Backwards thinking redux
We talked a lot about backwards thinking versus trial and error in the past. Now I learn to use the logical narrative, it becomes more and more a replacement of trial and error. Which feels really good!
Black to move |
5k2/5pp1/7p/3pn3/1r5P/4K1P1/1PR2P2/1B6 b - - 0 42
Pancevski, F. vs. Van Kooten, L., Hoogeveen 2013
The logic narrative revealed a new idea. The pieces are too fast to chase. But you might be able to chase the king into a fork with the rook.
Here fore you need to know two patterns.
Here you see where you need to chase the white king to in order to get a knightfork
And it helps to see that there is a mate in the position too.
Black to move |
This gives a clear idea what you need to learn from this position.
The logic: chase the king into a knightfork because it is slow
Pattern 1: the position of the king where it can be forked
Pattern 2: the mate
This prunes the tree of analysis drastically. From the patterns you can reconstruct backwards how to chase the king.
UPDATE
Here you see a position where the same logic applies. Chase the slow moving king into a discovered attack or a skewer. Again you need to see both patterns. Of course you already have absorbed those patterns in the past. But you need to connect it to the logical narrative.
That way, when you encounter such problem OTB, you are using system 2 to think logically about the position and system 1 kicks in when it recognizes the patterns. System 2 guides the attention and system 1 looks over the shoulder. No eagle required. It is not about absorbing the patterns but about connecting the patterns to the logic.
White to move |
8/5B2/2R5/1pP1p1k1/1P2P3/r5pK/3b4/8 w - - 0 51
Carlsen, M. vs. Ding L., Stavanger Armageddon Blitz 2019
Only AFTER you have found the solution OTB, you start to CALCULATE every possible answer to see whether there is a refutation.
UPDATE of the post in blue.
ReplyDeletePART I:
ReplyDeleteIn both positions, the “trick” is to eliminate as many alternatives as possible by using forcing moves, reducing the possible replies to a minimum on each turn.
Position 1:
FEN: 5k2/5pp1/7p/3pn3/1r5P/4K1P1/1PR2P2/1B6 b - - 0 42
Coach Yogi Berra (a great New York Yankees baseball player and coach) once stated, "When you come to a fork in the road, take it." Perhaps he was playing a game of chess at the time, and actually said, “When you SEE a fork on the board, take it.” Whatever.
When you are trying to drive the enemy king into the box, reducing the available moves by force is always a good indication that you ARE taking the correct fork in the road. After all: these moves are imaginary; “No harm, no foul.” [Or maybe that’s “No farm, no fowl.”]
In position 1, 1… Re4+ forces 2. Kd2 as the only legal move.
Having gotten this far, it’s fairly easy to find another forcing move: 2… Nf3+.
However, this time, White has 3 possible moves: 3. Kd1, 3. Kd3 and 3. Kc3 (on the way to b3). BTW, the diagram with three White kings on the board is brilliant! This is NOT “illegal” to do this: after all, it is an imaginary position. This is another “trick” that is quite handy when visualizing because System 1 gets to SEE (potentially recognize) three potential patterns at once. The first pattern is the potential Rook+Knight checkmate (3. Kd1 Re1#). HURRAH! We are 1/3 of the way toward reaching a “final solution!”
White has only two remaining choices. Since we have come to a “fork in the road,” take it. (This is what GM Tisdall suggested in variation processing when faced with alternatives: pick the simplest one [based on your intuition] and follow that fork, knowing that sometimes, you will pick the “wrong” [not simplest] alternative; stuff happens.) What if White plays 3. Kd3? System 1 should be triggered by the fork pattern with 3… Ne1+, forking the White King and Rook. Since losing the Exchange is NOT in White’s best interest, we can assume that we have completed another 1/3 of the “final solution!”
Having eliminated 2 of the 3 possible replies, again White is “forced” into choosing the least harmful alternative 3. c3. Here’s where a “stepping stone” position could be useful. Visualize the BRe4, BNf3, WKc3 with everything else unchanged from the original position.
We already know that White does NOT want to play his King to d3, allowing the fork. So, look for another forcing move. 3… Rc4+ gives White one of two choices, of which one has already been eliminated (4. Kd3 Ne1+ forking the King and Rook). The only alternative is 3. Kb3, but this allows a different knight fork of the King and Rook: 3… Nd4+, picking up the WRc2 (which is attacked twice and, after the White King moves away from b3, captured). Since there are no more alternatives AND all previously analysis shows Black checkmating or gaining material in all variations, we can confidently play 1… Re4+ and wait to see which “final solution” White will choose (presuming he does not resign immediately).
