Endgames from the Nationals by Graham Brightwell.

A welcome innovation at last year's National Championships was that the transcripts were collected at the end, and packaged together. Participants at the Nationals should receive the transcript collection with this newsletter, if they haven't already. Credit for this goes to David Shaman, who typed all the games in, and also to someone in France who shifted them into the now-standard GTHOR format.

Anyway, I got to look through more games than just my own, and I thought I'd go through the transcripts, find an instructively badly played one, and annotate it. But I didn't find one: almost all the midgame moves in almost all the games were sensible, thought-out moves. Not always good, but usually hard to criticise. But, an agonisingly large number of games featured a number of endgame errors, ranging from the elementary to the bizarrely complex. Of course, many of these will have been due to time trouble blunders, and all I can say is that, if you're prone to these, then leave more time for the ending. (Easier said than done, I know.) So, to repeat, all of the positions below are the outcomes of well-played midgames, but disaster is about to strike.

Marson v. Haigh; - First off, here is David Haigh, who had a good tournament overall, but here he falls victim to two careless moves. It's worth getting to know about how to play out the three-square region in the SE here. The rule of thumb is that you play whichever of g7 and h7 does not permit h8: if both do, it's often best to start with h8. Here, then White should be starting with h7. Should he play 56b8 57g1 first? Give it a moment's thought. Does it make any difference which order White's move to b8 and Black's to g1 come in? Yes, it does: if Black moves first, then b8 turns the whole b-column. So White should play 56h7. Which is Black's better response? The answer is that he should play g7 first, to keep the d4 disc. Check this: it's a help to be able to calculate this little detail quickly. Here it makes no great difference, but another time it could cost you the game.

For the record, the unfortunate David Haigh, no doubt in the throes of time trouble, played 56b8 57g1, and then compounded his error with 58h8 (P) 59h7 60g7 33-31 rather than 58h7 59g7 60h8 29-35. The difference is that, in the first line, Black gets the last move, and the four discs between d4 and g7, whereas White gets these with the last move in the second line. White-Black-White almost always works out better than White-White-Black.

Haigh v. Wormley; - David Haigh again, to play at 45. Black, I would say, is winning this easily enough. He's obviously worried about the c6-f3 diagonal, and White's threat of g2. Should he be? Well, Black is hardly in danger of running out of moves in any case, so you might even say that g2 will merely lose parity for White. E.g., 45g6 46g2 47d1 48b1 49b2 50h7 51g7, and Black would be winning even without the swindle coming his way in the SW. I think it's a good idea to play g6 before doing anything else, otherwise White is going to play there. You should have in mind the sacrifice g6-h7-g7. Suppose the game goes 45g6 46h7 47g7 48a2. Now play 49d1 50b1 51b7 and the game will be over. It may not be easy to calculate these lines, but Black's advantages ought to be clear: in fact, all the regions on the board play well for him. So this is not the time to panic. Our beloved Secretary chose this moment to play 45b7. This is superficially very attractive, as it seems that White will have to sacrifice to cut the diagonal. Furthermore, 46g6 47a7 is quite hopeless, so White has to play 46a2, met by 47g6, and now White is tremendously short of moves. This is in a way a nice tactical shot, and it would be entirely in order to try something like this from a bad position, but here it is a big risk.

White to play; - What should White do now? Again, this is not the time for panic. What White did do was see a way on to the diagonal, and take it. 48g8? But after 49h8, White discovered that being on the diagonal wasn't all that much good, as taking a8 would allow Black to wedge. Nevertheless, this line (e.g., 50a8 51a7 52b2 53d1 54b1 55a1 etc.) is definitely still a better chance than 50d1, as played, which guarantees Black both h1 and g2 at his leisure. Black won easily.

Back to the diagram. White's best move is the simple 48a7, challenging Black to find something to do. All there is is 49g2, and now White would really like to be on the diagonal, so 50g8!, with White winning easily now.

Wormley v. Selby; - Matthew Selby (White) has just played to b2, and Mark Wormley (Black) has responded by taking the a1 corner (this was in fact a mistake). Now, how is White to win this? This is quite a tough problem, requiring some accurate analysis. Matthew played the routine sequence 52a2 53b1 54a8, and now Mark won with 55b7, but even better would have been 55g7! 56g2 57b7 (P) 58h1 59g1 60h8 42-22. So here is the question: how is White to stop Black gaining parity with g7 at the key moment? Playing 52g7 himself sacrifices too much, but here are two other thoughts: (1) organise matters so that White plays one of the a2-b1 pair, and the response doesn't reflip b2-then g7 is not a diagonalisation; (2) try to play to b7 so that Black can't reply with a8, pulling the same stunt that Black is planning in reverse.

Is 52b7 the way to accomplish these aims? No, since unfortunately Black can reply 53g2 (the ``Principle of the Opposite X-square''), and White is now dead. We are thus led to try 52g2. This works beautifully. If 53g7, then 54b7, so Black does better to play 53b7 himself, met by 54a8. Now 55g7 can be met by 56b1 (or indeed by 56g1), but Black can avoid this too by playing 55g1 56h1 57g7. Unfortunately for Black, these contortions to gain parity cost too much, and 58a2 59b1 (P) is 29-34.

Brewer v. Haigh; - Black has, reasonably enough, played to reach this position. It looks an easy win, but let me start by telling you that, without doing anything outrageous, Phil Brewer lost from here. I would strongly recommend the following line: start with 49b1 50a1 51c1 52d1 53c2 54d2 55e1 56f1 57e2 (P) 58g1 59h1 60g2, then stop and count up. That is a classic example of an ``interior sweep'': Black progressively gives White the top edge, meanwhile taking all the discs in a huge triangle just below it. Another advantage of the sequence is that it costs no time on the clock. (Marginally `better', in fact, is 55e2 56e1 57g2 etc., but who cares? For the real Ebeneezers among you, the line from the diagram guaranteeing the most discs starts 49c1 50a1 51f1.)

Of course, there are other ways to win, but, especially when you're Black, the game won't win itself. The general principle I'd like to state is that, no matter how won the position looks, you should always have a plan, and always be trying to achieve something.

Selby v. Brewer; - Black (Matthew Selby) has gradually leaked away his advantage, and White (Phil Brewer, to move) is in a position to win. How? Phil played 54b2?, and now Black has a neat win with 55g7 56h8 57h7 58b1 59a1 60a2. In fact, Matthew played 55a1?, after which White could have won with 56h8 57h7 58a2 59b1 60g7. Instead, Phil played 56b1?, after which Matthew had a win with 57g7 58h8 59h7 (P) 60a2. In fact, Matthew played 57h7, which also won. But this isn't OQ, so I'll give you an explanation of all this. In a word: parity! Whoever plays first into the SE region gets two of the three moves. True, White can get three of the four in the NW in the meantime, but only by unwedging and giving Black the North edge (as well as the a1-h8 diagonal). So, Black ought to play g7 a.s.a.p., and White similarly with h8. No surprises, then, for spotting that 54h8 is the winning move.

After 54h8, 55g7 doesn't work. White plays, for instance, 56h7 (P) 57b2 58a2 59b1 60a1 29-35. So Black plays 55h7, which not only takes the h-file, but also puts a piece on the second rank ready to sweep across. This is now difficult again.

White to play and win.; - The usual way to treat a position with a one and a four square region, with the opponent having access to neither, is to play into the four-square region, letting your opponent take the next two moves in that region, meanwhile taking the single square, and finishing with a great last move. This technique, called `feeding', is sort-of illustrated by 56b1 57a1 58g7 59b2 60a2. But here the last move is not at all so great, and Black enjoys a substantial victory. The other try for a feed is 56b2, but then Black sweeps up with 57a2, taking the row and destroying the feed. Alternatively, we can think of playing g7 first and then sorting out the four: something like 56g7 (P) 57a2 58a1 59b2 60b1. That sequence makes sense, but reaps only 26 discs. In desperation, we are drawn to playing 56a2! This looks unthinkable, as it gives Black access to the hole, but so what? If Black does play 57g7 (best, indeed), then we get three out of four in the other region: 58b2 59b1 60a1 31-33. Meanwhile, 57a1 58g7 59b2 60b1 shows you how a feed is supposed to work.