Open chess diary 141-160

Recently on ICC I had, with Black, the position on the left. It was my opponent's move, and he played the winning 51.Rf8. After my 51...Rxf8 however, he didn't answer immediately, making it clear that he had seen the stalemate after 52.gxf8Q. Of course, gxf8B and gxf8N+ both win easily, but as I was waiting for his move, it dawned upon me that he had set the option "Always Queen", and that he did not know how to underpromote. Had I been a gentleman, I would have resigned, but as it was, I took pleasure in picturing him frantically looking for a manual, or typing things like "help underpromotion". In the end, he offered a draw - and I accepted.
The command, my good O*, is f8=B. But I don't know how this works on other servers. A reader, Dieter Huttenlocher,

sent me a funny thing that happened in the FIDE "Game Zone". The diagram on the right arose in a game noel - _next played there in 2001, with Black to move. He had enough time to see that he must not play c1Q, but c1R, and mate next move. Not having set 'always queen' (or that server's equivalent - I've never played there) he played c2-c1 and, getting a window with the four different pieces, carefully clicked the Rook - only to watch in horror as a Queen appeared on c1 anyway and the game was declared a draw.
    This is of course not possible in a regular game, but something else is possible there that might need a law to prevent it. It is perfectly legal for Black in the last diagram, to play c2-c1, toss that pawn aside, pick up the Queen from among his captured pieces, say "Queen", and even lower that Queen to the board, touching the promotion square - and still have the right to discard it and place a Rook on c1 after all. As to touch-move, the FIDE "Laws of Chess" specifically mention the touching of pieces "on the chessboard".
    There is nothing about naming pieces.

PS 22 December 2005: Christoph Pfrommer calls my attention to the fact that in the latest version of the FIDE Laws of Chess (October 2004), it says: "If a player promotes a pawn, the choice of the piece is finalised, when the piece has touched the square of promotion."
    So you can still pick up any piece and choose a new one, but only before you touch the board with it.

Just out, with Bloomsbury, the paperback edition of my novel The Cave.
    In a near perfect review, the London Evening Standard calls it "a near masterpiece", also writing: "rare delicacy and perception... a very satisfying psychological drama... almost too perfect... will merely have you tearing through the pages."
    The Face: "A tale of Mephistophelian friendship [that] ambitiously bursts its bounds... expert structuring ... Graham Greene-esque travelogue."
    Sainsbury's Reader's Magazine: "Tim Krabbé cranks up the tension from the first paragraph of his latest novel... As this cleverly constructed tale unfolds, we watch as Egon looks back on events that have led up to [his] defining moment..."
    The Observer: "This sombre, utterly absorbing story... a masterly work, concise, stylistically bold and full of surprises."
    The Guardian: "For such a short book, The Cave has an awful lot in it... the language is cold and precise."
    Compare item 76 in this Diary, which is also not about chess.

As a Christmas gift, Bader Al-Hajiri from Kuwait sent me a proof game for the well-kown problem on the left, a Mate in 2, by T.R. Dawson, first published in the Falkirk Herald, 1914. In June, of all months, even when it is an upside-down Christmas tree, and has become known as such.
    The solution is easy and difficult at the same time. Black can only have played d7-d5 or f7-f5 as his last move, each allowing an en passant capture and mate next move. But which which of the two did he play? The white pawns must have captured 10 times, which accounts for all the missing black pieces. That means Black's last move cannot have been d7-d5, because in that case the Bc8 was captured on its starting square, and a pawn could not have done that. Therefore, Black's last move was f7-f5, and White mates with 1.gxf6 and 2.f7 mate.
    Dawson composed this problem to show that in a symmetrical position, apart from the extra space on one side and the castling rule, there can yet be another motivation for one asymmetrical key.
    To reach the position, White must have made at least 27 moves - 4 by the King, 7 by the Knights, and 16 by the pawns. But it seems impossible to come close to a 27-move proof game, because in that ideal game, all the missing white pieces must have been captured on their starting squares and only a promoted black pawn can have done that. Still, Al-Hajiri needed only one extra move (16.Qb3) in this Christmas tree proof game:
1.h4 g5 2.hxg5 Nf6 3.gxf6 e5 4.a3 Bb4 5.axb4 c5 6.bxc5 Rg8 7.Nh3 Rg3 8.Na3 Re3 9.dxe3 d6 10.c4 Nc6 11.Kd2 Ne7 12.fxe7 Qa5+ 13.Kd3 Qc3+ 14.bxc3 a5 15.Nb5 a4 16.Qb3 axb3 17.Nf4 b2 18.Nd4 bxa1Q 19.Ng6 Qxc1 20.Nc6 Qxf1 21.g4 Qxh1 22.f4 Qd5+ 23.cxd5 Be6 24.dxe6 d5 25.g5 e4+ 26.Kd4 Ra4+ 27.Ke5 Rd4 28.cxd4 f5 and White mates in 2 moves.

