I don't think that the isolation is too much of a problem, because it helps me to focus.
DeleteThe funny thing is that I often see the same problems somewhere on facebook or youtube and say, "hey, that was an Anastasia's mate from the game between Euwe and Vidmar!".
A cogent example of the vicissitudes of playing chess [from lichess.org]:
DeleteFEN: 8/8/1p6/1Bp5/P1P1k3/2K1PR2/bP3R2/3r4 b - - 2 49
White’s previous move was 49 R5f3. Notice the “blunder” indication: the complete reversal of the GM Stockfish evaluation.
What pattern did White NOT “SEE”? WHY did he fail to see it: was it lack of time on the clock, lack of knowledge of that particular pattern, an insufficient trigger (context) for the appropriate pattern, or focusing on the wrong thing (expecting his material superiority to “win” the game for him)?
Black, on the other hand, did “SEE” the appropriate pattern on his turn to move.
“There's none so blind as those who will not [or cannot] SEE.”
Puzzle #ftpXj
Rating: 1335
Played 9,787 times
From game 3+0 • Blitz
NM username20 (2371)
Duraco7 (2402)
1 d4 Nf6 2 Nf3 b6 3 Bf4 Bb7 4 e3 d6 5 Bd3 Nbd7 6 h3 e6 7 Nbd2 Be7 8 c3 h6 9 Qc2 g5 10 Bh2 g4 11 hxg4 Nxg4 12 Bg3 Bf6 13 Bh4 Qe7 14 Bxf6 Ndxf6 15 Bb5+ c6 16 Be2 O-O-O 17 Rf1 e5 18 dxe5 dxe5 19 Qf5+ Kb8 20 O-O-O c5 21 Nc4 e4 22 Nfe5 Nxe5 23 Nxe5 Rde8 24 Qxf6 Qxe5 25 Rd8+ Bc8 26 Qxe5+ Rxe5 27 Rxh8 Rg5 28 Rxh6 Rxg2 29 Rh8 Kc7 30 Re8 Be6 31 Bb5 a6 32 Bxa6 Kc6 33 a4 Rg5 34 Bb5+ Kd6 35 Rb8 Kc7 36 Ra8 Kd6 37 f4 exf3 38 Rxf3 Bb3 39 Rd8+ Ke7 40 Rd2 Rg1+ 41 Rf1 Rg3 42 Re2 f5 43 Kd2 Ke6 44 c4 Ke5 45 Rf4 Rg1 46 Ref2 Rd1+ 47 Kc3 Ba2 [+13.1] 48 Rxf5+ [#23] Ke4 [#3] 49 R5f3 [#-3] Rd3+ [#-2] 50 Kc2 [#-2] Bb3+ [#-1] 51 Kb1 [#-1] Rd1#
Yes, that is the right formulation. You are ready with studying a position when you SEE the scenarios.
ReplyDeleteThere is a lot of "FUN" going on along the h2-b8 diagonal, with "a cherry on top" pin on e5 (B.A.D. [3:3]). The complications come from having to "SEE" (or figure out) the exchange sequence(s) required to gain the most advantage.
ReplyDeleteIn an ideal world, you see a position for the first time, and then a few possible scenarios pop up. Then you look if you you can make a scenario work. Only once you have a scenario that might work, you start to calculate, by checking all relevant lines of attack and their interactions.
ReplyDeleteIf no scenario pops up, you conclude that there is no tactic. Then you start looking for a positional scenario.
I'm curious about how thoroughly we should examine ALL possibilities during training.
ReplyDeleteIn the given position, GM Stockfish reviewed the top 10 moves (at my request) and gave GM Warmerdam’s move Rxb8 the third highest score of +2.81 (1.5 hours of analysis time). As you noted, the highest rated move is Rxe5 [+7.36], followed by c4 +3.9]. The tenth highest move is rated +1.87, which is still winning for a GM.
I understand the desire to exhaustively analyze all possibilities DURING TRAINING in order to glean as much as possible during the training time. BUT, GM Warmerdam was PLAYING, not training.
Once I have found ONE way that absolutely works, regardless of how long the process takes, I’m content to see it through to the end without agonizing over “Woulda, Coulda, Shoulda” variations. Am I being too pragmatic for my own good?
Is it worth trying to SEE the point of ALL possible variations during training? Is there a rapidly decreasing return on time invested trying to analyze ALL possible variations in every training position?
GMKSGM Stockfish (GigaMegaKiloSuperGrandMaster) ^^
DeleteFrom this specific position I'm only bothered by the three responses to Rxe5. Those three scenarios I want to be able to see.
ReplyDeleteThe move Rxb8 I might have a look at somewhere in the far future out of curiousness, to understand what happened there. I don't consider it to be essential.
The time consuming part is twofold:
- it takes a lot of time to investigate and formulate the logical narratives
- I'm not easy satisfied with the result of the absorption. I know when I have absorbed the scenarios, and when I have not, I'm going to repeat it, no matter what.
3 hours per position at average is how it goes in practice. Since the amount of scenarios is finite and for life, I don't care how long it takes.
Given my tactical ability, I'm on a par with players that have a rating 250 points higher then me. So I stopped extending the problem set. I only am going to repeat the current set until I'm satisfied.
Today I started with a new set with problems, but now it are positional problems. Since in the positional area, I'm not on a par.
I used to struggle with the value administration of a combination. How much value did I gain in a certain line. When absorption is well done, you SEE the gain of value without calculating it.
ReplyDeleteTo get an impression.