Correction:
DeleteHaving eliminated 2 of the 3 possible replies, again White is “forced” into choosing the least harmful alternative 3. c3. Here’s where a “stepping stone” position could be useful. Visualize the BRe4, BNf3, WKc3 with everything else unchanged from the original position.
should read:
Having eliminated 2 of the 3 possible replies, again White is “forced” into choosing the least harmful alternative 3. Kc3. Here’s where a “stepping stone” position could be useful. Visualize the BRe4, BNf3, WKc3 with everything else unchanged from the original position.
PART II:
ReplyDeletePosition 2:
FEN: 8/5B2/2R5/1pP1p1k1/1P2P3/r5pK/3b4/8 w - - 0 51
Again, forcing moves help to restrict the number of alternatives that must be considered. It also helps—a lot!—that the Black King has very restricted alternatives to evade attack.
1. Rg6+ gets the Black King moving.
There are only two alternatives: 1… Kh5 or 1… Kf4. Take the “simplest” fork in the road: 1… Kh5, setting up a discovered check. Already System 1 should be screaming about capturing with 2. Rxg3+ (discovering check), followed by 3. Rxa3.
1… Kf4 requires an additional move to set up the skewer (also incidentally protecting WPd4): 2. Rg4+ forcing the Black King to the 3rd rank, to be followed by the skewer 3. Rxg3+ and then 4. Rxa3. When does System 1 recognize the potential skewer? Perhaps only after realizing that the Black King MUST move to the 3rd rank in response to 2. Rg4+.
As you noted, the same “logic” applies (the logical scenarios are very similar). That’s certainly enlightening, because the pieces-on-squares are completely different, yet there are just two distinct (and different) tactical patterns that System 1 must recognize at the appropriate point in the logical process in both positions. It sure does help improve our “vision” to eliminate trial and error!
I advocate to not accept calculation as the standard. The patterns of the motifs are already absorbed. What is not absorbed, is to see these patterns immediately in this position. That is why I plea to fiddle around with this position for some time, in order to learn more about the most principled variation.
ReplyDeleteHence I urge you to limit the pace to one or two problems per day.
Limit the use of calculation to refutation only. The most principled variation (the solution) must emerge from the interaction between the patterns (system 1) and logic reasoning (system 2) solely. Replace the eagle by logic reasoning.
Hence the "backwards thinking" in the title of this post.
ReplyDelete"goal" should read "positional goal"
ReplyDeleteBack in early March, 2017, Aox commented: "There is no real difference between positional and tactical thinking. There are weaknesses and you need to find methods to make use of them.
ReplyDeleteA slight expansion of Aox’s aphorism might be helpful:
There are weaknesses (and the potential to create weaknesses) and the goal is to find methods to create and take advantage of all possibilities, no matter how those possibilities arise (whether accidentally or via logical deliberation and intention).
With regard to forward thinking or backward thinking, I think you have resolved the conundrum between tactical (forward) thinking and positional (backward) thinking.
It has always been a dichotomy that combination-oriented players tend to “think forward” (usually based on dynamic factors such as the initiative, active piece play, etc.) to discover a favorable endpoint, whereas positional-oriented players tend to “SEE” a desirable goal (usually based on static factors, primarily pawn structure) and then “think backward” to the current position in order to determine a viable path to that end goal. Mister Lasker opined that the master player synthesizes these two contrasting approaches, but he provided only indirect “clues” as to how that is to be accomplished.
Your recent approach resolves the conundrum of which direction to think by synthesizing the two directions into one masterly process.
I'm not talking about positional play yet. I'm talking about the cooperation of system 1 and system 2. System 1 projects the future goal. From there you go backwards to the current position. Then system 2 starts thinking. System 2 thinks forward. It can't do otherwise.
DeleteThis is what I advocate: absorb the solution to a degree that you SEE the future position before your minds eye. The fork, the mate, the skewer. Educate the eagle. Then go to the position and work out the narrative that leads to solution. Work out the chase.
In positional play, it works equally so. You picture which LoA's you want to dominate (the end). Then you go to the beginning and use system 2 to get there.
In an ideal world, both the picture of the end AND the logical elements how to get there are absorbed in the study room. That should be doable, because both are finite.
What is not finite, is the amount of ways to COMBINE the elements.
Compare it with language. The amount of words is finite. Vast, but finite. But the amount of sentences you can make is infinite.
System 2 works better if it has a clear goal. Otherwise the result is accidental, just like ordinary trial and error.
DeleteYou have to be aware of that during study. You study better when you know what the aim is. Otherwise you think that you are ready with a problem while there is still something you should learn from it.