The Best Moves Never Played - Key Positions 1

  • WIM energia
  • | 18.01.2013
  • | 16917 Aufrufe
  • | 18 Kommentare

A strong chess player calculates variations well but also knows when to stop and evaluate the resulting position, resorting to experience and knowledge of typical positions. An experienced player will know that certain positions are better for him and will not attempt to calculate everything to the end but rather use these final positions as a guide for which path to take. This topic will span several articles in this column, "The Best Moves Never Played".

The easiest way to illustrate the concept of key positions is through endgames. A recent example from the 75th Tata Steel Chess Tournament will serve our purpose here. Nakamura misplayed the opening, ending up in losing endgame, in which Sokolov was convincingly pressing. After putting up a tough defense, which Nakamura is famous for, Sokolov missed a win and the game ended in a draw.

Just how losing the position was can be perceived from the players' after-game interviews. Hikaru criticized his own play in the opening and then commenting on the endgame he mentioned that "he [Sokolov] played the one plan where I had some hopes of a draw. I mean it's still completely losing but I managed to get a little bit lucky". Sokolov in his 13-second interview resorted to humor: "I had a completely winning game. If I don't kill myself tonight I am going to live 1000 years". Nevertheless, the game ended in a draw and it is our job to figure out where Sokolov went wrong.

Let us start from a rather late stage of the game.


In the above position black is up an exchange for a pawn. The bishop defends two weaknesses at the same time, but the black rook and king are active and black's position is winning. White is in zugzwang - he cannot move his bishop because either the g or the a-pawn will fall, so he can only move his king on the f2 and g2-squares. If the king goes to h2 the e2-pawn will fall. There are two plans possible to make progress in this position:

1) the f4-break, which Sokolov played in the game;

2) Bringing the king to h3 to take away the g2-square from white king and then winning the g3 pawn using zugzwang.

Let us start from the second plan, which to me seems easier to implement and does not give white any counterplay.

It seems that the above variations are straightforward and not hard to follow. Now, let us examine the continuation that Sokolov chose. Surely, his choice required tons of calculation, so my guess is that he misevaluated some resulting position or missed some defensive idea by Nakamura.

Why is the last move a mistake? Black MUST not allow f5-f6. The far advanced f6-pawn will deflect black's rook from a checkmating pattern. This is black's main winning idea:

And this is what happens when the rook is tied to guarding the f6-pawn:

Let's look at the variations where black did not allow f5-f6 but before doing so, we shall point out the defensive set-up for white. The bishop has to defend the a3-pawn, and it should also control the h2-b8 diagonal. If black's rook steps on the d-file it should go to b4 to control the d2-square.


From the discussed line, there are two critical zugzwang ideas that one had to see, in order to figure out how to win this endgame:

1)  It is white to move and white is in zugzwang. The bishop has to move away from its ideal square d6. If it goes to b4 then Rf4 wins a pawn, and if it goes to c7 then Rf2-Re2 wins a pawn (the f6-f7 is not scary because the f8-square is not defended by the bishop).


2) Rf2 and push h2 is a threat, and if bishop goes to d6 then f6 falls.


In the two positions above white's king couldn't move. The line below is an alternative win for black if white plays actively.

Once again, there were two critical positions that one had to evaluate correctly, when going for the above line.

1) The way to make progress is to push the h-pawn and then trade it for the e-pawn, which will result in...


2) Generally, the position with fixed a- or h-pawns is theoretically complicated and according to Dvoretsky's "Endgame Manual" even GMs have trouble winning it. However, in this particular situation one had to see the plan of cutting white's king along the g-file, running with the king to the a-pawn and capturing on a3 with the rook.


Now, back to the game, where white retained both the a and e-pawns and black had no way to make progress.

I hope this article illustrated the idea of critical positions, where one had to know how to evaluate them before going for a particular line. Next week we will look at middlegame positions where similar concepts apply.


  • vor 4 Jahre



  • vor 4 Jahre


  • vor 4 Jahre


    that was so easy

  • vor 4 Jahre

    NM Splane

    I think variation 2 is overly complicated. After 51... Rg3+! 52. Kf2 h4 53. Bd6 Rf3+ 54. Kg1 h3 55. Be7

    Black has a simple win with 55 ...Kf4   56. f6 Ke4. Now White has nothing but king moves in the g1-h1-h2 triangle. Black moves his king to e6 and if necessary to f7 to force White's king to the first rank. As soon as White's king leaves h2 the rook sacrifice on f6 works.

  • vor 4 Jahre



    I think what happens is 91. e8=Q+ Kxe8 92. Bd6, and White uses his king to prevent Black's king from coming over and prying out the pawn.

  • vor 4 Jahre


    in the game

     Nakamura, Hi (2769) vs. Sokolov, Ivan (2663)

    Variation 3 | Wijk aan Zee NED | Round 2| 13 Jan 2013 | ECO: A22 | 1/2-1/2
    couldn't black just promote the pawn and win?
  • vor 4 Jahre


    :) Nice endgame !:)

  • vor 4 Jahre



  • vor 4 Jahre


    its played today :)

  • vor 4 Jahre


    from today HOU game ?

  • vor 4 Jahre


    Flayy, it looks like Bd6 holds things together.

  • vor 4 Jahre


    Flayy, the bishop can go to d6 safely.

  • vor 4 Jahre


    It was a draw after the last move 90... Rc4? How can white defend now? There is the Rxb4 threat with passing the a pawn, so white has to give up the e7 pawn. Or what am I missing?

  • vor 4 Jahre


    Nice article!  Complicated and very interesting endgame.

  • vor 4 Jahre



Nach oben

Antwort senden: