Monday, 18 March 2019

046. Creating Problems

White: J. Shepley - C&DCCC Sinclair Trophy, 2019

As mentioned in Game 9: “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).” The same article also had brief analysis of a variation I dubbed ‘The Calabrese Counter-Gambit Deferred’, in that it goes 1 e4 e5 2 Nc3 Nc6 and if 3 Bc4 then 3...f5!?.

This can transpose to the Calabrese itself if, as usually occurs, White responds with 4 d3. But there some independent lines, and I gave a couple of paragraphs on 4 exf5 and 4 Bxg8. The latter of those I appear to have mangled:

“(b) 4 Bxg8 Rxg8 5 d3 (5 Qh5+ g6 6 Qxh7 Rg7 7 Qh3 fxe4 or 7 Qh8 Nd4) 5...d5 6 Qe2? (6 Qh5+ g6 7 fxg6 Rxg6 8 Qxh7 Qf6) 6...Bxf5 7 Nf3 Qd6 8 Bd2 0-0-0 Bixby-Curt, USA 1904 (via 2 Bc4 f5).”

Obviously 5 d3 d5 doesn't make any sense. It should have been 5 exf5 d5, but then 8...Qf6 (in the second bracket) drops the d5-pawn; and I can't find any reference to Bixby-Curt elsewhere, after either 5 d3 or 5 exf5. My original file in fact gives (respectively) 5 exf5 and 8...Bxf5 (Stockfish says 8...Qg5 is better) and doesn't mention “Bixby-Curt” at all. Okay, this was back in the day when we were – or I was – still typing stuff manually into word processor programs instead of entering them in ChessBase files. Mangling was an occupational hazard.

At least the bracketed 5 Qh5+ etc seems to be correct. My own file carried this further: 5...g6 6 Qxh7 Rg7 7 Qh8 Nd4 “-/+” 8 Kd1 (8 d3 f4) 8...Qg5 9 Qh3 (9 g3 fxe4 10 Nxe4 Qf5) 9...d5! 10 Qg3 (10 Nxd5 fxe4; 10 Nf3 Nxf3 11 Qxf3 fxe4) 10...Qh5+ 11 f3 fxe4 12 Nxd5 Rf7 “-+”. Stockfish doesn't have much of significance to add to that. And in the game below, played over 20 years later, my opponent made only four more moves before resigning.

Returning to the opening, I find it a little odd that 3...f5!? hasn't been seen more in chess praxis. MegaBase 2018 has just five games (dating back to I.Rabinovich-A.Flamberg, Triberg 1914), despite Black scoring 4/5. CCDatabase 2018 has another three, with Black scoring 2½/3. In total that's 6½/8 (81.25%) for Black. Even if we add the moves 4 d3 Nf6
5 Nf3, as recommended in Ovetchkin & Soloviov's book The Modern Vienna Game (Chess Stars 2015), and which can arise via various move orders, the figures are still in Black's favour: P51 W28 D5 L18 (59.8%).

I guess it's the professional mindset. After 1 e4 e5 2 Nc3 Nc6 3 Bc4 Black can play simply 3...Nf6 4 d3 Na5 and has no problems. Whereas for us non-professionals, “no problems” may be less appealing. Personally, I'm willing to weigh potential problems (for me) against practical problems (for my opponents). And if they solve theirs, then I have to solve mine. But if they don't...

Wednesday, 27 February 2019

045. AlphaZero nand AlphaZero

Black: AlphaZero - all-play-all tournament,, 2018

I've been reading Game Changer (New in Chess 2019), Matthew Sadler and Natasha Regan's book on AlphaZero: tech company Deep Mind's self-learning AI. AlphaZero was given the rules of chess, played 44 million games against itself in nine hours, and was then the strongest chess player in history, as demonstrated by its trouncing of the world champion computer program, Stockfish 8, in a 1000-game match: W155, L6.

AlphaZero is now my favourite player ever. Garry Kasparov loves it because he sees a mirror of his own style: dynamism as a priority, material secondary to the initiative. But AlphaZero is also like Karpov: subtle prophylaxis, restricting the opponent's counterplay, the opponent's pieces, keeping control. So yes, AlphaZero favours dynamism but only for itself. This is shown in a very simple way in a given graph plotting the number of legal moves each side has available at any point in the game. AlphaZero always seems to have more. It's Kasparov combined with Karpov, enhanced by a computer's incomparable tactical ability. Nigel Short called it God.

