Chess Piece

FIDE Candidates Tournament
Yekaterinburg, Russia
March 15–April 5, 2020
First Half Standings (round 7 of 14)

1–2 Maxime Vachier-Lagrave FRA 2767, Ian Nepomniachtchi RUS 2774, 4.5/7

3–6 Fabiano Caruana USA 2842, Alexander Grischuk RUS 2777, Anish Giri NED 2763, Wang Hao CHN 2762, 3.5/7

7–8 Kirill Alekseenko RUS 2698, Ding Liren CHN 2805, 2.5/7

Note: Tournament suspended after 7 rounds

The FIDE Candidates Tournament began on March 16th, in Yekaterinburg, Russia, with eight of the top players of the world to play a double round-robin over 14 rounds to decide who would challenge the reigning world champion, Magnus Carlsen, in December for the title. Due to the coronavirus pandemic though, the tournament had to be suspended at the halfway point, to be resumed later.

In the meantime let us take a look at this very difficult endgame from round six.

The Russian-born Dutch GM Anish Giri (born June 28, 1994) participated once before in the 2016 Candidates Tournament 2016 (Moscow, Russia) where he drew all 14 games. Here in Yekaterinburg he broke the draw streak by losing to Nepomniachtchi in round 1. After four draws he finally scored his first win ever in the Candidates.’ And he really earned it — the game was even for a long time but he persisted into a difficult knight-and-pawn endgame and finally won after a bad mistake by his opponent who was exhausted after seven hours of play.

These endgames are really difficult. In Dvoretsky’s Endgame Manual the author says that “We shall not be making a systematic examination of the endgame in which a knight faces a knight and pawn: its theory is quite complex, and in my view, rather chaotic. There are no principles which are operative for many positions; the evaluation and the course of the struggle depend entirely upon the concrete details.”

Just some rules of the thumb (taken from Dvoretsky’s book):

• The peculiarities of the knight, such as its “distaste” for rook pawns, or its ability to fork pieces or win tempi by checking the enemy king, function here.

• The knight must quite often be sacrificed in order to obtain a “knight versus pawns” endgame.

• Botvinnik’s Formula. “Knight endgames are pawn endgames”: that’s something Botvinnik once said. What he had in mind, is that many of the laws of pawn endings apply equally to knight endings. The same high value is given, for instance, to the active position of the king or the outside passed pawn. Such techniques as the pawn breakthrough, shouldering, the various methods of playing for zugzwang, and so forth, are seen constantly, not just in pawn endgames, but also in knight endgames.

• Pawns on the same side. Is it possible to convert an extra pawn, if all the pawns are on the same side? Practice has shown that the chances of success are greater in knight endgames than in any other type of endgame — with the exception of pawn endgames. For example, the “four vs. three” position is considered a win.

Alekseenko, Kirill (2698) — Giri, Anish (2763) [C54]
FIDE Candidates 2020 Yekaterinburg (6.2), 23.03.2020

1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5

IM Roven Vogel reviewed Wesley So’s chessbase DVD on “My Black Secrets in the Modern Italian” (which, by the way, all Filipino chess players are required to buy. Go to and click “shop”) points out some essential motifs and ideas expounded on by Wesley:

• The very modern concept of pushing back the Bg5 with h6/g5 followed by Nh7 and h5/h4/g4 to attack the king.

• He also suggests going for Kg7 followed by Ng8/Nge7/Ng6.

• Furthermore, Wesley advises black to play the important move …a5 instead of …a6 to gain more space, because …a5 allows you to try out the optional maneuver Qb8/Qa7.

You will see from current tournament praxis that these ideas are regularly being implemented.

4.0–0 Nf6 5.d3 d6 6.c3 0–0 7.Re1 a5

Motif no. 3, right?

