War on lists
This position was first published on April 21, 2017
2r3k1/5ppp/p7/1prRq3/4n3/P1N1P2P/1P3PP1/2RQ2K1 b - - 0 1
[solution]
I have declared the war on lists. List of targets, list of points of pressure, list of lines of attack, list of attackers, list of candidate moves, list of defenders, that kind of stuff.
System II is very fond of making lists. But once they are made, it lacks the skills and the Short Term Memory slots to use them. Putting the mind on halt.
I start at the end of the line with the king:
Is it a target?
Yes
Is it a target because it is immobile?
Yes
Is it part of a potential duplo attack
Yes
Which is the other target of the duplo attack?
Queen
What is the point of pressure?
c1
Is c1 B.A.D. (Barely Adequate Defended)?
Yes
Is c1 part of a duplo attack?
Yes
What is the other target?
Knight
Is the knight immobile?
Yes
Why?
It is defending the other rook too
So it is overloaded?
Yes
Is the other rook B.A.D. (Barely Adequate Defended) too?
Yes
The first stage of the process is starting with the target at the end of the line, and then working back to the beginning of the line, to the attacker. You see that there is a very limited set of questions needed.
 |
| Diagram 1 black to move |
[solution]
I have declared the war on lists. List of targets, list of points of pressure, list of lines of attack, list of attackers, list of candidate moves, list of defenders, that kind of stuff.
System II is very fond of making lists. But once they are made, it lacks the skills and the Short Term Memory slots to use them. Putting the mind on halt.
I start at the end of the line with the king:
Is it a target?
Yes
Is it a target because it is immobile?
Yes
Is it part of a potential duplo attack
Yes
Which is the other target of the duplo attack?
Queen
What is the point of pressure?
c1
Is c1 B.A.D. (Barely Adequate Defended)?
Yes
Is c1 part of a duplo attack?
Yes
What is the other target?
Knight
Is the knight immobile?
Yes
Why?
It is defending the other rook too
So it is overloaded?
Yes
Is the other rook B.A.D. (Barely Adequate Defended) too?
Yes
The first stage of the process is starting with the target at the end of the line, and then working back to the beginning of the line, to the attacker. You see that there is a very limited set of questions needed.
- Is it a target?
- Is it immobile?
- Is it part of a potential duplo attack?
- What is the point of pressure?
- Is it B.A.D. (Barely Adequate Defended)?
- Is it overloaded?
I solved it in about five minutes. This is one of those problems that are few in combinational books, but more common in tactics exercises, where the candidate moves are pretty much known up-front.
ReplyDeleteAt first, I looked at taking the knight, but then noticed that the real problem isn't White's back-rank, but Black's. Then I almost took on d5 with the rook, but decided to blunder-check it all the way through first. Once again, Black's back-rank is at fault. So, my third "try" was with queen takes rook, which I realizes works because it leaves the rook-battery intact, and there are no back-rank issues for Black.
If I had spent long enough, I might have remembered that QxR was the solution, but I've noticed lately that memory is more hurtful than helpful because I will remember the problem, but then I'll remember the solution to a different problem!
1. Material : white is a pawn up
ReplyDeleteconclusion : its either about checkmate or winning a piece
2.Tactical weaknesses
a. Kingsafety : king is formaly tactical weak (h2,f2) but no real danger in sight
temporary conclusion its not (primary) about checkmate
b. virtually every pieces is under attack, especially blacks queen
looking closer ( making some low deepth calculations )
c ) recognising c1 is weak ( Rc5xRc1 pin , Rd5-d8 check ( mate ) ) both backranks are weak
temporary conclusion any white move has to prevent Rd5-d8 with checkmate or winning the queen with tempo
d ) Suspecting it starts with the queensac QxR ( partly because i was looking at RxR,RxN,.. for half a second before )
e) looking a little into QxR
-> beeing convinced that this might be the right move
3) calculating seriously
PART I:
ReplyDelete@ Tempo:
"I have declared the war on lists. . . . System II is very fond of making lists. But once they are made, it lacks the skills and the Short Term Memory slots to use them."
How true! This has implications for GM Kotov's so-called Tree of Analysis. Since a list can be used to represent any type of tree (in computer science), the same "war" must be declared on trees. Thinking in terms of lists (or trees) places the emphasis on the organizing and recording mechanism rather than on the "thinking process" (if System 1 in combination with System 2 can be designated as a "thinking process").