Thanks to Joose Norri for showing me a game he played in the recent European Team championship, in which one of my favorite themes, the 'Unguarded Guard' occurred twice, and which ended with another, 'Catastrophe in the Queen's Ending'.

Norri - Borgo (Finland - Italy), Leon 2001
1.d4 d5 2.c4 e6 3.Nc3 Nf6 4.Nf3 Nbd7 5.Bg5 c6 6.cxd5 exd5 7.e3 Be7 8.Bd3 Nh5 9.Bxe7 Qxe7 10.Qc2 g6 11.O-O-O Nb6 12.h3 Be6 13.Kb1 Ng7 14.g4 O-O-O 15.Na4 Nxa4 16.Qxa4 Kb8 17.Rc1 Ne8 18.Rc3 Nd6 19.Rhc1 Bd7 20.Qb4 Rhe8 21.Rb3 Rc8 22.Ne5 Rc7 23.g5 Rec8 24.h4 f6 25.gxf6 Qxf6 26.a4 Nf7 27.Ba6 Bf5+ (see diagram) 28.e4 With the intention to lure away the Bishop from d7. It doesn't

come to that. 28...c5 29.Rxc5 Qxa6 30.exf5 Nxe5 31.Rxc7 Rxc7 32.dxe5 gxf5 33.Rg3 Qf1+ 34.Ka2 Qc4+ 35.Qxc4 Rxc4 36.Rg8+ Kc7 37.Rg7+ Kc6 38.Rxh7 Rxa4+ 39.Kb3 Rf4 40.Rf7 Rxf2 41.h5 f4 42.h6 Rh2 43.h7 f3 44.e6 f2 45.e7 Rh3+ (see diagram) In this interesting position, Norri found the brilliant 46.Rf3! "the only move to avoid immediate defeat," as he wrote. Indeed, something like 46.Kc2 loses to Kd7 etc., and Black remains two pawns up. But after 46.Kb4(!) Kd7 47.Rxf2 Rxh7 48.Kc5 a drawn endgame would result. However, in this variation, Black plays 46...Rh4+ Now, 47.Kb3 Kd7 is lost, but although White has a new Unguarded Guard: 47.Rf4!, he probably loses: Rxf4+ 48.Kc3 Rc4+! 49.Kd2 (or Kb3) Re4 49.h8Q f1Q 50.e8Q+ Rxe8 51.Qxe8+ and the Black King having access to d4, Black has good chances of winning.

Norri adds: "But that's a detail. More interesting is the fact that although I more or less decided to play 46.Rf3 while my opponent was still considering his 45th, it never occurred to me to consider 46.Kb4 Rh4+ 47.Rf4. I just thought 'Rh4+ wins'... Strange."
After 46...Rxf3+, the game continued: 47.Ka2 Re3 48.h8Q f1Q 49.Qh6+ Kd7 50.Qxe3 Qc4+ 51.b3 Qa6+ 52.Kb1 Ke8 53.Qe5 (see diagram) After Qd3+ or Qb5 Black might still have some winning chances, but after 53...Qc6?? 54.Qg7 Kd7 55.Qf7, he had to resign.

A friend gave me Josef Breuer's problem bible Beispiele zur Ideeengeschichte des Schachproblems, something like: 'An illustrated History of Ideas in the Chess Problem.'
It is a 400-page, 1800-problem behemoth of a book, leaning heavily toward the 'Logical New-German' school - which is fine with me because, as far as I know anything about problem composition schools, that is my favorite one.
Browsing, with a special eye for moremovers, and

Mate in 8
A. Popandopulo, 1964

names like his, I saw this eightmover (see diagram) by the Russian Awenir Popandopulo (1920 - 1988) whom I admire as one of the greats of what I like to call the romantic mechanism.
    The first things you notice are the two mates g5 and Nf7 which, for the moment, are both prevented by a black Rook; Ra4 pinning Pg4, Rb7 guarding f7. This suggests bothering these Rooks, which can be done in a number of ways.
    To start with the Ra4, its pinning line can be interfered, for example with 1.d4. The Rook must capture (Rxd4), but is thereby lured to a square where it can again be attacked: 2.Rd3 If, after Rf4, White then plays 3.Rxd6, intending Rd8, Nf7+ and Rh8 mate, Black wins too much time with 3...fxg4 After 4.Rd8, he then has g3+ 5.Ng4+ Rxg4+ 6.Kxg4 Rb4+ and White is too late, and after 4.Nxg4+ Rxg4+ 5.Kxg4 Rb4+ 6.e4 Rxe4+ 7.Bxe4 g6 White is one move late, too.
    1.Rxa3 also doesn't help White; Black plays 1...Rf4 and White hasn't progressed.
    Perhaps then, White must attack the other Rook. 1.Rb3 is obvious, because if the Rb7 moves, for instance to c7, there follows 2.Rb8 and mate with Nf7+ and Rh8. However, Black plays 1...Nb6! (2.axb6 a6) and again, White has just lost time. Another obvious move is

1.a6!, threatening 2.axb7 and 3.Nf7 mate, and that in fact is the solution. The Rb7 must keep guarding f7, but whichever square along the 7th rank it chooses, it has a drawback. If 1...Rc7, then 2.c4! (why not d4 or e4 will become clear), 2...Rxc4 3.Rc3! Rxg4+ (if Ra4, then 4.Rxc6 fxg4 5.Rxc7 g3+ 6.Ng4+ Rxg4+ 7.Kxg4 and 8.Nf7 mate. Had White played 2.d4?, and therefore 4.Rxd6 in this line, Black could then stave off the mate one extra move with 4...fxg4, e.g. 5.Rd8 g3+ 6.Ng4+ Rxg4+ 7.Kxg4 and White needs two more moves.) 4.Nxg4+ fxg4 5.Rxc6! Bursting through the second line of defense. The Rc7 must flee once more, and wherever it goes, e.g. 5...Rd7, there follows 6.Rc8, 7.Nf7+ and 8.Rh8 mate.
    It will be clear what happens if after the key, the Rb7 chooses another flight square:
1...Rd7 2.d4! Rxd4 3.Rd3! Rxg4+ 4.Nxg4+ fxg4 5.Rxd6! R- 6.Rd8, 7.Nf7+ and 8.Rh8 mate
1...Re7 2.e4! Rxe4 3.Re3! Rxg4+ 4.Nxg4+ fxg4 5.Rxe6! R- 6.Re8, 7.Nf7+ and 8.Rh8 mate.
    The white Rook smashes through two pawn walls, a white one and a black one, and past two Rook guards to get at the King. This theme, where two pieces fight a duel, is called opposition in problem jargon, a term which has a different meaning there than in ordinary chess. Officially, this is a multiple double Rook-Rook opposition.
    I prefer to picture a medieval battling tower, trampling foot soldiers, knocking down two enemy towers, before conquering the castle.

The sequence: Queen sacrifice, double check, mate, has happened many times in many forms. The most famous example is the coffehouse game Réti - Tartakower, Vienna 1910, which is also one of the nicest because each of the doublechecking pieces mates in one variation. Thanks to Richard Vondruska for showing me an example with reversed colours. Strangely, these are the only games I could find in my database with those twin Rook & Bishop mates on d8/d1.


Réti - Tartakower, Vienna 1910 1.e4 c6 2.d4 d5 3.Nc3 dxe4 4.Nxe4 Nf6 5.Qd3 e5 6.dxe5 Qa5+ 7.Bd2 Qxe5 8.O-O-O Nxe4 (see diagram) 9.Qd8+ Kxd8 10.Bg5++ and Black resigned.	Dutch - Sugden, London 1964 1.e4 c5 2.d4 cxd4 3.c3 d5 4.exd5 Qxd5 5.Nf3 Bg4 6.Qa4+ Nc6 7.Nxd4 Bd7 8.Nb5 O-O-O 9.Nxa7+ Nxa7 10.Qxa7 (see diagram) Qd1+ and White resigned.

PS 22 December: I would have been naive if I had thought this had never been published before. Rik van der Heiden pointed out to me that both diagrams had already appeared side by side in a Kurt Richter book, Chess combination as a fine art. A strange book - both the English version (1976) and the original GDR Sportverlag book Schönheit der Kombination (1972) have covers that mention Golz and Keres as the authors, when Keres only wrote a foreword, and Golz only collected Richter's pieces from SCHACH. I have the German version, and now discovered I had seen those twin mates there, and had even made a note of it. Can't remember when my memory last worked.

