FACED a Spaniard.
Spanton (1913) - Manuel Llopis de Aysa (2000)
QGD Exchange
1.c4 e6 2.Nc3 d5 3.d4 Nf6 4.cxd5 exd5 5.Bg5 Be6 6.e3 0-0 7.Bd3 h6!?
There was a time, not so long ago, when ...h6 had a dubious reputation in the Exchange Variation of the Queen's Gambit Declined. The problems were thought to be twofold: i) it provides White a target for a g-pawn push, or a possible piece sacrifice; ii) it takes away a square a black rook might want to use, after joining the third rank at d6 or e6, for putting pressure on White's kingside.
So what has changed? Nothing really - those criticisms are still valid, BUT 7...h6!? comes with engine approval, eg it is the top choice of Stockfish17.1 and Dragon1. Therefore, the argument goes, the benefits of ...h6 must outweigh its drawbacks.
 |
| The position after 7.Bd3 occurs 11,650 times in ChessBase's 2026 Mega database |
The continuation 7...c6 occurs 3,916 times (34%), 7...Nbd7 occurs 2,939 times (25%) and 7...h6 occurs 2,251 times (19%).However, make the same search, but from 2023 onwards only, and the results are rather different. There is still quite a substantial number of games, 2,227, and 7...c6 is still most popular, but its 695 appearances represent a more modest 31%. In second place, with 630 games (28%). is 7...h6, while 7...Nbd7's 265 games is 12%.
In other words, in recent praxis, 7...c6 has slipped three percentage points, and 7...Nbd7 has dropped 13 percentage points, but 7...h6 has gained nine percentage points.
This is surely an example of how engines are affecting opening choices, even in non-tactical positional lines.
Back to the game.
8.Bh4 Re8 9.Qc2
 |
| Guess which move is Black's most popular continuation (hint: it is also liked by the engines) |
*****
*****
*****
*****
*****
Well done if you plumped for 9...c5!? - you either know your (modern) theory, or you have a natural affinity for the Exchange Variation of the QGD.
I suspect most club players would give 9...c5!? little or no consideration. After all, what happens if White captures on c5, saddling Black with an isolated queen's pawn? Well, after 10.dxc5, the engines continue 10...d4!?, and if 11.exd4 (the most popular move in Mega26) Qxd4 (attacking the h4 bishop as well as the c pawn) 12.Nf3, then 12...Bd8+!?, when the engines reckon Black is at least equal.
The engines' alternative to 10.dxc5 is 10.Nf3 or 10.Nge2, but then 10...c4!? comes with tempo. For those who do not like the ...c4 move in such positions, 10...Nc6 looks reasonable.
9...c6
This, to me, is the most natural continuation, or, at least, the one I regard as being most 'normal' in this opening.
 |
| The position occurs 2,716 times in Mega26 |
*****
*****
*****
*****
*****
10.Nf3 Ne4 11.Bxe7 Qxe7 12.Bxe4!?
This is very much second in popularity to 12.0-0, but it has been played by Magnus Carlsen and other strong grandmasters.
12...dxe4 13.Nd2 Bf5
ML offered a draw.
 |
| Who, if anyone, stands better? |
*****
*****
*****
*****
*****
Equal chances, according to the engines.
14.0-0 Bg6 15.a3 Nd7 16.b4 Nb6 17.Na4 Nxa4
The engines reckon Black has at least the better part of equality after 17..Nd5.
18.Qxa4 a6 19.Qc2 Rad8 20.Nc4?!
The engines strongly dislike this, suggesting 20.Qc5 or 20.Rfc1 (one of the points of 20.Rfc1 is that it frees the f1 square for the knight).
20...Qg5
Best, according to the engines, is 20...Rd5, eg 21.Ne5 Bh5 22.h3 f6 23.Nc4 Rg5, when Black has promising kingside play.
21.Kh1 Bf5 22.Ne5 f6 23.Nc4 Be6 24.Rae1 Bxc4 25.Qxc4
 |
| How would you assess this heavy-piece middlegame? |
*****
*****
*****
*****
*****
Completely equal, according to the engines.
25...Qd5 26.Rc1 Qxc4 27.Rxc4 Rd5 28.Rfc1 Kf7 29.Rc5 Ke6 30.Kg1 Rb8 31.Kf1 Rxc5 32.Rxc5
 |
| How would you assess this rook-and-pawn ending? |
*****
*****
*****
*****
*****
Completely equal, according to the engines.
32...Kd6 33.Ke2 b6 34.Rc2 Rc8 35.f3 exf3+!? 36.gxf3 Re8 37.Kd3 Re7 38.e4 Re8?!
Black should not do nothing - the engines suggest 38...g5, maintaining equality.
39.Rg2
The engines suggest 39.f4 or 39.a4.
39...Re7?!
White's rook should almost certainly not be allowed to occupy g6, so again ...g5 is called for.
