There are four potential targets:
- King
- Queen
- d8Q
- Bishop
In the diagram of the previous post, we couldn't discover anything that signals that we should start with 1. Ng4. This means that we can only find 1. Ng4 by trial and error in combination with luck. If you are lucky you find 1. Ng4 within a few minutes. If you are not lucky you can think for days on end.
That is of course the horror scenario that we want to prevent. That is why I start the analysis now after 1. Ng4. If we can determine fully what is going on after 1. Ng4, then we might be able to find out how that is telephoned to the position before 1. Ng4.
 |
| BLACK to move (after 1. Ng4) |
Case A
No matter what black does, 2. Nxf6 wins.
Except for two scenarios:
- 1. ... Qxd7 (A1)
- Bishop moves to anywhere but stays in contact with the promotion square d8 (A2)
- It captures the only defender of the promotion square
- It defends pawn d7
- it removes the promotion threat
- it attacks the rook
- It saves the bishop
- Hence the promotion square remains defended
Case B
No matter what black does, 2. Rg8+ wins.
Except for three scenarios:
- Queen moves to a square where it is in contact with d8 (B1)
- Queen moves to a square where it is in contact with the bishop (B2)
- Bishop moves to anywhere but stays in contact with the promotion square d8 (B3)
Ad B1
When the Queen is defending the promotion square, we cannot sac the rook since the rook is needed to protect the promotion square.
Ad B2
When the Queen defends the bishop, we cannot sac the rook since the rook is needed to protect the promotion square after for instance 1. ... Qf5 2. Rg8+ Kxg8 3. Nxf6+ Qxf6
Ad B3
- It saves the bishop
- Hence the promotion square remains defended
- Hence f6 has an additional defender
UPDATE
The bishop as such isn't a target in this position. Play d8, and you will win it. But you will find out de power of blacks a-pawn. You probably must sac a piece back to stop it.
That leaves us with the K, Q or d8Q as potential targets.
A target must be either immobile OR part of a duplo attack. Otherwise it will escape.
The natural target of this position seems to be the promotion square. There is nothing more immobile than a promotion square at a moves distance.
There seems no more unnatural move than 1. Ng4, since it abandons the main trump of white. So what's the power of 1.Ng4 ?
Although it abandons the pawn, it poses a threat. The bishop is the defender of d8. When it is captured, promotion becomes possible.
Black can take the d-pawn, but that "schlucks" up a tempo.
When the pawn is gone, the promotion square stops to be a target. So we are in need of a either one new immobile target OR two targets for a duplo attack.
For me this is a total new way of thinking (system II) with total new patterns (are you awake, system I?).
It is a new pruning method. You look for one immobile target OR two targets for a duplo attack. There is nothing else. There can't be. You can abandon a target only with tempo. But you must know what your new target(s) are. Everything else can safely be pruned as being not potential viable.
If black decides to NOT take on d7, only moves that relate to d8 OR the bishop are relevant. Other moves can't influence the threat. Hence they can be pruned from the tree of analysis without further worrying.
Let us grow a brain for target awareness AND tempo's!!
PART I:
ReplyDeleteI think this may be a pursuit of the white rabbit down the rabbit hole. Back into the briar patch, Brer Rabbit, with Alice!
I recognize that the following “analysis” is merely a long-winded recap of your quite succinct assessment above. I am still convinced that our conclusions regarding PoPs, LoAs and Funs are the correct starting point for finding solutions to tactical problems, whether in a game or as an isolated problem.
Lay on, MacDuff! (I.e., begin a vigorous attack!)
(Please note that the original Shakespeare is NOT “Lead on, MacDuff!”, implying that if you take the lead, I will follow you.)
In the original position, d7 (WN vs BQ) and d8 (WR vs BB) are B.A.D. These are the available PoPs. The d7 PoP is defensive for White, whereas the d8 PoP is offensive for White. Since we are looking to increase the White advantage, our high regard for the initiative tells us to focus on the Pawn promotion as the highest priority goal.
As a side note, Ne5 appears to be just BAD, because it is DIRECTLY attacked twice and DIRECTLY defended once (by Bg3) (simple counting). But this initial impression of e5 is mistaken because the ratio of attackers/defenders is NOT 1:2 – the Function of the Black Bishop to stop the d7-Pawn promotion removes it as an attacker on e5, and the Black Queen cannot capture first without just losing it for insufficient compensation (1. … QxNe5 2. BxQe5 BxBe5+ 3. g3 Bf6 4. d8(Q) Bxd8 5. Rxd8). Ergo, the Ne5 is “safe” (at least for the moment).
This preliminary identification of PoPs, LoAs and Funs is the “filter” mechanism for eliminating spurious moves from consideration. These considerations in conjunction with “WHO HAS THE MOVE?” will often drastically reduce what must be considered as candidate moves.
PART II:
ReplyDeleteSo what do we use to orient ourselves toward these preliminary points? If given a B.A.D. situation, we can increase the number of attackers on the B.A.D. square(s) OR we can “thin” the ranks of the defenders of the B.A.D. square(s) through exchanges or diversions. We have no resources for increasing the number of attackers on either d7 or d8, so the only option is to either exchange or divert the defenders. There is no way to “attack” the Black Queen (for either exchange or diversion) because it really has no critical function and is therefore maximally mobile. By that process of elimination, we arrive at either an exchange of the Black Bishop or a diversion of it. Since there is nothing we can do to divert it initially away from covering the d8-square, the only remaining option is to exchange it. We CAN “attack” it DIRECTLY in one move with 1. Ng4.