Van der Heiden also showed me another nice example of the doublechecking mate. (See diagram.) In the Bundesliga (Germany) game Tal - Ramos, 1991, Black resigned because 29...Kd7 30.Qd8+ Kxd8 31.Bf6++ etc.is mate. After 29...Ke7, 30.Bf6+ is quickest, but of course Qd8+ works in that case, too.
My surprise remains that Réti - Tartakower and

Dutch - Sugden seem to be the only games where this twin Rook-and-Bishop mate occurred. Looking for more, I did find this: 1.e4 Nf6 2.Nc3 d5 3.e5 d4 4.exf6 dxc3 5.fxg7 cxd2+ 6.Bxd2 Bxg7 7.Qh5 Qd4 8.Nf3 Qxb2 9.Rd1 Qxc2 10.Qd5 (see diagram). These moves were played in Linder - Krugliakov, Moscow 1965, and in Maitzegui - Alio, Calella de Mar, 1985 Winning two pawns brought Black some difficulties, which were not solved by 10...c6? 11.Qd8+ and mate.

PS 27 December: David Paulowich sent me a third game with the

twin Rook- and Bishop mates - a remarkable game, because the characteristic Queen sacrifice could have happened earlier.
Galula - Andor, Paris 1954
1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.Bd2 dxe4 5.Qg4 Nf6 6.Qxg7 Rg8 7.Qh6 Qxd4 8.O-O-O Ng4 In four games, Kilchsperger - Waldhauser, Basel 1951; Miles - Marshall (Tony, but not Frank), Wolverhampton 1968; Lundquist - Philipps, Ostrig 1980, and Mitchell - Hamilton, cr Scotland, 1980, there followed 9.Qh4 Qxf2?? 10.Qd8+ and mate. But Galula played 9.Qxh7 and the game went on: 10...Rh8 10.Qxe4 Bxc3 11.Qf3 Bxb2+ 12.Kb1 Nxf2 13.Bb5+ c6 14.Ne2 Qb6 15.Kxb2 Qxb5+ 16.Ka1 Nxh1 17.Qf6 Qxe2?? 18.Qxh8+ (see diagram) 18...Ke7 Now Bb4+ is mate in 2, and Bg5+ is a pedestrian mate in 3, but of course 19.Qd8+, with its twin mates in 3, was the nicest. 19...Kxd8 20.Bg5++ Kc7 21.Bd8 mate.

A nice idea by Pim Blijlevens, which I hadn't seen before: what is the

shortest game leading to a position where selfmate is possible? (In a selfmate, one side, normally White, forces the other to mate him.) He gives this 10-ply solution: 1.f4 e5 2.g4 Ke7 3.Nc3 Kf6 4.Nf3 Kg6 5.Nd5 Nh6 (see diagram on the left) and now 6.Nh4+ Qxh4 mate.
And its counterpart: 1.e4 f5 2.Ke2 g5 3.Kf3 Nc6 4.Kg3 Nf6 5.Nh3 Nd4 (see diagram on the right) and 6.Qh5+ Nxh5 mate. Of course in both solutions, many interpositions are possible.
Blijlevens thinks a shorter helpselfmategame than 10 ply should be possible. I'm not so sure. Anyone?

PS 11 December: Sasho Kalajdzievski showed yet a different solution: 1.e4 h5 2.f4 g5 3.Kf2 f5 4.Kg3 gxf4+ 5.Kh3 (see diagram) and Black doesn't need his next move, so it could be argued this is shorter: 5...a6 6.Qxh5+ Rxh5 mate.

It wouldn't seem possible that a 21th century top chess program could miss a mate in 2 in lengthy analysis, but Pascal Losekoot showed me a funny glitch in Fritz 6. He discovered it when he fed Fritz6 the historical Stamma position on the left - historical because it is the first known minor promotion, from Stamma's 1737 booklet Essay sur le jeu des échecs.
White can mate in 6 with 1.g6+ Kh8 2.g7+ Kh7 3.Nf8+ Rxf8+ 4.gxf8N+ Kh8 5.Bc3+ and mate next move. Fritz6

doesn't see that; at move 3, it hesitates between Bxf2 and g8Q+, seeing both will lose - only after 3...Rxf8+, it discovers the mate, now in 3.
Wondering what exactly was the decisive factor in this glitch, Losekoot found that in analysis, Fritz6 does not see a capturing Knight promotion when it answers a check. He constructed the test position on the right (slightly varied by me.) In the diagram, Fritz6 plays Nxe4, barely drawing. But with the King on a1, it plays 1.Nc8+ Rxc8 2.bxc8N mate.
This must somehow have been introduced in the upgrade; Fritz 5 doesn't have this. PS: And neither, as ChessBase informs me, does Fritz 7.

