U2000
QGD Semi-Tarrasch
1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.Nf3 c5 5.cxd5 cxd4!?
The mainline in ChessBase's 2022 Mega database runs 5...Nxd5 6.e4 Nxc3 7.bxc3 cxd4 8.cxd4 Bb4+ 9.Bd2 Bxd2+ 10.Qxd2 0-0 11.Bc4 with a slight edge for White, according to Stockfish15 and Komodo13.02.
6.Qa4+!?
Magnus Carlsen has played 6.Qxd4, which is the main move, and also 6.Nxd4, but not, as far as I can discover, the text.
6...Bd7 7.Qxd4 exd5 8.Bg5
One of the points of 6.Qa4+!?, as opposed to 6.Qxd4, is White can win a pawn with 8.Nxd5!? However Black gets strong counterplay with 8...Qa5+ 9.Nc3 Nc6.
8...Nc6
Also possible is 8...Be7, when 9.Bxf6 Bxf6 10.Qxd5 0-0 gives Black reasonable compensation for a pawn, but 8...Be6?! 9.Bxf6 Qxf6 10.Qxf6 gxf6 11.0-0-0 is slightly better for White, according to the engines.
9.Qd2
9.Bxf6 Nxd4 10.Bxd8 Nc2+? 11.Kd2 Nxa1 12.Bg5 gave White a large advantage in Salo Landau - Rudolf Spielmann, Amsterdam 1936 (1-0, 27 moves). However Black has a major improvement with 10...Nxf3+ 11.exf3 Rxd8, and if 12.Nxd5 then 12...Bc6, when White's extra pawn is largely meaningless.
9...Be6 10.e3
Not 10.Bxf6?! Qxf6 11.Nxd5?? Bxd5 12.Qxd5 Bb4+ 13.Nd2 Qxb2 with ...Rd8 to come.
10...Be7 11.Be2
 |
| How should Black proceed? |
*****
*****
*****
*****
11...Ne4!?
This is a typical move in this type of IQP position, but usually comes after both sides have castled, eg Artur Darzineks - Mikhail Tal, Latvian Championship (Riga) 1954, went 11...0-0 12.0-0 Ne4 13.Bxe7 Qxe7 14.Qe1!? f5!? 15.Rd1 Rad8 16.Nd4 Bc8 17.Bb5 with at least a slight advantage for White, according to the engines (but 0-1, 49 moves).
12.Bxe7
Also equal, according to the engines, is 12.Nxe4 dxe4 13.Bxe7 Qxe7 14.Nd4.
12...Qxe7 13.Qd1!?
The engines marginally prefer keeping contact with the queen's knight by 13.Qd3 or 13.Qc2.
13...Rd8!?
The wrong rook? It may be that the queen's rook is better placed on c8, especially after 13...Nxc3 14.bxc3, leaving d8 for the king's rook.
14.0-0 0-0 15.Qe1?! Nb4?!
Threatening a fork on c2, but that is easily stopped.
16.Nd4 Nxc3 17.bxc3
The engines slightly prefer 17.Qxc3, which would justify White's 15th move.
17...Nc6 18.Qd2 Rc8 19.Rfc1
The engines reckon White is slightly better after 19.Nxe6 fxe6 and may be slightly better after 19.Nxe6 Qxe6!?
19...Rfd8 20.Qb2 Ne5 21.Rab1 Rd7 22.Qb4 Qf6 23.Rb2 Bg4 24.Bxg4 Nxg4 25.Nf3 Rc4 26.Qb3 Rdc7 27.Qb5
Threatening a back-rank mate, as well as the d pawn, but the engines prefer 27.Rd2!?, eg 27...Rxc3 28.Rxc3 Rxc3 29.Qxd5 Rc1+ 30.Rd1 Rxd1+ 31.Qxd1, reckoning Black's farside pawn-majority is of no significance.
27...Qc6
Black has a slight edge after 27...Qd8!?, according to the engines.
28.h3 Nf6 29.Qxc6 R4xc6 30.Nd4 Rb6?!
Probably better is 30...Ra6.
31.Rxb6 axb6 32.Rb1 Rxc3?!
Probably another inaccuracy. The engines reckon White's edge is smaller after 32...Ne4 33.Rxb6 Nxc3.
33.Rxb6 Rc7 34.a4
 |
| In this position SB offered a draw |
*****
*****
*****
*****
34...Kf8!?
34...Kf8!?
The pawn-structure means White must be better, but the engines disagree as to how much. Komodo13.02 reckons for quite some time White has the upper hand, while Stockfish15 gives only a slight edge, but with more time both engines become less enthusiastic about White's chances.
35.Kf1 Ne4 36.a5 g6 37.a6?!
Too intent on a draw. After ...
37...bxa6 38.Rxa6
... the position is completely equal, according to the engines. The problem with White's play is Black has more chance of conjuring a win from an equal position than from the previous slightly inferior one, albeit the chance is a small one.
38...h5 39.Ra5 Rc5 40.Rxc5 Nxc5 41.Ke2 Ke7 42.Nf3 f6 43.Nh4 Kf7 44.Nf3 g5 45.Nd4 Kg6 46.Nc6 Kf5 47.Nd4+ Ke5 48.Nc6+ Kd6 49.Nd4 h4!?
Gaining space, and hoping to manoeuvre White into some sort of zugzwang, must be Black's best hope, but Stockfish15 at first does not like the text.
50.Nf5+
Stockfish15 initially claims a slight White edge after 50.f4!?, but fairly soon reverts to rating the position as completely equal.
50...Ke5 51.Nd4 f5 52.Nc6+ Kf6 53.Nb4
Now it is Komodo13.02's turn to claim a slight edge for White, after 53.f4!?, but it also quickly reverts to equality.
53...Ke6 54.Nc6 g4 55.Nd4+ Kf6 56.Nc6 Ne4 57.Nb4 Ke6 58.Nc6 gxh3?!
 |
| Guess what happened next |
*****
*****
*****
*****
0-1
SB resigned, presumably thinking White is lost after my intended 59.gxh3 Ng5? However 60.f4 wins for White, eg 60...Nxh3?! 61.Nd4+ Kf6 62.Kf3 wins the Black knight. Slightly better is 60...Ne4, but 61.Nd4+ Kf6 62.Nf3 wins the h4 pawn, and with it the game, according to the engines. Instead of 59...Ng5? Black should play 59...Kd6 or 59...f4!?, according to the engines, but any advantage lies with White.
No comments:
Post a Comment