I prefer to think of the chess move selection process as more of a Hegelian dialectic: thesis, antithesis, synthesis. Quoting from Wikipedia:
"Within Hegelianism, dialectic acquires a specialised meaning of a contradiction of ideas that serves as the determining factor in their interaction; comprising three stages of development: a thesis, giving rise to its reaction; an antithesis, which contradicts or negates the thesis; and the tension between the two being resolved by means of a synthesis."
Aox's step-by-step "reasoning" demonstrates this approach (although I'd be the first to accept that Aox probably does NOT have a dialectic approach consciously in mind, much less a Hegelian approach). It is simply a sequence of steps which present themselves to the conscious mind, informed (silently) by the subconscious "reasoning" built upon pattern recognition and prior experience. Since our individual experiences are unique, there is no way to simply copy the "reasoning" process and apply it ourselves. Alas, we MUST build our own processes by trial and error.
GM Kotov's "think like a tree" concept MAY be the correct approach for carefully documenting every possible variation during home analysis (where every variation can be examined ONCE at leisure and written down as the analysis proceeds), but it is a misleading (wrong?) way to approach either solving problems or playing a game. The human's "monkey mind" does not climb trees!
My own processes have changed since this position was originally posted. (I know this because I looked at the position as if it was new and only after thinking about it, I went back to check my comments on the original post. I was somewhat pleasantly surprised to see the difference in my approach.)
Good point. It is not a war on lists but a war on trees! System II has no problems with long sequences, but with the administration of branches.
DeleteCoincidentally I have been sawing down the tree in the front yard today.
DeletePART II:
ReplyDelete1. I formerly started (as did Aox) with a material count to get a "feel" for the tension between material and other elements. When solving problems, I no longer start with material. Why not? Because it "colors" the thinking toward consideration of material gain as if material is the primary consideration in every position, and it is NOT, especially in tactical problems. The primary consideration is: What can the pieces DO? In this example, this time I did not count the material until after reading Aox's comment.
2. My second step involves determining if the King is in the "box". White's King is not in the "box", but Black's King IS in the "box". Whatever solution is found must address this issue.
2. I started with the possible LoAs. White threatens Black's back rank (2:1 on d8, which is "BAD" for Black) and the Black Queen (1:1 on e5, which is B.A.D.). Black will have to address both threats with tempo. The first thought is simple: chop off the WRd5, "killing two birds with one stone." Consider that to be the "thesis" because it removes both White threats at once.
3. The "antithesis" is that White can counter by capturing on d5 in two ways, with either WQd1 or WNc3. Given that the original ration on d5 was (2:2), it appears that Black loses material by capturing first.
4. Contra 3, there is a Function that constrains the White Knight: it is overloaded protecting WRd5 AND protecting (by blocking the LoA of the TWO Black Rooks) the WRc1. In short, it is relatively "pinned" by its defense of WRc1. The ration on c1 is (1:2) without the WNc3 in place, or (2:2) if it remains in place. A pinned piece loses its defensive Function. So, only the White Queen can recapture on d5, but this leaves the WRc1 totally unprotected.
5. After the White Queen captures, Black recaptures on d5 with the BRc5. White can capture on d5, opening the WRc1 to capture.
6. The "synthesis": Black wins material by 1. ... Qxd5.
I really do wish things were this simple!
There are at least 2 reasons to count the material
Delete1) you look at evry! piece and see where it is, easier to not to forget any piece in "the corner"
2) you can make conclusions about the goal of the puzzle
if you are already a piece up then it can be about defendin this piece asf..
I stopped with counting material. It serves no goal and it consumes Short Term Memory slots. System II is very fond of counting. System I cannot process the results. Nor can system II, for that matter. System II should behave itself a bit.
DeleteWar on trees. Why does this sounds so wrong? I'm sure it must be correct! It's the only way to stop system II from proliferating its bad habits.
ReplyDeleteA tree becomes a list if you say at every branch: yes, as you do above.
ReplyDeleteThat is exactly the plan. To cut all the branches of the tree of analysis down and lay them in one long sequence. This means that the administration of the branches must be incorporated in the method. I don't use a list of targets at runtime. I start with the king and when I run out of options I continue with the queen. Da capo al fine. Lather, rinse, repeat.
DeleteThe questions itself are related to the initiative (duplo attack and/or immobility due to lack of space or function). That means that I only continue with a target when there are relevant moves related to it.
https://chesstempo.com/chess-tactics/13884
ReplyDelete@ Aox: A nice problem, illustrating some good points!
Delete1. Black is to move and gain wood (see 2 below for the reason it's about wood gathering).
