Monday, June 06, 2016

Grand scheme of an attack

This weekend I have been thinking about what is the best way to proceed. Once, I analyzed a lot of sacrifices at f7 from Polgars middlegame brick. I could distil a few rules out of it about when such sacrifice will work and when not. These rules help me to judge such positions fast. The accompanying patterns are burned into the brains, so it works fast too.

At first sight, this seems a logical way to go. Knowledge is acquired by looking at common factors, and both the knowledge and the corresponding patterns are stored, and easy to retrieve. Yet there are a few drawbacks with this. I studied about 120 positions, and only about 28 or so fitted into the same template of an attack on f7. This means that there are 90 positions that don't fit into any general scheme. Maybe it is possible to categorize these too, but I might very well end up with 70 categories or so.

Although the method of categorizing positions with the same features into templates works in itself, there are so many exceptions that it is not practical. If we look at the tactical themes like pin, double attack etcetera, these are so common that all tactics fit in only a few templates. That is the kind of stuff I'm looking for.

I once made a general scheme for duple attacks. It contains (a subset of) the following elements:

  • Attacker
  • Target
  • Attacking square
  • Target square
  • Route from attacker to attacking square
  • Route from attacking square to target square
  • Defender

This grand scheme is very useful to describe any duple attack. If the attacker is far yet from its attacking square, it needs preliminary moves to get there. All these moves must be with conservation of tempo. If the target isn't on its target square yet, it must be forced there by coercion, or attracted to it by a capture or so. All preliminary moves need to conserve the initiative (by CCT).

At CT I had three failures, which I initially denominated under the category "knight fork". But with a closer look, there were quite different elements involved.

Diagram 1 White to move
3r2k1/8/p2r1Q1p/1p5q/2p3pN/P1N5/1P3PPP/4R1K1 w - - 0 1

The targets are Kg8 and Qh5. They are already standing on their target squares.
The attacker is Nc3, which is still a long way from its attacking square f6.
The defender Rd6 is overworked. It has to defend the attacking square f6, and its brother in arms Rd8. The attacking square is occupied by whites own queen.

First the attacker Nc3 must get to the attacking square f6 with tempo.
1.Ne4 attacks both Rd6 and attacking square f6. In fact it is a knight fork too.

Blacks rook is under attack. There are 5 standard answers to an attack which must be checked:
  • Capture the attacker. The knight on e4 cannot be taken.
  • Counter attack. The options are very limited. The white queen cannot be taken, since it frees the attacking square for the white knight. The white queen is a desperado, which can take on d8 any time with check.
  • Escape. Moving one of the targets from their target squares looses the black rook on d6.
  • Interference. A knight cannot be interfered.
  • Pin the attacker. Qe8 pins the attacking knight on e4 to the mating square on e1. Yet that is not enough. The black queen has moved from one target square to another, and the white queen is still a desperado that can take on d6. 1.Ne4 Qe8 2.Qxd6 Rxd6 3.Nf6+ wins the exchange.
The following "knight fork" uses the same grand scheme, but the accents differ.

Diagram 2. White to move.
4r2k/6pq/p2n3p/3Pr3/1Pb3PN/7P/5RB1/Q2R2K1 w - - 1 1

Target 1, Kh8 is already sitting on its target square.
Target 2, Re8 is not yet on its target square e5.
The attacker is Nh4.
The attacking square is g6.

1.Qxe5 is a preliminary move designed to put the black rook e8 on the target square e5 with tempo.
Check the 5 standard answers:
  • Capture the attacker
  • Counter attack
  • Escape
  • Interference
  • Pin the attacker
 None of these is sufficient.
1. ... Rxe5 2.Rf8+ distracts the black queen from defending the attacking square g6.
2. ... Qg8 3.Ng6+ forking. The move order is important in ortoe conserve the initiative.
3 ... Kh7 4.Rxg7 getting the queen back.

I found that all duple attacks can be described with this grand scheme and the five standard answers. All tactical patterns are related to this scheme. Unless the position is about a trap or promotion, of course.


  1. Today I managed to surpass the 1800 blitz landmark at CT. It took me 74 days. 100 down, 200 to go.

  2. Congratulations! I am really surprised what drives you so much and what is your source of perseverance. I could not stand such a long period of testing ideas on my own! You are a really tough man Tempo! Good luck reaching next goals!

    1. I feel not bound to an opinion. When new facts arise, I exchange my opinion for a new one. And so my opinions always reflect the latest state of my knowledge. This means, that for me, my opinions are always true. Since if not, I simply adjust them. A lot of people have preferences how the truth should look like. I have not.

      This strong belief in my opinions causes that there is little difference between my knowledge and my actions. I act according my latest knowledge. No energy is wasted to doubt. When in doubt, I come into action in order to resolve the doubt. I simply cannot help that.

      Take for instance the salt mines. The reasoning behind it was full of marginal comments which casted doubts. But the logic within the reasoning itself was consistent. Hence I act as if it is true, until more knowledge arises. And that was exactly what happened. After two months salt mining, I understood that the salt mines could indeed lead to improvement in certain areas. But not in the areas where I need to improve NOW. And so the doubt is resolved. I'm sure there is a role for salt mining in the future, IF I manage to improve in the more relevant areas now.

      If you read my blog now, and compare it to 2007, you will see that I reach in most areas the same conclusions now, as I did then. But now these conclusions are tested in depth, and all the good willing advice produced by my fellow chess players and well intended chess authors is stripped from all nonsense, leaving a 0.01% or so of useful information. The diagnosis of the problem to solve is crystal clear now, beyond any doubt. So all energy goes into the search for the remedy now.

      Gee, how often did I use the word "now"?

  3. Congratulations on your new rating high!

    Also, belated congratulations on finding a way to activate the Fusiform Face Area in your May 29 post. I hope Aox in particular appreciated that.

    Back on April 20 you said you'd reached 1791 at CT. Was this a different measure, or did you just go down and up? If so, to what do you attribute the oscillation? And what role does the vulture strategy play at the moment? --mfardal

    1. "All time high" is an extreme by nature. Oscillating around an average ± 50 is normal. My average now is about 1750. Was 1650 in December, when I started again after a two year break.

      Sometimes I remember the vulture view, sometimes I forget. The accent lies now on the post mortem, where I investigate how the position fits in the grand scheme, as described in this post.