A Fortress Revisited

I am not a very technical person and I do not check how many people actually read the blog posts I write. I know that it is hundreds when I put up a new publishing schedule. We got about 300 votes for the Gelfand covers – and I would assume that most people voted, as it is so easy.

Obviously I like that my books have a greater readership than the free stuff, which probably has a lot to do with the effort I put into them (while the blog posts are meant to be contemplations and opening up for debates that will help me write BOX next year more than anything).

So I always get surprised when a really strong player says that he has read it. This is another case of this (based on this post).

Apparently I was not right when I believed that Black could maybe hold the fortress in this position:

 Giri – Kraemer, Germany 2014

111…Kg8? 112.Nd6 Rf6 113.Rc8+ Kh7 114.Rd8!

Black is in zugzwang.


Apparently White was winning all the same. The following interesting analysis was sent by email from GM Karsten Mueller, maybe the world’s leading authority on the endgame.


I foolishly speculated that Black was holding, but look at this:

112.Rc8+ Kg7 113.Rg8+!

113…Kh6 114.Ng5 Rf5

114…Rd7 115.Kf3 Re7 116.Kf2 Ra7 117.Ke3 and White will be able to activate his king. Black is too passive.

115.Kg1 Rf6 116.Rd8 Kg7

117.Rd2 Rc6 118.Rd7 Rf6 119.Ne4 Ra6 120.Nd6

White wins the f7-pawn and eventually the game.

8 thoughts on “A Fortress Revisited”

  1. Fool that I may be, I don’t see the win of the f-pawn in the final diagram. Doesn’t Black just continually check the King along the a-file with his Rook or if the King hides just locate on the f-file? I played the line against Fritz and although it insisted that it maintained a nearly 3 point advantage, it started repeating positions without finding a break through.

  2. Of course I used a computer looking at it. But it is really pretty useless unless the forced win is within its horizon. It just says +2 or something no matter what. Karsten clearly spent a bit of time working out the solution, obviously with the help of a computer, but because he is great, not because of the engine…

  3. Very interesting! I guess these kind of variations are very hard to find over the board, especially after 4 hours of play.

  4. I think you got that wrong Jacob: Most of us, your frequent readers, could not be bothered to even make a single click in a poll! Your posts appear often referenced by websites like Chesscafe and the like, so you are talking to thousands each time.

    By the way your “Chuchelov” discussion about computer use went viral via Twitter.

    Greetings from Mexico

  5. @Murrel: I see the white K walking g2-f3-f4-e5-d4-c5-b4 and then he wins the pawn without the black R being able to cover it on the f-file since both f4 and f5 are under control of white.

  6. And another point might be that with black’s f pawn exchanged for white’s g pawn, the fortress is no longer a fortress and therefore white should win with extra piece and h pawn.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top