Tuesday, 31 January 2017

011. Chess Correspondence

White: D.M. Andrew - BCCA Premier, 1991

Not only did we used to play chess by post – with cards, envelopes, stamps and so forth – we used to correspond about chess by post as well. My two most regular correspondees, both now deceased, were Otto Hardy and Donald Andrew. Otto was a significant opening theoretician and his letters were full of his own games and analysis, of which I still have three files worth. Donald was... well, I'm not sure what Donald was exactly. Perhaps the most apt word is “enthusiast”. Donald used to collect other players’ game scores and pick their brains about opening analysis. As well as me, I think he also wrote to and received stuff from Jeff Horner and John Littlewood, and very likely there were more of us. For services rendered, as it were, Donald used to send me an occasional book of first class stamps, which I used for my postal games.

Googling today, I discover that Donald was Yorkshire Champion in 1949, came joint second in the British Major Open the same year and, much later, was joint British Senior Champion (in 1985). A game of his features in the updated edition of John Littlewood's book How to Play the Middle Game in Chess (Batsford 2001), where John wrote:

[I]nexperienced players have somehow acquired the erroneous belief that middle games with opposite-coloured bishops are also drawish. Nothing could be further from the truth! In fact, I am tempted to generalize by stating that, in middle game situations, opposite-coloured bishops tend to unbalance play and tilt it even more in the favour of the player with the initiative. As an instructive example of this, I quote a correspondence game played recently by an old friend of mine, Donald Andrew:

That game was Andrew-Roach, correspondence 1999: 1 e4 e5 2 Bc4 Nf6 3 d4 exd4 4 Nf3 Nc6 5 e5 d5 6 Bb5 Ne4 7 Nxd4 Bd7 8 Bxc6 bxc6 9 0-0 Bc5 10 f3 Ng5 11 f4 Ne4 12 Be3 Qe7 13 Nd2 Nxd2 14 Qxd2 Bxd4 15 Bxd4 c5 16 Bf2 d4 17 c3 Bb5 18 Rfe1 d3 19 c4 Bxc4 20 Rac1 Ba6 21 Rxc5 0-0 22 Ra5 Qe6 23 f5 Qc8 24 f6 h6 25 Re3 Rd8 26 Rg3 g5 27 Rxg5+ hxg5 28 Qxg5+ 1-0.

I never actually met Donald in person, but I played him five times in BCCA tournaments. All the games were drawn. The first of these followed (or rather, transposed to) the same line of the Two Knights Defence as the one above, until Black deviated at move seven. As it happens, Donald could have gone for opposite-coloured bishops in our game too.

Here 13 Bxc5 gxf3 14 0-0! was possible, after which my handwritten notes give a terse “14...Bh3!” and no further. Twenty-five years on, Houdini continues this line with 15 Re1 Bxg2 16 e6 Rd8 17 exf7+ Kxf7 18 Re7+ Kg6 19 Bd4 Qf5 20 Rxg7+ Kh5 21 Rxg2 c5 22 Rg3 Rhg8 23 c3 and claims a slight advantage for White. This is perhaps why virtually everyone else has preferred 11...Qe7.

Tuesday, 17 January 2017

010. The Weakest Square

Black: L.T. Ellis - BCCA Gambit Tournament, 1999

As all beginners quickly learn, the weakest square in each sides' position at the start of the game is KB2 – that is f2 for White and f7 for Black – which is only protected by the king. We learn this by losing (often more than once) to sequences like 1 e4, 2 Bc4, 3 Qh5 and 4 Qxf7 mate. One of the aims of the King's Gambit, too, is to open the f-file and target this square with as many pieces as possible.

In the following game, White did indeed aim for f7 with 6 Qxf3 and 7 Bc4. For instance, 7...Nxd4? 8 Bxf7+! Kxf7 9 Qh5+ Kg7 10 Be5+ Nf6 11 Bxd4 and wins, or similarly 7...d5 8 Bxd5 Nxd4 9 Bxf7+! Kxf7 10 Qh5+ Kg7 11 0-0 with strong play for the two sacrificed pieces in the style of the Double Muzio. But Black forestalled any such ideas with his early queen manoeuvres.

Instead, the weakest square turned out be Q2 (d7), which is covered initially by four pieces: b8-knight, c8-bishop, queen and king. In fact three of them (knight, bishop and king) were still defending it in the key position at move 31.

The trouble was that the knight and bishop had not moved throughout the game, nor had the a8-rook, nor did they. Meanwhile White had amassed five attackers: knight, bishop, queen and two rooks, and could theoretically add the e-pawn as well if required.

In the end, the "weak" Q2 square didn't collapse after all. But preventing that cost Black too much and he soon resigned.

Friday, 6 January 2017

009. Patzer Sees a Check

White: stormytlc - thematic tournament, ChessWorld.net, 2011

Talking of articles, I once wrote an article on 1 e4 e5 2 Bc4 f5!? (the Calabrese Counter-Gambit) for a special issue of Tim Harding's magazine Chess Mail (May 1997). Nineteen years ago. Damn. My theoretical investigations have moved on quite a bit since then.

Jänisch's "refutation" 3 d3 Nf6 4 f4, which I indicated as still being the critical line, had in fact already been neutralized by Mark Lyell: 4...Nc6 5 Nf3 fxe4! (I concentrated on the inferior 5...exf4, utilizing some transpositional analysis from Matthias Wahls) 6 dxe4 Nxe4 7 fxe5 Nxe5! 8 Bd5 (or 8 Nxe5 Qh4+) 8...Nxf3+ 9 Qxf3 Nf6 and Black is fine, J.Emms-M.Lyell, British Championship, Southampton 1986.

In his later book on the Italian Game and Bishop's Opening, Beating 1 e4 e5 (Everyman 2010), John Emms suggested two other possibilities for White: 3 d4!? exd4 4 e5 d5 5 exd6 Bxd6 6 Ne2 Nc6 7 0-0 Na5 8 Bxg8 Rxg8 9 Nxd4 “with advantage, J.Pietrasanta-K.Shirazi, Pau 2008”; and the simple 3 d3 Nf6 4 Nf3.

The first, a sort of reversed Falkbeer, was missing from my article. I subsequently faced 4 e5 three times (1998-2001), though White always retreated the bishop in my games. After 5 exd6 etc, Shirazi's play might be improved by 7...Qh4!?, when something like 8 g3 Qf6 9 Nd2 Ne5! 10 Nxd4 Bd7 and ...0-0-0 is quite unclear.

The second line, 3 d3 Nf6 4 Nf3, can indeed be tricky if Black develops "normally". I gave (among other things) 4...fxe4 5 dxe4 Nxe4 6 Qd5 Nd6, following L.Bledow-P.von Bilguer, Berlin 1839, which looks extremely dodgy to me now, especially if White just plays 6 0-0!. Instead, Emms notes that 4...Nc6 5 0-0 Bc5 6 Nc3 d6 7 Bg5 “is a King's Gambit Declined with reversed colours, and 7...Na5 8 Bxf6! Qxf6 9 Nd5 Qd8 10 b4! Nxc4 11 bxc5 was somewhat better for White in D.Fryer-M.Lyell, Hastings 2003/04.” The problem is the combination of Bg5 and Nc3-d5 which the natural ...Bc5 does nothing to prevent. Trying to solve this led me to the patzer's variation 4...fxe4 5 dxe4 Bb4+!?.

White has four reasonable ways to block the check, all of which rule out the Bg5 and Nc3-d5 plan: (i) 6 Nc3 sees the knight pinned; (ii) 6 Nbd2 puts it on the wrong square; (iii) 6 Bd2 allows the bishop to be swapped off; (iv) 6 c3 Bc5 leaves the c3-square obstructed by a pawn. Almost all of my opponents have chosen option four, when White's position does look rather good. It will take at least four moves for Black to evacuate the king from the centre, while the c5-bishop is an obvious target for space-gaining advances on the queenside with b2-b4 and a2-a4-a5. Nevertheless, it turns out that it's not so easy for White to prove anything, while Black gets on with the slow plan of ...d7-d6, ...Qe7, ...Be6 and ...0-0. It's often possible (and better) to insert ...Nc6-d8 before ...Be6 as well.

The game below was one of my earliest with this set-up. As it happened, my opponent managed to keep me from castling short, but by that time it was okay to go long. Note that 18 Qxd6?? would lose for White after 18...Nb8, while 18 Na3 Rhe8 19 Qxd6 Qf4 20 Qd2 Nf6 21 Qxf3 exf4 is fine for Black. And also that no one has yet managed, in my 16 further games, to cast doubt on Black's position after 5...Bb4+. It may be possible to cast doubt on the whole idea of 2...f5, but that will have to wait for a future post.