In the December issue of the British Chess Magazine, John Roycroft gives the position on the left, taken from a chess column by J.A. Porterfield Rynd in the Dublin Herald of 19 October 1895. It is from a game Porterfield Rynd vs. Yates, which had a brilliant finish: 37...Nd3 38.Qxf3 Qb3 39.Qxd3 Qxd3 40.Bf7 Qxc2 41.Ka2 f4 42.gxf4 Qc4+ 43.b3+ Qxb3+ 44.Bxb3 mate.

    The position on the right is a famous endgame study by H. Cordes - White to play and win. The solution is: 1.Bc7 Qe1+ 2.Kh2 Qxf2 3.Bd6 Qf4+ 4.g3+ Qxg3+ 5.Bxg3 mate. This study was first published in 1895 in the Rigaer Tageblatt, where it won second prize. This was mentioned in the July 1895 BCM.
    The position is an exact, be it mirrored, copy of the position after 39...Qxd3 in the Rynd vs. Yates game. Cordes being first, this means that Porterfield Rynd was a plagiarist, and that my great scoop The messenger - the Saavedra myth exposed where I related how Rynd had anticipated the famous Saavedra study in a simul game, was in fact a canard.
    I'll have to rewrite that piece.

Schachwoche showed a new case of the ultimate blunder: resigning in a won position. Thanks to René Olthof for sending me the whole game, which was played in the first Albanian grandmaster tournament.

Mitkov (2413) - Vl. Georgiev (2564), Dürres 2001
1.c4 c5 2.Nc3 Nc6 3.Nf3 g6 4.e3 Bg7 5.d4 d6 6.b3 Bg4 7.Be2 Qa5 8.Bd2 Nh6 9.Nb5 Qd8 10.d5 Nb8 11.Bc3 0-0 12.Bxg7 Kxg7 13.0-0 Nd7 14.Qd2 a6 15.Nc3 Qa5 16.h3 Bxf3 17.Bxf3 b5 18.Be2 bxc4 19.Bxc4 Nb6 20.Rab1 Rfb8 21.Rfc1 Nxc4 22.bxc4 Ng8 23.Rb3 Nf6 24.f3 h5 25.e4 Rb4 26.Nb1 Qb6 27.Qb2 Rb8 28.Nd2 a5 29.Qc2 h4 30.Kh2 Nh5 31.Rcb1 a4 32.Rxb4 cxb4 33.Qxa4 Qf2 34.Qc2 Nf4 35.Qb2+ f6 36.Nf1 Ne2 37.Nd2 Rc8 38.Rf1 Qg3+ 39.Kh1 Nf4 40.Nb3 Rxc4 41.Qf2 Rc3 42.Nd4 Rc1 (see diagram) and White resigned. I'd like to know how they found out what he could have played instead.
For more examples, see The ultimate blunder, elsewhere on this site.

On my records page, I have a game Zoran Djordjevic - Milorad Kovacevic, Bela Crkva 1984, which went: 1.d4 Nf6 2.Bg5 c6 3.e3 Qa5+ (see diagram) and White resigned. This shortest decisive game in a serious tournament was copied in Vassallo (2454) - Gamundi (2457), Spanish team championship, Salamanca 1998, a tournament in which, as Antonio Torrecillas informs me, the likes of Shirov, Lautier, Speelman and Illescas were also playing. Both players, he adds, are "strong IM's".

PS 22 November: Unbelievably, as David Paulowich pointed out, the "Djordjevic Bishop sacrifice" also happened in (quite an interesting) game where White managed to draw with it:
Mory (2105) - Hareux (2195), Open French ch, Besançon 1999
1.d4 Nf6 2.Bg5 c6 3.e3 Qa5+ 4.Qd2 Qxg5 5.h4 Qg6 6.h5 Nxh5 7.Nf3 d6 8.Nh4 Qe6 9.Bd3 g6 10.Nc3 b6 11.O-O-O Ba6 12.d5 Qd7 13.dxc6 Qc8 14.c7 Qxc7 15.Bxa6 Nxa6 16.Qd4 O-O-O 17.Nd5 Qd7 18.Qxh8 Bg7 19.Qxh7 Rh8 20.Nxe7+ Qxe7 21.Nxg6 Qe4 22.Ne7+ Qxe7 23.Qf5+ Kb8 24.Rxh5 Rxh5 25.Qxh5 Qf6 26.Qb5 Nc7 27.Qb4 Qxf2 28.Qxd6 and a draw was agreed.

Checking this further, I found four other occurrences of the moves 1.d4 Nf6 2.Bg5 c6 3.e3 Qa5+ where White had played on; but in Aracil - Fernandez, Cajas open 1989; Wagner - Hartleif, Wurmans 1995; Fernandez - Crespo, