QGD Tarrasch
1.d4 d5 2.c4 e6 3.Nc3 c5
Two years ago in the same tournament our game went 3...Nf6 4.Bg5 c5!? 5.cxd5 cxd4 6.Qxd4 Be7 7.e3!? exd5 8.Nf3 0-0!?, with an equal position, according to Stockfish17 and Dragon1. Although I won that game in 22 moves, I did not want to find out if VL had prepared an improvement.
4.cxd5 exd5 5.e3!?
This is the second choice of the engines, and has been played by grandmasters, but is much less successful than the mainline 5.Nf3.
5...Nf6 6.g3!?
This gives a strange impression, but the engines are OK with it, and there are 39 examples of the resulting position in ChessBase's 2025 Mega database.
 |
| How should Black respond? |
*****
*****
*****
*****
6...Nc6
The engines also like 6...cxd4 and 6...Bg4.
7.Bg2 Bg4
The engines suggest 7...cxd4 or 7...h5!?
8.Qb3?!
This is strongly disliked by the engines. They reckon White is slightly better after 8.Nge2.
8...cxd4
The only game in Mega25 to reach the position saw 8...c4!?, which is the top choice of the engines.
9.Nxd5
Possible is 9.exd4!?, the point being 9...Nxd4 10.Qa4+ Nc6 11.Bg5 is sharp and may give enough compensation for a pawn, although Dragon1 is less sure than Stockfish17. The engines fluctuate between the two moves, and also 9.h3!?
9...Nxd5 10.Bxd5 Bb4+
 |
| How should White proceed? |
*****
*****
*****
*****
11.Kf1!?
Not 11.Bd2? Bxd2+ 12.Kxd2 Qa5+! (this is even stronger than 12...dxe3+), after which the white king has no good square as both 13.Kc2 and Kd3 can be met by 13...Nb4+, eg 13.Kd3 Nb4+ 14.Ke4 f5+ 15.Kxd4 (15.Ke5 Nc6+ 16.Kf4 Qc7+ 17.Kg5 h6+ 18.Kg6 Ne5#) Nxd5 16.Qxd5 Rd8 etc. That leaves 13.Kc1, but 13...Qe1+ 14.Kc2 Nb4+ wins the white queen.
11...Qe7 12.a3!?
This is Stockfish17's top choice, although it comes to also like 12.Kg2. Dragon1 suggests 12.e4, marginally ahead of 12.Kg2.
12...Bc5 13.Qa4?
White is fine, according to the engines, after 13.h3, with 14.e4 possible whether the bishop retreats to h5, f5 or e6.
 |
| How should Black proceed? |
*****
*****
*****
*****
13...0-0!?
The engines also like 13...Rd8!? as there is no need to defend c6.
14.e4
After 14.Bxc6 bxc6 15.Qxc6? Rac8 White has very little chance of surviving Black's attack.
14...Ne5 15.Qb3 d3!? 16.h3 Bh5 17.Bf4 Qf6?!
Developing the queen's rook makes sense, and even the speculative 17...Bxf2!? is probably good for Black after 18.Kxf2 Rac8. However the text, which at first glance looks good as it pins White's dark-square bishop, gives White the chance to beat off Black's attack.
 |
| What should White play? |
*****
*****
*****
*****
18.Qc3!
The engines agree this is best.
18...Be2+?
Better is 18...Nd7, although that allows 19.Qxd3!? as 19...g5 can be simply met by 20.Be3. Dragon1 suggests 19...Qxb2, but 20.Ra2 (necessary to protect f2) leaves Black with at best a slight edge, according to the engines.
19.Nxe2 dxe2+
 |
| Can you see why 18...Be2+? was wrong? |
*****
*****
*****
*****
20.Kxe2
White has a winning advantage, according to the engines, after the remarkable 20.Kg2! Black can save both hanging pieces with 20...Nd7, but 21.Qxf6 Nxf6 22.Bxb7 Rae8 23.b4! Bd4 24.Ra2 will see White emerge material ahead and with the bishop-pair, eg 24...Nxe4 25.Rxe2 g5! 26.Bc1. Slightly better may be 20...Bd6, but it is still good for White, eg 21.Qe3 Ng6 22.Bxd6 Qxd6 23.Qxe2.
20...Qa6+ 21.Kd2 Nd7
The engines give 21...Bd4!, the point being 22.Qxd4?? loses to 22...Nf3+. White has to play 22.Qb3, when 22...g5! 23.Bxe5 Bxe5 leaves White a pawn up, but with an exposed king in a position with opposite-coloured bishops. Dragon1 reckons White has the upper hand, but Stockfish17 gives White only a slight edge, and it would be a difficult middlegame for White to negotiate.
22.Be3
The engines give 22.Rac1!?, the idea being to meet 22...Bxf2 with 23.Rhf1, when White has given back the pawn but is ahead on development and has threats, along with the bishop-pair. Instead of immediately grabbing the f pawn, the engines give 22...Rad8!? 23.Kc2, and now 23...Bxf2, but still prefer White after 24.Kb1!?
22...Rac8 23.b4?!
The engines still like Rac1.
23...Bxe3 24.Qxe3
 |
| Black has full compensation for a pawn, according to the engines |
*****
*****
*****
*****
24...Nf6 25.Rhc1?!
The engines prefer 25.Qd3!?, meeting both 25...Qd6 and 25...Qb6 with 26.Ke2.
25...Nxd5
25...Rcd8!? is slightly stronger, according to the engines.
26.exd5 Rcd8 27.Rc5?
Equal is 27.Qd3!? Qxd3+ 28.Kxd3 Rxd5, as is 27...Qh6+ 28.Kc3 Qxh3 29.Kb3, according to the engines.
27...b6 28.Rc3 Rxd5+ 29.Rd3 Rfd8?
This looks natural, but throws away the win, which was to be had, it seems, with more than one continuation, including 29...Rxd3+ and possibly 29...Qb7, but most convincing is 29...Rdd8!, to which there appears no satisfactory answer. Perhaps the best try is 30.Kc2, but the engines give 30...Qc4+ 31.Rc3 Qb5! 32.Rd1 Qa4+ 33.Rb3 Rxd1 34.Kxd1 Re8 35.Qd3 Qc6 36.Rc3 Qh1+ 37.Kc2 Qxh3, after which Black is a pawn up and still has much the safer king. There is a long way to go to convert the extra pawn, and there are plenty of alternative moves before reaching this point, but the engines are convinced Black is winning.
30.Rxd5 Rxd5+ 31.Ke1 Qb5 32.Rd1 Rxd1+ 33.Kxd1
VL offered a draw in my time.
 |
| Dead equal? |
*****
*****
*****
*****
Yes, according to the engines, one point being Black sooner or later has to spend a tempo preventing back-rank mate.
33...g6 34.Kd2 Qd7+ 35.Ke2 Qxh3
Winning a pawn, but only temporarily.
36.Qe8+ Kg7 37.Qe5+ Kf8!? 38.Qb8+ Kg7 39.Qxa7
After 39.Qe5+ the only way to keep the game going is 39...Kh6 40.Qf4+ Kh5!?, although 41.Qxf7 gives White a tiny edge, according to Dragon1 (Stockfish17 reckons the ending is still equal).
The text means White has the smaller of the two pawn-majorities, and has the majority that is on the farside from the kings, but Black still has the draw in hand.
39...Qe6+ 40.Kf1 h5 41.Qa6 Qc6 42.Qd3!?
Centralising the queen. It allows ...
42...b5!?
... which means one pawn is now holding up two on the queenside, but that does not change the engines' assessment of complete equality.
43.Qd4+ f6 44.Qe3 g5 45.Qd3 h4!? 46.gxh4 gxh4
 |
| A passed pawn is normally a major asset in a queen-and-pawn ending, even though here it has come at the cost of accepting three isolanis, but the engines steadfastly maintain their verdict of completely equal |
*****
*****
*****
*****
47.Qf5 Qc1+ 48.Ke2
Also fine is 48.Kg2!? as following 48...Qg5+ 49.Qxg5 fxg5 White has 50.a4 (50.f4!? may also hold the draw), after which both sides queen a pawn.
48...Qc4+ 49.Ke1 Qc3+ 50.Kf1 Qa1+!? 51.Kg2 Qe5!?
Continuing to take risks in the hope of somehow conjuring up a win.
 |
| How should White proceed? |
*****
*****
*****
*****
52.Qd7+
The Syzygy endgame tablebase shows 52.Qxe5 fxe5 53.Kh3 Kf6 54.Kxh4 is a draw, provided Black finds 54...Kf6 55.Kg3 e4!
Another try is 52.Qg4+ Qg5 53.Kh3 Kg6 54.Qxh4, but it seems Black holds with, among other moves, 54...Qd5 and 54...f5, but 54...Qxh4+?? 55.Kxh4 is a winning pawn-ending for White.
52...Kh6 53.Kh3 f5 54.Qd2+
Not 54.Kxh4?? Qh2#.
54...f4!?
 |
| White to play and gain a slight edge (Stockfish17) or even the upper hand (Dragon1) |
*****
*****
*****
*****
55.Kxh4??
After 55.Qc2! White is the one pushing for a win, although it seems 55...Qg5 56.Qe4 Qf6 holds, and other lines too, according to the engines.
The text was presumably played in the light of the reply ....Qg5+ only drawing, thanks to the f5 pawn being pinned after the black queen leaves g5. However that misses:
55...Qh5#
No comments:
Post a Comment