Lockdown Chess

THERE are few, if any, signs of a return of over-the-board chess in Britain.
Perhaps hopes will rise if the government switches from two-metre social-distancing to the World Health Organisation-recommended one-metre.
Meanwhile many players are making do with playing over the internet, often at fast time limits although there have been some attempts at more-normal time controls.
The former does not interest me as, apart from the fact I have BT disconnection-prone home broadband, I play over-the-board chess too quickly and fear playing at even faster time limits would encourage me to play even faster OTB.
And more-normal time limits online are, as many people expected, plagued by cheating and suspicions of cheating.
However I was fortunate to be invited to play in the 6th British Webserver Championship, which is being run by the Fide-approved International Correspondence Chess Federation.
ICCF chess is virtually immune from cheating, for the simple reason that engines are allowed.
People who do not play correspondence chess often ask: what's the point if engines are involved?
Well the point is that human+engine is stronger than engine alone, or at least should be, which is why every game of correspondence chess does not end in a draw.
I am playing on board thee for Lean Thinkers in division two, which consists of seven teams, and is run as a single round-robin, ie you play each opponent once rather than once with white and once with black.
My first two games to finish were draws. Here is the third one.
Mel Suffield (2060) - Spanton (2256)
Maróczy Bind
1.e4 c5 2.Nf3 g6
Engines are notoriously weak at playing openings. They are so weak that I believe I am right in saying there has never been a match between a top computer and a strong player in which the engine was allowed to pick its own moves from the start. Instead an engine is supplied with an opening book, usually compiled by a grandmaster in collaboration with computer programmers, that it is forced to follow until it is trusted to 'think' for itself.
At correspondence chess, however, players will usually pick their own opening moves while consulting books and databases. One problem with this approach is that the game may quickly reach a position an engine is not comfortable with.
3.c4 Nc6 4.d4 cxd4 5.Nxd4
We have reached the basic starting position of the Maróczy Bind. Black's decision to allow the bind used to be regarded as dubious strategy, or as downright bad. For instance Reuben Fine in The Ideas Behind The Chess Openings (Bell 1948) wrote of the Sicilian Defence: "Black must never allow White to play P-QB4 in the opening because he then has no counterplay on the QB file and is thereby doomed to passivity." Current thinking is that the Maróczy Bind does indeed give White an edge, but that the edge is within the drawing margin, so allowing c4 is viable. Engines do not like it for Black, however, so this arguably is a case of where human+engine is weaker than engine alone.
5...Bg7 6.Be3 Nf6 7.Nc3 d6
More popular in ChessBase's 2020 Mega database are 7...0-0 and 7...Ng4!?
8.Be2 0-0 9.0-0 Bd7
This move appears more than 10,000 times in Mega20, but a major alternative is 9...Nxd4, transposing into a line of the Gurgenidze System, although in that system Black usually captures on d4 at an earlier point.
10.Qd2
Easily White's most-popular continuation, but avoiding an exchange of knights with Nc2 has been recommended in repertoire books, although normally Nc2 is recommended earlier in the move-order to avoid ...Nxd4.
10...Nxd4 11.Bxd4 Bc6
We are quite some way into the game, but this position has been reached almost 3,500 times in Mega20 - opening 'knowledge' tends to last longer in correspondence chess!
12.f3 Nd7!?
This is the start of a common manoeuvre in the Maróczy Bind, but the move may surprise players unfamiliar with this opening system as Black is 'risking' having his fianchettoed bishop exchanged.
13.Be3!?
This is White's normal reply. 13.Bxg7 is certainly playable but is unpopular for two reasons: a) exchanges normally favour Black in the Maróczy as Black has less space, b) exchanging dark-square bishops leaves White weak on the dark squares, and flies in the face of the rule-of-thumb that when you have a bad bishop (White's light-square bishop here) you should avoid swopping off your good one.
13...a5 14.b3 Nc5 15.Rab1 Qb6!?
Another strange-looking move, but Andrew Greet in Staring Out: The Accelerated Dragon (Everyman 2008) explains: "Station(ing) the queen opposite an enemy bishop ... is only temporary. Black is planning ...Rfc8, after which the queen will drop back to d8. Then the black rooks will be ready to anticipate any opening of the queenside (typically involving a3 and b4), while the queen may find a useful role on the kingside."
16.Rfc1 Rfc8 17.Rc2
Greet gives this an exclamation mark, pointing out that 17.a3 runs into 17...Nxb3! 18.Bxb6 Nxd2 19.Rb2 Nxc4 20.Bxc4 Bd7.
17...Qb4
Greet recommends this "useful refinement," which threatens 18...a4, over the much more popular 17...Qd8. The idea is that after the "more or less forced" ...
18.Qc1 Qb6
... the white queen is misplaced. However he does not say what to do if White replies ...
19.Qd2!?
... which could be a tacit draw offer. Engines prefer White, so I guess objectively Black should play 19...Qb4, but I let the game carry on with ...
19...Qd8!?
So the game has transposed to a mainline position of the Maróczy that occurs 435 times in Mega20.
20.Bf1 b6!?
Greet gives a game in which he played ...Be5 but says he would, if having the position again, "probably prefer" ...h5, which is the most-popular choice in Mega20. The idea is to follow up with ...Kh7 and ...Qh8!? The text is a useful solidifying move that basically waits to see what White does next.
21.Rd1 Rab8!?
The game Emir Dizdarević (2520) - Duško Prelević (2328), Bosnia & Herzegovina Team Championship 2004, saw ...Qd7, Bd4!? Bxd4, Qxd4 when, despite my comments at move 13, White is better, according to Stockfish11 and Komodo11.01.
In 2017 the game Jinshi Bai (2556) - Di Li (2485), Chinese Rapid Championship, was drawn after ...Be5, Bg5 Ne6, Be3 Nc5, Bg5 Ne6 etc.
The idea of the text, which may be a novelty, is to protect b6 so the black queen can move without fear of the b6 pawn becoming en prise.
22.g3 Qf8 23.Bh3 Rd8 24.Bh3 Re8 25.Bf4 Rb7
The engines prefer 25...h6 or 25...Kh8. The text gets the queen's rook off the same diagonal as White's dark-square bishop. Generally speaking in the Maróczy, Black has to bide his time, hoping to get in the pawn-break ...b5 or ...d5 or ...f5. Meanwhile, engines like White's space advantage.
26.Bg5
White too generally manoeuvres, probing for weaknesses.  Stockfish11 suggests 26.Bg2, while Komodo11.01 likes 26.Qf2 and 26.Nc3.
26...h6 27.Bf4 Kh7
White's 26.Bg5 followed by 27.Bf4 gave Black two tempi to play 26...h6 and 27...Kh7. The question is, who has this helped? I cannot give a definitive answer, but the engines are starting to diverge - Stockfish11 reckons White now has the upper hand, ie about three-quarters of a pawn, while Komodo11.01 reckons White's edge is more like a quarter of a pawn. Over the board I doubt if either advantage is all that significant at club level, but it can be a different matter in correspondence chess.
28.Rcc1 Qh8
This is sometimes known as a Réti queen in honour of Richard Réti, who popularised the idea of manoeuvring a queen to a corner square. In this position there is no doubt Black controls the long dark diagonal, but it is less clear whether this gives any special benefit. However it is not as if the queen had much scope on f8.
29.Be3 e6!?
Forcing the knight from d5 must help Black's cause ... provided the black position can stand the weakening of d6.
30.Nc3 Rd7 31.Ne2 Rdd8 32.Qc2?!
The engines suggest 32.Nf4 or 32.Nc3. The problem with the text is it allows Black to get in one of the three desirable pawn-breaks.
32...d5! 33.Bxc5 bxc5 34.cxd5 exd5 35.Qxc5 Ba8 36.Qa7
This is the engines' choice. Clearly White cannot capture on d5 because the e2 knight will hang, and if 36.Qxa5?! then 36...dxe4 when the white king starts to look vulnerable, eg 37.Rxd8 Rxd8 38.f4 Bd4+ 39.Kh1 Bb2 40.Re1 Rd3 with lots of pressure for a pawn, or 37.f4 Rd3, again with plenty of compensation, while 37.fxe4?! Bxe4 seems even worse - the black bishops are very powerful on an open board.
36...Qf8 37.e5
Returning the pawn like this is how the engines want to play it.
37...Bxe5 38.Nd4
Black's passed pawn must be blockaded, according to the engines, eg 38.Qxa5?! d4 39.Rd3 Qe7 is virtually winning for Black, they reckon.
38...Qe7 39.Qxe7 Rxe7 40.f4
Also very playable, and certain to get rid of the bishop-pair, is 40.Nc6 Bxc6 41.Rxc6.
40...Bxd4+!?
Giving up the bishop-pair and leaving White with the better remaining bishop, but Black gets a rook to the seventh.
41.Rxd4 Re2 42.Ra4
42.a4 lets White maintain a blockader on d4, but the position is equal in either case, according to the engines.
42...d4