8.Nbd2 Be6 9.Bb5 Ba7

This move is actually new in this position. As given in no. 3 above, the logical follow-up to 7…a5 is …Qd8–b8–a7. Let’s take a look at an example: 9…Qb8 10.Nf1 Qa7 11.Be3 Bxe3 12.Nxe3 Ne7 Black is preparing …c6 and if possible also …d5. 13.a4 Ng6 (13…c6 14.Bc4 d5 15.exd5 cxd5 16.Bb5 there is some uncomfortable pressure on e5) 14.Bc4 Bxc4 15.Nxc4 At the moment there are equal chances for both sides. White can still go for d3–d4. So,W (2810)-Ding,L (2777) Tbilisi 2017 1/2 47.

10.Nf1 Ne7 11.Ng3 c6 12.Ba4 Ng6 13.h3?!

This was the perfect time to play 13.d4 for one, it prevents Black’s next move.


Black has equalized and can look forward to a fight on even terms.

14.exd5 Nxd5 15.Bc2

Not 15.Nxe5? Nxe5 16.Rxe5 Bxf2+! 17.Kxf2 Qf6+ Black wins material.

15…Qc7 16.d4 exd4 17.Nxd4 Rae8 18.Bg5

Don’t fall for 18.Nxe6? Bxf2+! 19.Kxf2 fxe6+ 20.Kg1 Qxg3 Black is clearly better, maybe even winning.

18…Ndf4 19.Qd2 Bd5 20.Rxe8 Rxe8 21.Re1 Re5 22.Bxf4 Rxe1+ 23.Qxe1 Qxf4 24.Qe8+ Nf8 25.Bb3 Bxd4 26.cxd4 Bxb3 27.axb3 Qf6

[27…Qxd4?? is refuted by 28.Nf5 (threatening the queen and mate starting Ne7+) 28…Qd1+ 29.Kh2 g6 30.Nh6+ Kg7 31.Qxf7+ Kxh6 32.Qxf8+ Kh5 (32…Kg5 33.f4+ Kh5 34.g4+ Kh4 35.Qh6#) 33.g4+ Kg5 34.f4+ Kh4 35.Qf6+ and mate]

28.Qe4 g6 29.Ne2 Ne6 30.h4 h5 31.g3 Qd8 32.Qe5 Qb6 33.d5

Removing his isolated queen’s pawn from the board. That’s one less weakness to worry about.

33…cxd5 34.Qxd5 Kf8 35.Nc3 Qc7 36.Ne4 Qc1+ 37.Kg2 Qxb2 38.Qd7?

White should have taken either the pawn on a5 or the one on b7. Alekseenko probably wanted to create some threats against the enemy King while Black is in time trouble. He probably reasoned that he can capture the pawn anytime.

38…b6 39.Nd6 Qf6 40.Qe8+ Kg7 41.Qd7 Kg8 42.Qe8+ Nf8 43.Qc6 Qd8 44.Nc4?!

But this is wrong — 44.Nb7! wins back the pawn. After 44…Qd4 45.Nd6! Qc5 (to prevent Qc7) (45…Kg7 46.Qc7 Qd5+ 47.Kh2 Qxb3 48.Ne8+ Kh7 49.Qe5 Ne6 50.Nf6+ Kh8 (50…Kh6 has the same effect: 51.Ng8+ Kh7 52.Nf6+) 51.Nxh5+ Kg8 52.Nf6+ Kg7 (52…Kf8 53.Qb8+ Ke7 54.Ng8+ Kd7 55.Nf6+ Ke7 draw) ) 46.Qxc5 bxc5 47.Nc4 the material balance is restored.

44…Ne6 45.Nxb6 Nd4 46.Qc5 Nxb3 47.Qb5 Nd2! 48.Qxa5 Qd3 49.Qa1 Qe4+ 50.Kg1 Nf3+ 51.Kf1 Nxh4 52.Qa8+

[52.gxh4?? Qh1+ wins the white queen]

52…Qxa8 53.Nxa8 Nf3

Black has three pawns against White’s two all on the same side of the board. This should probably be a draw but Giri shows that it can be tricky.