Here we encounter the first sidetrack of our attention. 1. Ng4 REMOVES the defense of the d7-Pawn, allowing the Black Queen to capture it. Since our primary objective is the (eventual) promotion of the d7-Pawn, our first “gut” (System 1) reaction is “This CANNOT be the right move – I lose the d7-Pawn!”. This is where System 2 MUST come to the rescue and OVERRIDE this reaction: “Maybe System 1 is right and it’s NOT the right move, but there is still something to investigate because it does threaten to capture the defender of the d8-square.”
Black cannot allow the d7-Pawn to remain threatening to promote AND also allow the Black Bishop to be eliminated.
The most obvious variation is simply to hack off the d7-Pawn with 1. … Qxd7. This also attacks the White Rook. Now the Black Bishop can be allowed to be captured without any adverse effect on Black.
Now System 1 comes back into play (with pattern recognition) – “Hmmm, I can attack the Qd7 by capturing on f6. Too bad there isn’t a fork available on f6, because f6 is a tactical base for a Knight fork with one tactical target on d7.” System 2 picks up the thread and continues logically, “The Black King is now totally immobile. That means it is a candidate for a forcing move that makes it a second tactical target. Is there a forcing move available? YES! 2. Rg8+ FORCES the Black King to capture it. This now sets up all the preconditions for a successful Knight fork on f6.” 1. Ng4 Qxd7 2. Rg8 Kxg8 3. Nxf6+ Kf8 4. Nxd7. Evaluation: White is winning. So the first variation “works” for White.
PART III:
ReplyDeleteBUT (there’s always one of these stinking up the logic), what if Black does something different? As long as he can control d8 with the Black Bishop, he doesn’t need to capture that pesky d7-Pawn immediately.
What possible Black Queen moves are available that reinforce the control of d8?
1. … Qd5
1. … Qa5
1. … Qb6
1. … Qg5
Let’s assume that White just captures the Black Bishop, because that was the purpose of moving 1. Ng4. After 1. Ng4 Qd5 2. Nxf6 the d7-Pawn is protected and will promote on the next move. 2. … Kxf6 3. d8(Q)+ Qxd8 4. Rxd8 winning.
The same logic applies for all four Black Queen moves.
By process of elimnation, we arrive at a move of the Black Bishop to maintain the defense of the d8-square. But the Black Bishop has very limited mobility – it cannot leave the h4-d8 diagonal without allowing the d7-Pawn to promote, and there are only two “safe” squares on that diagonal – e7 and g5. Either move allows the White Bishop to check on e5, forcing Black to block the check (because the Black King is already totally immobile). Black can either interpose the Black Bishop (allowing the immediate promotion of the d7-Pawn) or the f7-Pawn. White then captures on f6 with the White Bishop with check, again with only two alternatives: capturing on f6 or moving the Black King to f7. The capture on f6 allows a recapture by the white Knight, and the d7-Pawn is protected and will be promoted next move. Moving the Black king allows White to capture the Black Bishop on e7. If the Black King recaptures, the d7-Pawn promotes with check. If the Black Queen captures on d7, then White Knight forks King and Queen on e5.
White is winning.
Sorry: I left out something in my analysis, so I deleted the comment and then re-entered it here.
ReplyDeleteRegarding Ad B2:
After 1. Ng4 Qf5, instead of trying to force the Knight fork idea, simply continue with the previous logic: chop off the defender, then promote the d7-Pawn. 2. Nxf6 protects the d7-Pawn, which will promote next move. 2. ... Qxf6 3. d8(Q) Qxd8 4. Rxd8 and the Rook has plenty of time to stop the Black a4-Pawn. There is no requirement to make the Knight fork trick work in every variation.
I did look at a different fork idea, but it doesn't work (by one tempo!). 1. Ng4 Qf5 2. Rg8+ Kxg8 3. Nh6+ Kg7 4. Nxf5+ g6xf5 5. Bc7 a3 6. d8(Q) Bxd8 7. Bxd8 a2 and the White Bishop cannot get on the long diagonal a1-h8 in time to stop the a2-Pawn from promoting. I stopped looking along this line at this point. This morning, I gave this position to GM Stockfish, who recommends 8.f4 Kg6 9.Kg3 a2 10.Be5 Bd8 -/+ (-1.21). Not the result we are looking for - cut it off!
This illustrates another pruning concept. Our expectation (based on our preliminary evaluation of the position) sets a lower boundary on what moves can be considered. As soon as we find something that concretely contradicts our expectation, we can prune that variation. In the original position, our expectation is set by the fact that WHITE IS TO MOVE - and therefore White SHOULD end up in a better (if not winning) position. At any time, if we find that White is no longer winning, we can stop looking at that variation.
I know that is (maybe!) NOT the right thing to do in a game, but it IS the right thing to do when trying to solve a tactical puzzle under a time constraint.
Sorry, one more correction: GM Stockfish recommends 6. f4 Kg6 7. Kg3 a2 8. Be5 Bd8. Mistakes were made. . .
ReplyDeleteUpdate in blue
ReplyDeleteATH 2000 at CT in Blitzmode :))
ReplyDelete@ Aox: Congratulations!
ReplyDeleteIs there anything particular you are doing differently?
i simply learn! them all. in a few months i will look if i am realy better at puzzles i never saw before ( i doubt that ). i can calculate my performance on puzzles i never saw before from the statistics of CT for premium members.
DeleteI dont only learn position+mainline, im looking for the tactical weaknesses, learn the refutation of my errors, try to "understand" the general concepts ( forcing moves first, more attackers to the goal ... ), think about how to find the solutien quicker asf.
But, most of that, i do since a long while.
second thing i do is some very!! moderate!! visualisation training with several completly different tools. And i play some blitzgames as warmup for my puzzletraining