White only has one good move

*****
*****
*****
*****
*****
43.Bf1?!
Worse is 43.Rxa5? d3, eg 44.Rd1 Bf3 45.Bf1 Rxh2! 46.Rxd3 Rh1+ 47.Kf2 Rxf1+ etc, or 44.Bf1 d2 45.Rd1 Bf3 etc. But White seems fine after 43.Rd1, eg 43...d3 44.Bf1 Re3 45.Rd2 (or 45.Kf2). MS said after the game he got his moves "in the wrong order."
43...Rd2!
This prevents a blockade. True, after ...
44.Rxa5
... White has a pair of connected queenside passed pawns, but the black passer is far enough advanced to beat them.
44...d3 45.Ra7
Komodo11.01 reckons White can draw with 45.Ra4 but eventually comes to agree with Stockfish11 that Black is much better after 45...Bf3 47.Rc7 Kg8 48.Raa7 Re8 49.Re7 (not 49.Rxf7+ Re1) Rxe7 50.Rxe7 Rd1, which virtually transposes to a position - winning for Black - reached in the game.
45...Bd5 46.Re1?!
White almost certainly puts up better resistance with 46.Rac7, eg 46...Kg7 47.R7c3 Be4 48.a4 Rb2 49.Rd1 d2 50.Kf2 Bc2 51.Rxc2 Rxc2 when White has a pawn for the exchange. The engines agree Black is much better, but there is a lot of play left.
46...Kg8 47.h3 h5 48.Ra5 Bf3 49.Rae5 Rd1 50.Re8+ Rxe8 51.Rxe8+ Kg7 52.Kf2 Rxf1+ 53.Kxf1 d2 54.Rd8 d1=Q+ 55.Rxd1 Bxd1
White has two pawns for a bishop but his position is hopeless.
The game finished:
56.Kf2 Kf6 57.Ke3 Kf5 58.b4 Ba4 59.Kf3 Bc6+ 60.Ke3 Bg2 61.g4+ hxg4 0-1