ACO 65+ Game Seven

Martin Schaefer (1921) - Spanton (1804)

QGD Semi-Tarrasch

1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.Nf3 d5 5.cxd5 cxd4!?

This seems to have become popular round the same time strong engines started evaluating the move as roughly on a par with 5...Nxd5 and 5...exd5.

6.Qxd4 exd5

How should White proceed?

*****

*****

*****

*****

*****

7.Bg5

This looks natural and is not a bad move, but the most popular continuation in Chessbase's 2023 Mega database is 7.e4!?, after which the mainline in Mega23 runs 7...Nc6 8.Bb5 dxe4 9.Qxd8+ Kxd8 10.Ng5 Be6!? with a slight edge for White, according to Stockfish15.1 and Komodo14.1.

7...Be7 8.e3 Nc6 9.Bb5 0-0 10.Qd3!?

Most popular in Mega23 is 10.Qa4, while Anatoly Karpov in a 1998 simul preferred 10.Bxc6.

10...Be6 11.0-0 Qb6

Andrey Zhigalko (2586) - Artyom Manukyan (2046), Asrian Memorial (Yerevan, Armenia) 2015, went 11...h6 12.Bh4 Nb4 13.Qb1 Qb6 14.Rd1 Rad8!? 15.Nd4 with a slight edge for White, according to the engines, but the game, despite the huge rating gulf, was agreed drawn after 43 moves.

12.a3 Rfd8!?

The engines strongly dislike this, preferring 12...a6 or 12...h6.

13.Rfd1

White has the upper hand after 13.Na4!?, according to the engines.

13...h6 14.Bh4 Rac8 15.h3 Ne4

This resource typically arises for Black in Semi-Tarrasch positions with an IQP.

What should White play?

*****

*****

*****

*****

*****

16.Na4?!

Best, according to the engines, is 16.Bxc6 Bxh4 17.Nxh4, and after both 17...bxc6 and 18...Nxc3 they reckon the game is equal. Also reasonable is 16.Bxe7 Nxe7 17.Nd4, although the engines reckon 17...Nxc3 18.bxc3 gives Back the better side of equality.

16...Qc7 17.Bxe7 Nxe7!

Seeking to take advantage of the awkward placing of the white minor pieces on the queenside

*****

*****

*****

*****

*****

18.Qf1!? a6 19.Bd3 Qa5

This seems better than 19...b5, one point being 20.Nc3 Nxc3 21.bxc3 Qxc3 22.a4!? is equal, according to the engines.

20.b3 b5 21.Nb2 Qc3 22.Qe2 Qxb3

The engines prefer building up with 22...Nc6!?

23.Nd4 Qc3

How can White equalise, or at least come close to doing so?

*****

*****

*****

*****

*****

24.Rdc1

The engines reckon White should play 24.a4!?, eg 24...Qc5 25.Nxe6 fxe6 26.Qg4 with good counterplay.

24...Qa5!?

I rejected 24...Qxc1+ 25.Rxc1 Rxc1+ because I feared the imbalance would give White practical chances. However, after ...

25.Nxe6 fxe6 26.Bxe4 fxe4 27.Qg4 Rxc1+ 28.Rxc1 Qxa3 29.Qxe6+

... White has counterplay.

29...Kf8

The engines reckon safety-first with 29...Kh8!? is better.

30.Rc2 Qd6

Letting the e4 pawn go but keeping a 2-0 queenside majority.

31.Qxe4 Qd5

The engines reckon Black should be getting on with it on the queenside, claiming Black has the upper hand after 31...a5.

32.Qf4+ Qf5 33.Qxf5 Nxf5 34.g4 Ne7 35.Kf1 Kf7

Again the engines prefer ...a5, claiming the text leaves the game equal.

36.Ke2 Ke6 37.Nd3 Kd6 38.Nf4 Rc8 39.Rd2+ Kc6 40.Ne6 g6 41.f4 Re8

Komodo14.1 reckons White now has a slight edge, although Stockfish15.1 still calls the game equal

*****

*****

*****

*****

*****

42.e4 Nc8?!

This is probably too slow. The engines much prefer 42...a5 or 42...b4.

43.f5 Nd6

White to play and get a winning advantage

*****

*****

*****

*****

*****

44.Ra2?

Good seems to be 44.Kf3, but best almost certainly is 44.e5!, eg 44...Nc4 45.Rd8 Re7 46.fxg6! Rxe6 47.g7 Rxe5+ 48.Kd3 Rg5 49.g8=Q Rxg8 50.Rxg8, after which the engines reckon Black cannot cope with White's kingside pawns and exchange advantage.

44...Nxe4 45.Kf3

After 45.Rxa6+ Kb7 46.Ra3 gxf5 47.gxf5 Ng5 the position is equal, according to the engines, in a similar way as to happens in the game.
45...gxf5 46.gxf5 Ng5+ 47.Nxg5 hxg5 48.Rxa6+ Kd5 49.Kg4 Ke5!?

Stockfish15.1 is happy with this, but Komodo14.1 claims White now has the upper hand. The latter engine prefers 49...b4.

50.Rb6 Rg8 51.Rxb5+ Kf6

The Syzygy endgame tablebase shows this position is drawn, but it clearly is much easier for White to play

*****

*****

*****

*****

*****

52.Rb6+ Kf7

Other moves lose.

53.Rh6

Black to play and draw

*****

*****

*****

*****

*****

53...Rg7?

Six moves draw: Ke7, Ra8, Rb8, Rc8, Rd8 and Re8.

White to play and win

*****

*****

*****

*****

*****

54.Kh5

In the postmortem grandmaster Spyridon Skembris did not see anything wrong with this move. But later, when he asked me to show the game as part of his afternoon lecture, I was able to tell him White has a winning move in the diagrammed position, and after a little thought he came up with the correct 54.Rh8! The point is the black rook cannot sensibly move, and if  54...Kf6 White has 55.Ra8 (other moves also win), eg 55...Rh7 56.Ra6+ Ke5 57.Kxg5 Rxh3 58.f6 Rg3+ 59.Kh6! (only-move) Rf3 60.Kg7, after which White reaches a winning Lucena position. Alternatively 54...Ke7 55.Ra8 Kd6 leaves White with a host of winning moves, including the simple 56.f6.

54...Rg8

An only-move, but a fairly obvious one.

55.Rb6 Rh8+ 56.Kg4

If 56.Kxg5 Black captures on h3 and will reach a drawing Philidor position, eg 56...Rxh3 57.Rb7+ Kf8 58.f6 Rg3+ 59.Kf5 Ra3 60.Ke6 Ra6+ etc.

56...Rh4+ 57.Kg3 Rf4 ½–½