54.Kg2 Ne5 55.f4? <D>


White should have maintained his pawns where they were. Now Black has the possibility of an additional pawn lever with ….h5–h4.

55…Ng4 56.Nb6 Kf8 57.Nd5 Ke8 58.Nc3 Ke7 59.Ne4 Ne3+ 60.Kf3 Nc4 61.Ng5 Kf6 62.Ne4+ Kf5 63.Nf2 Nd2+ 64.Ke3 Nf1+ 65.Kf3 Nh2+ 66.Kg2 Ng4 67.Nh3

You remember the old rule that if you are ahead exchange pieces, and if you are behind exchange pawns, right? If White exchanges pieces then 67.Nxg4?? hxg4 68.Kf2 g5 liquidates quickly to a won pawn endgame.

67…f6 68.Kf3 Ke6 69.Ke4 Kd6 70.Ng1 Kc5 71.Kd3 Nh6 72.Ke3 Nf5+ 73.Kf3 Kc4 74.Nh3 Nd4+ 75.Ke3 Nf5+ 76.Kf3 Kd4 77.Nf2 Nd6 78.Nh3?

It is starting to get tricky. Correct is 78.Nd1! Kd3 79.Nb2+ (White shouldn’t let his knight be stuck to the corner of the board. For example if 79.Nf2+? Kd2 80.Nh3 Nb5! 81.Kf2 Nd4 82.Ng1 Kd1 83.Nh3 Nf5 84.Ng1 Kd2 85.Nh3 Kd3 86.Kf3 Nd4+ 87.Kf2 Ke4 88.Ng1 h4! wins. This “pawn lever” is what I was talking about back in move 55.) 79…Kd2 80.Na4 Ke1 81.Kg2 Ke2 82.Nc5 this knight can now attack the black pawns from behind.


[78…Kd3! wins with the maneuver I just showed last move 79.Kf2 Nf5 80.Kf3 Kd2 81.Kf2 Nd4 etc]

79.Ng1 Kd3 80.Kg2 Nd2 81.Kf2 Ke4 82.Ne2 Nb1 83.Ng1 h4 84.Nh3 Kf5 85.gxh4 Kg4 86.f5!

The correct drawing maneuver, doubling Black’s pawns.


[86…Kxf5 87.Kf3 Nd2+ 88.Kg3 draws with White maintaining a strongpoint on f4]

87.Ke3 Nc3

[87…Kxh3? 88.Kf4 white wins both black pawns and it is a dead draw]

88.Nf2+ Kg3 89.Nd3?

Loses. 89.Nh1+ Kg4 (89…Kxh4 Malcolm 90.Kf4 Nd5+ 91.Kxf5 Kh5 92.Ke6 draw) 90.Nf2+ Kh5 91.Kf4 and, once again, a draw.

89…Nd5+ 90.Kd4 Nf4

Black manages to save his pawns and now it is a win for him.


[91.Nxf4 Kxf4 92.Kd3 Kg4 93.Ke3 (93.h5 Kxh5 94.Ke3 Kg4) 93…f4+ 94.Ke4 f3 95.Ke3 Kg3]

91…Kxh4 92.Ke3 Kg3 93.Nb3 Ne6 94.Nd2 f4+ 95.Ke2 Ng5 96.Kf1 f3 97.Kg1 f2+ 98.Kf1 f5 0–1

On the 78th move White erred by playing Nh3 instead of Nd1, and on the 89th move his mistake was playing Nd3 instead of Nh1. So, on the first instance the knight should have gone to the center instead of the corner, and on the second instance it should have gone to the corner rather than the center. Yup, those knight endgames are really difficult to play.


Bobby Ang is a founding member of the National Chess Federation of the Philippines (NCFP) and its first Executive Director. A Certified Public Accountant (CPA), he taught accounting in the University of Santo Tomas (UST) for 25 years and is currently Chief Audit Executive of the Equicom Group of Companies.