2. Neither King is in immediate danger of checkmate BUT the White King is within range for a check by BNc5. THIS IS THE FIRST CLUE WHICH REDUCES THE SET OF MOVES TO BE CONSIDERED! The "solution" must (somehow) make use of this fact. That means we have a clue that BNc5 will be involved in the solution at some point AND it will probably involve a capture on d3.
3. Material is even. Pawn structure is roughly equivalent.
4. Weaknesses: The first White move 1. Ba3 creates a hanging (unprotected) piece on a3. THIS IS THE SECOND CLUE WHICH REDUCES THE SET OF MOVES TO BE CONSIDERED! WBa3 will be involved in the solution. Since it is a "weakness," it is likely to be one of the targets.
I presume that White thought that the "pin" on BNc5 (against the undefended BBe7) was sufficient to hold the balance. The "fly in the ointment" is that there is a potential CHECK by that BNc5, which makes the pin illusory - the BNc5 is relatively, not absolutely pinned. THIS IS THE THIRD CLUE WHICH REDUCES THE SET OF MOVES TO BE CONSIDERED!
The WNd3 is B.A.D.; as a consequence, d3 is a critical PoP. Captures WILL occur on d3. There are two Black pieces (BNc5, BBf5) and two White pieces (WBe2, WRd1) which can capture on d3. We "destroy" targets simply because they ARE targets! THIS IS THE FOURTH CLUE WHICH REDUCES THE SET OF MOVES TO BE CONSIDERED!
5. "Finger" it all out!
The immediate capture with check 1. ... BNc5xWNd3+ is "tit-for-tat": Black exposes the LoA by BBe7 on WBa3 BUT allows a balancing LoA after 2. WBxBNd3 against the unprotected BBf5. THIS IS THE FIFTH CLUE WHICH REDUCES THE SET OF MOVES TO BE CONSIDERED!
If Black captures on d3 with 2. ... BBf5xWBd3, then White can counter with 3. WBa3xBBe7. If Black captures on a3 with 2. ... BBe7xWba3, then White can counter with 3. WBd3xBBf5. In either case, material remains balanced, and potential tactical action is over for the moment. Tit-for-tat, 1:1 - not good for Black.
There MUST be something better for Black, because the problem condition implies that Black is to move and "win".
First thought: change the order of captures on d3 - capture with the unprotected BBf5 with 1. ... BBf5xWNd3, which accomplishes TWO things at once. THIS IS THE SIXTH CLUE WHICH REDUCES THE SET OF MOVES TO BE CONSIDERED!
(1) The BBf5 no longer comes into play (is not unprotected) when (if) the BNc5 captures (WITH CHECK!) on d3.
(2) White MUST recapture equivalent material SOMEWHERE to maintain material parity.
White has two possible responses to attempt to maintain material balance: either capture on d3 or c5.
2. WBe2xBBd3 is the simplest response. It is refuted because Black can capture on d3 again, this time WITH CHECK, leaving Black the time to capture on a3: 2. ... BNc5xWBd3+ 3. WRd1xBNd3 BBe7xWBa3 and Black is ahead in material.
2. WBa3xBNc5 postpones the day of reckoning. A simple "counting" technique suffices to demonstrate why this does not work: Black has a chain of TWO potential captures remaining (2. ... BBd3xWBe2 and 3. ... BBe2xWRd1); White has only ONE potential capture remaining (3. WBc5xBBe7. So, Black continues the capturing sequence with 2. ... BBd3xWBe2 3. WBc5xBBe7 BBe2xWRd1 4. WNc3xBBd1 and the potential captures are exhausted - with White down an exchange.
This seems extremely time-consuming; it is NOT. It took less than 30 seconds for me to run through this entire "thought" sequence (a significant portion of which was spontaneously "thrown up" by System 1), followed by confidently stepping through the solution on Chess Tempo with no errors.
Well i chose this puzzle because in my eyes its the same cheme as the puzzle above. The Solving method is similar to http://temposchlucker.blogspot.com/2018/09/war-on-lists.html?showComment=1537596639332#c303775803374460730
DeleteThere is an exchange orgy at several potential squares f5,e6,e7,c5,a3,d3,e2,d1..?..
The point here ist that you dont calculate CCT but you calculate: maximize potential followup captures for black CCT says: Calculate first Ne4+ and Nd3 check, Maximisation od followup captures say calc Bd3 first ( follow up captures are Be2 and Bd1 )
Exchange orgys can only win when you start AND end the series of captures or you win the exchange. The only possible way do win the exchange in a forced manner is BXd3,Bxe2,Bxd1
So in my eyes both puzzles are extrem similar