Matthew Sadler called it a lunatic. That was while going through a game where AlphaZero, as Black, sacrificed three pawns and its kingside structure for the rather nebulous gains of a well-placed knight, open files for its rooks, and an offside enemy bishop.

Its long-term initiative proved to be so great that even Stockfish was unable to contain it. Sadler's (and my) delight reflects the notion that technological advancement can sometimes seem like magic.

But then Sadler gets down to the business of trying to understand AlphaZero's chess, analysing how it plays, how it thinks, identifying its recurrent ideas, and making it comprehensible to a human readership. Regan does the same thing from the technical side, interviewing the Deep Mind team to gain insight into how AlphaZero actually functions, how it learns, and again how it thinks. And these two approaches “meet in the middle”. Game Changer is an extraordinary piece of work.

You'll notice that my opponent in the game below is AlphaZero. But not the AlphaZero. Just someone who has taken the name “AlphaZero” as their moniker. All the same, "human" AlphaZero is a strong and original player. We've had six games so far (four still ongoing) and I don't look like winning any of them. In the one below "it" adopts an unusual defence to the King's Gambit: 2...exf4 3 Nf3 Nc6 4 d4 f5. This is supposed to be bad, but winning a pawn and wrecking Black's structure was not enough to get me past opposite-coloured bishops, even with rooks on.

Well, there's no use in wondering what the real AlphaZero would have done because AlphaZero doesn't play the King's Gambit. Or 1 e4. AlphaZero plays 1 d4 and sometimes 1 Nf3. In other words, after due and deep consideration, AlphaZero has definitively pronounced: by test, 1 e4 not best.

Monday, 24 December 2018

044. Surprising Openings

White: J. Rudd - C&DCCC Minor Counties, 2002

Whether or not an opening, or a variation of an opening, is a surprise naturally depends on whether an opponent expects it or is otherwise prepared. In an information-neutral context – i.e. without knowledge of someone's repertoire or, conversely, their knowledge of your own – the element of surprise can therefore be a matter of chance. My openings tend towards the disreputable, so they tend to have decent surprise value – most pleasingly demonstrated when an opponent starts thinking on move four.

But in this post I'm not concerned with opponents’ reactions. Instead, I've been looking through my old CC games, sorted variously by ECO code and notation, and have been surprising myself with some of my opening choices. For instance, as White: 1 e4 c5 2 a4!? (one game), 1 c4 (three games); and as Black: 1 d4 e5!? (one game), 1 d4 c5 2 d5 Nf6 3 c4 Ne4!? (one game), 1 e4 c5 2 Nf3 d6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 a6 (another three games). Okay, I can deduce that the single instances there were prompted by the writings of Hugh Myers and Stefan Bücker. But what I was doing playing the English and the Najdorf, I really have no idea.

In the blog-relevant C20-C99 category, there are surprises too: 1 e4 e5 2 Nf3 d5!? (three games), 1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 f5!? (three games). I don't remember those either, especially the three Elephants. The games with the Schliemann Deferred (all from 2002) are slightly less surprising. I have occasionally played this over the board, usually aiming to take opponents (who might be ready for 3...f5) away from standard paths, since the seemingly benign inclusion of ...a7-a6 and Ba4 actually changes the position considerably. To obtain an advantage White has to play 5 d4 (a sideline in the Schliemann itself), and if 5...exd4 then 6 e5 with a favourable sort of reversed Falkbeer. I see that one of my CC games did go this way – and I lost – so whatever new ideas I might have had probably weren't very good. I'll have to look at that again sometime.

The game below took a less critical path: 5 d4 exd4 6 Nxd4 Nxd4 7 Qxd4 c5!.

Here the obvious threat to the a4-bishop (with ...b7-b5 etc) enables Black to escape the opening with a satisfactory position. This increased by degrees in the middlegame, and reduced more sharply in the endgame, until my opponent (now an OTB IM) was rightly justified in offering a draw (at move 35), which I declined – a decision equally and instantly justified when they blundered by reply. Sorry, Jack.