40.Rg6
 |
| How should Black proceed? |
*****
*****
*****
*****
*****
40...Kc7?!
The engines suggest 40...Ke6, or play on the queenside with 40...b5 or 40...a5, or 40...h5!?, but in each case with a large advantage to White.
41.f4 Rd7 42.Kc4
The engines suggest clamping down on the kingside with 42.h4, intending h5, or advancing on the queenside with 42.a4. If, for example, 42.h4 is met by 42...c5, then 43.bxc5 bxc5 44.d5.
42...Kd6!?
This is the best try, according to the engines, but they reckon White is winning.
 |
| How should White proceed? |
*****
*****
*****
*****
*****
43.e5+?!
Winning a pawn, but letting the black rook become active. Instead, White should increase the pressure, eg 43.f5!? Kc7!? 44.a4 Rf7 45.a5 h5!? 46.h4 Kb7 47.Rg1 Re7 48.Re1 Rf7 49.Kd3 Kc7 50.Rc1 Kb7 51.c5 cxd5 52.exd5 bxa5 53.bxa5 Re7 54.Kd4, when White's advantage is worth more than a rook, according to the engines - Black's three kingside pawns are held up by two White pawns, leaving d5 a monster.
43...Ke6 44.exf6
The engines prefer 44.h4!?, and if 44...Kf6, then 45.Rg1!?, the point being 45...Kxf4? loses to 46.e6 Re7 47.Rf1+ Ke4 48.Re1+ Kf5 49.d5 etc.
44...gxf6 45.Rxh6 Rg7 46.Rh5 Rg1 47.f5+ Kd6 48.Rh6 Rc1+ 49.Kb3 Kd5 50.Rxf6 Kxd4 51.Kb2
 |
| Black has three plausible rook moves - which should he play? |
*****
*****
*****
*****
*****
51...Rf1?
The engines agree this loses; they also agree 51...Rc4! draws. Stockfish17.1 reckons 51...Rh1!? also draws, but Dragon1 reckons it loses. Even after the 51...Rh1!? line is continued for several moves, eg 52.Rxc6 Rxh2+ 53.Kb3 b5 54.Rxa6 Rh3+ 55.Kc2 Rh2+ 56.Kd1 Rf2 57.f6 Rf5 58.Rd6+ Ke4, both engines persist in their different evaluations. One reason 51...Rc4! draws is that 52.h4 can be met by 52...Ke5.
52.Rxc6 Rf2+
 |
| How should White proceed? |
*****
*****
*****
*****
*****
53.Rc2?
Both 53.Kc1 and 53.Kb3 apparently win, but the text does not. One winning line runs 53.Kc1 Rxf5 54.Rxb6 Kc3 55.Rc6+ Kb3 56.Rxa5, when White is three pawns up, and the Syzygy endgame tablebase confirms Black is lost.
53...Rxf5 54.h4 Rh5 55.Kb3!?
 |
| Black to play and draw |
*****
*****
*****
*****
*****
55...b5?
The one plausible move that loses, according to the engines. Of course not 55...Rxh4?? 56.Rc4+ etc, but both 55...Kd5 and 55...a5!? leave White with at best a slight edge, according to the engines, although it takes Dragon1 a lot longer to realise this than Stockfish17.1. After 55...Kd5, White can protect the h pawn with 56.Rh2 or 56.Rc4, but there seems no way to make progress, while 56.Rd2+ Kc6 57.Rd4 does not change anything.
 |
| After 55...b5? it is White to play and win |
*****
*****
*****
*****
*****
56.Rc6?
Thanks to 55...b5?, White's protecting the h pawn with 56.Rh2!? now wins, eg 56...Ke4 57.Kc3 Ke3 58.Rh3+ Ke4 59.Kc2!? Kd4 60.Ke2 Ke4 61.Kf2 Kf4 62.Rh1 Kg4 63.Ke4 Kf5 64.Kd4!? Kf6 65.Rh3!? (White's ability to temporise with the rook is what makes the endgame winning) Kf5 66.Re3!? Rxh4+ 67.Kc5 Rh6 68.Kd3 with Rd6 and Rxa6 to come. There are lots of alternatives along the way, but the engines are sure White always has a win.
56...Rxh4 57.Rxa6
The main difference between the position in the game and the one reached in the previous note is that in the game White's king is confined to the back three ranks.
57...Rh3+ 58.Kc2 Rh2+ 59.Kd1 Kd3 60.Ke1 Kc4 61.Rg6 Kb3 62.Rg3+ Ka4 63.Rd1 Ra2 64.Kc1 Rxa3 65.Rxa3+ Kxa3 66.Kc2 Kxb4 ½–½
 |
| With the b pawn in its half of the board, this is only a win for the attacking side if the king can get to one of three squares two ranks ahead, ie in this case a3, b3 or c3, and since White easily stops that with the obvious 67.Kb2, the ending is drawn |
.jpg)