An Endgame from New York International

This position arose in Lund – Yedidia, New York International 2017 (8). The game contains interesting DEB material as well as typical N vs. B endgame ideas.

This position is only slightly better for White. Black has the isolated d5–pawn and a potential bad bishop in Bc8 (right now it’s a Double-Edged Bishop, DEB – see my recent book The Secret Life of Bad Bishops). However, the presence of dark-squared bishops are to Black’s advantage, and the pawns on a3/b4 gives Black a point to attack with a7–a5. Regrettable for Black, a move he never makes in the game.

23.Bb2 Ba6 24.Nd4 Bd6 25.f4!?

My plan is to try to create weaknesses on the kingside with the help of the queen. With the queens still on the board, Nd4 can prove more flexible than Black’s light-squared bishop (right now on a6, far away from e6/ f5 and the kingside). At the right moment then, I can exchange queens and enter a favorable version of the endgame B+N vs. 2B. In the game, this plan worked well – but Black could have defended better.

The move played prevents Black from going 25…Be5 to exchange the knight on d4. The resulting opposite-colored bishops with queens on the board would not be winning for White.This is a good example of how the presence of dark-squared bishops is to Black’s advantage, and how it restricts White and give Black additional options.

25…Bc4 26.Kf2

Freeing the queen by protecting e3.

26…Qe8 27.Qd1


27…a5! 28.bxa5 bxa5 29.Qg4 Among many moves, Black has the boring 29…Bc5= aiming to exchange on d4. However, Black can also be ambitious and play something like 29…g6!? There is an attacking point on a3, and the open b-file can turn out to be an occasional problem for White. Black is not without chances here, and White suddenly needs to be alert too (unlike the game).


With the bishop on c4 shut off from the kingside, Black has to be a little careful here.

28…Bf8 29.Nf5 Qg6 30.Bd4 a6 31.Bb6?!

Here I missed an excellent chance to exchange dark-squared bishops with 31.Bc5! In beginning time-trouble I mistakenly thought that I would get another chance later.


I like 31…Qxg4 32.hxg4 Bd3 more: the doubled pawns on the g-file is bound to hem White to some degree – doubled pawns just are less flexible.

32.Qxg6 hxg6 33.Nd4 Bd6

White has a slight advantage, but nothing tangible yet. The following exchange of dark-squared bishops is a misjudgement by Black.

34.Bc5 Bxc5?

After this exchange, White gets real winning chances.

Better is 34…Bc7 35.Ne6 Bb8 36.Bb6 Bf5 37.Nc5 Bc8 White is slightly better, but these endgames with a Black bishop pair are more likely to be drawn than won for White.

35.bxc5 Kf7


Placing the knight on a light square in front of the passed pawn. From here, the knight covers important squares (a5–b4–d4–e7–d8) and threatens several knight forks. It’s a typical scenario in N vs. B place the knight on the same square as the bishop like this.


36…Be4 was a better square for the bishop.


Placing the pawns on the dark squares and fixing the doubled pawns on the g-file.

37…Ke6 38.Kf3?!

A careless move in time trouble.

38.Ke1! On the way to d4. 38…d4 39.e4!+–


38…Bd3! Now the fork 39.Nb4 doesn’t quite work in view of 39…Be4+ check.

39.Nb8+ Kc7 40.Nxa6+ Kc6

Black is hoping to passify the knight on a6, tied down to c5. Other ideas are d5–d4 with counterplay, or bringing the bishop back to c8, winning back the c5–pawn. Unfortunately for Black, White can counter all these ideas.


The White king is on the way to d4, at the same time putting Black in zugzwang by covering both the e2– and f1–square.

41.g3? Bf1 42.Kf2 Bh3 43.Ke2 Bc8 44.Nb4+ Kxc5.

Again a slight advantage for White, but a draw as the more likely result.


  • 41…f5 42.g3 Zugzwang again.
  • 41…Bb3 42.Ke2! within distance of d4. (42.Ke1?! allows 42…d4 43.exd4 Kd5 44.Nc7+ Kxd4 45.c6 Kc5 46.Ne8 Kxc6 47.f5! [47.Nxg7 f5!] 47…gxf5 48.Nxg7 Kd6 Is this winning for White? Probably, but it seems unnecessary compared to the game where I keep control with one precise move.) 42…Bc4+ 43.Kd2 Bf1 44.Kc3! (44.g3?! Bh3 45.Kc3 Bc8 46.Nb4+ Kxc5 47.Nd3+ Kd6 48.Kd4± Advantage, but enough for a win? Again, I would say probably not.) 44…Bxg2 45.Kd4+–
  • 41…Bd3 fails to a knight fork: 42.Nb4+

42.Nxb4+ Kxc5 43.Nc2 Bb5

Black cannot go for the a-pawn: 43…Kb5 44.Nd4+ Ka4 45.Ne6 f5 46.Nxg7 Kxa3 47.h5 gxh5 48.Nxf5+–

44.Nd4 Bd7 45.Ke2 Bg4+

45…Kc4 46.Kd2 Black is in a kind of zugzwang again and will have to allow the king to c3. 46…f5 (46…Kc5 47.Kc3 Bc8 48.Nb3+ Kb5 49.Kd4 Ka4 50.Nc5+ Kxa3 51.Kxd5+–; or 46…Bc8 47.Nc6 Bd7 48.Ne7 Be8 49.f5 gxf5 50.Nxf5+– Threatening g7 and a knight fork on d6.) 47.g3 Zugzwang.

46.Kd3 Bd7 47.Nb3+ Kb5 48.Kd4 Be8

48…Ka4 49.Nc5++–


The regrouping is perfect.

49…Bf7 50.g4

Taking more spacing on the kingside before releasing the knight to attack the g7–pawn.

50…Bg8 51.g5 Bf7 52.Nb7!

This is the 3rd light square for the knight from where it is able to perform a knight fork (beside a6 and c6 earlier). Now the knight will jump to d6–e8.

Also possible was 52.gxf6 gxf6 53.Nd7 f5 54.Ne5 Be8 55.Kxd5+–


52…Kc6 53.Nd8++–

53.Nd6+ Kc6 54.Ne8 fxg5 55.hxg5


55…Kd7 56.Nxg7 Ke7 57.a4 Kf7 58.a5 Kxg7 59.a6+–

56.Nxg7 Kd6 57.a4 Kc6 58.f5 Kb6 59.fxg6


1,170 total views, 5 views today

Leave a Reply