Wednesday, 12 December 2018

043. Form is Temporary

White: R.W. Bauld - BCCA Championship, 1994/95

I'm playing very badly at the moment and don't know why. I've hardly won a league game in over a year and am now losing to people who haven't beaten me in a long time - or even at all. So far this season, I've had a lost position in every one of my nine games. That I've drawn five and even won a game doesn't make me feel much better. From a peak rating of 221 (ECF) eighteen months ago, I'll have dropped 30 points by the time the next list comes out.

Well, form comes and goes, I suppose. And drastically so for me in correspondence chess. For instance, after winning the BCCA Championship for the second time (and consecutively) in 1993/94, in the following year I came last. The game below is from the 1994/95 tournament.

It features another "deferred" Open Game. To be precise, it's a type of Réti – a reversed Classical Pirc – to which ECO duly assigns the code A07. But it could just as well be C42 or C44. Indeed, the first game in the database to reach the position at move seven (B.Ivkov-L.Rellstab, Bled 1950) took the route 1 e4 e5 2 Nf3 Nf6 3 d3 Nc6 4 g3 d5 5 Nbd2 Be7 6 Bg2 0-0 7 0-0. So there you are.

It's also a slightly unusual game in that, having built himself a fortress, my opponent took to "passing" from move 25 on, giving me unlimited time to try and find a way in. My written notes give no indication as to why I didn't take the e-pawn on move 29. Instead, I sent my knight on a trip from f6 to a8 and then round again to g5. This did in fact enable me to break through, though it was only a mistake on move 47 that gave me a winning position. And then I fell for a stalemate trick.

It was another three years before I was back in the Championship and won it again, this time with my best ever score of +6.

Wednesday, 28 November 2018

042. Days of Innocence

Black: C.S. Thomson - BPCTC, 1993-94

Once upon a time, when people used to play 1...e5 against me (which no one has done OTB for three years now), I got to utilize a lot of "unexpected" variations in the King's Gambit. One of these was the Rosentreter: 2...exf4 3 Nf3 g5 4 d4!?, intending 4...g4 5 Bxf4 gxf3 6 Qxf3 with practical chances (see here, for instance), based on the fact that Black has no pieces out at all.

Some players, perhaps surprised by 4 d4, would try to reach a standard set-up with 4...Bg7, intending ...h7-h6, ...d7-d6 and so forth. Only to be confronted by 5 h4 h6 6 hxg5 hxg5 7 Rxh8 Bxh8 8 g3! (Keres), after which they found themselves instead in a fringe line of Becker's Defence, not knowing what to do.

I used to do quite well with this line (8...g4 9 Nh2! is the main point), even against FIDE titled (or future titled) players, such as: Gregory Kaidanov, Mark Hebden, James Cobb, and Robert Eames. My opponent in the game below, Craig Thomson, has a title too (he's an FM – whether at the time or later, I'm not sure), though the Becker wasn't a surprise to him because he started it off with 3...h6. (In any case, the element of surprise tends to be more potent in rapid chess than postal chess.) All the same, he didn't manage to defend correctly, and I won – as I thought then – a nice game.

They feel like days of innocence now. Neither that game, nor any of the others mentioned, stands up to engine scrutiny – especially the one against Hebden, which exemplifies Tartakower's maxim: that the player who wins is the player who makes the next-to-last mistake; i.e. it didn't matter that Hebden blundered first (17...Qf5??), nor that several other moves by both sides were also blunders, because I blundered last (38 Bf4??) and therefore lost.

If there's a reason I stopped playing 5 h4 and 8 g3, I've forgotten what it was. Maybe there wasn't a reason. Today, Stockfish gives Black the edge, but not fatally so. Or as John Shaw put it in his book: “The King's Gambit – not quite as stupid as it looks.” On the other hand, on the few occasions in recent years I've gotten as far as 2...exf4 as White, I've only played 3 Nf3 twice. Coincidentally, the last of those was against Hebden again (in 2016), where he blundered first again, and won again. Some things never change.

Thursday, 1 November 2018

041. My First CC Tournament

Black: D.J. Finnie - BCCA Major, 1988-89

I can't remember why I started playing correspondence chess. Likely I picked up a leaflet somewhere and decided to send it off. After playing chess as a junior – and packing it in at secondary school – I returned to it again in 1985 when the local chess club (Mansfield) suddenly turned up in my regular, the now defunct Stag & Pheasant (which is not the Wetherspoons of the same name). Three years later I entered the BCCA pyramid tournament for the first time.

By then my OTB rating was 163, which meant I was placed directly in a Major section, two divisions below the Championship. All-play-all, both colours, five opponents, ten games. I won nine. Going through them with Stockfish thirty years on, the engine inevitably points out a few mistakes but surprisingly few serious ones by me. It seems I actually played quite well (see here, for instance). Except in the game below.

In this one Philidor's Defence arrived via a Pirc move order: 1 e4 d6 2 d4 Nf6 3 Nc3 Nbd7 4 Nf3 e5 5 Bc4 Be7 6 0-0 0-0.

I've faced this (Hanham) system twice as White and lost both times (once OTB in 1993). Since then I've studiously avoided it, opting variously for (3...Nbd7) 4 f4 (or if 3...e5 then 4 Nge2), or 3 f3, or 2 Nc3 without d2-d4 at all. It's interesting... Switch the colour of every piece in the diagram and I'd be perfectly content to grind this out as "Black", so why should it be more difficult with White? I guess it's a psychological problem. Or else just me. I'm sure Magnus Carlsen would be more than happy to take the white pieces here, and as often as he possibly could.

Well, anyway, looking over my BCCA game again now, I have to say I'm struggling to understand some of my moves:

13. Why not play Qf3 straight away, instead of wasting time with the dark-squared bishop?
15. Why then leave the bishop on f4 to be captured?
17. Why not take on e6 at once, instead of wasting time with the light-squared bishop?
23. What on earth was Rd4 for?
29. Okay, this was just a blunder.

But I can't be sad (can't be sad), ’cause nine out of ten ain't bad. I even won a game with Alekhine's Defence.

Monday, 15 October 2018

040. Old Novelties, New Novelties

White: Glawurtz89 - all-play-all tournament,, 2018

Twenty years ago in an ICCF thematic I played the novelty 19 Rf5 h6 20 Bf4 h5!. This was my improvement on analysis (20...Rfe8 21 Rg1) by Victor Ivanov and Alexander Kulagin in their very interesting book Play the Schliemann Defence! (Olbrich 1994).

It was indeed good, but I failed immediately to back it up. After 21 Bg5, my next move 21...Kc7? could have been met by 22 Qd2!, since 22...hxg4? 23 Qa5+ Kc8 24 Qxa7 g3! (as I'd analysed) fails instead to 24 Nc4! and if 24...g3 then 25 Nxd6+ Rxd6 26 Bf4 Rd7 27 Qxa7 wins for White. Fortunately, my opponent played 22 Ng6?, made some more mistakes, and finally allowed me a win with 31...f2! 32 d5 R8e4! 33 dxc6+ Kc7 – which I missed. I didn't even realize I was winning. This is all according to the engine of course.

Going back, the correct move was 21...Bc7, which seems to work out for Black. In the game below I intended to give it a go – until my opponent played 19 Rg1!, planning simply to trap the queen with Rg3.

No worries. I'd looked at this back in the day. Ivanov & Kulagin analysed the further 19...Bxe5 20 dxe5 Rxe5 21 Rg3 Qxg3 22 hxg3 Rxg5 23 Qe3 Rxg4 24 Qxa7 Re8, giving the appropriate symbol for “with counterplay”.

Okay, I'll just check that with the engine... 25 Rd4 Rg5 26 Rb4 b5 27 Rb3 Rh5 28 Ra3 e3 29 Rxe3 “+- (2.22)”. In human-speak: “White is winning”. Bollocks!

I had to leave the g4-pawn alone, grovel with 23...Re5 24 Qxa7 Rfe8, and hope that the e-pawn would still offer enough.

Well, as it turned out, my rooks managed to get in on the second (at the cost of a couple of pawns) and eliminate the white queenside, leading to a rook plus one vs. knight plus two endgame where the black e-pawn was enough. For instance, at the end, 46 Rxh7 c4+! 47 Kxc4 e2 48 Rh1 Ne3+ 49 Kd3 Nf1 draws.

All the same, if there are any novelties to be found post 23...Re5, they're going to be White's. I think I'll look for an improvement earlier on next time.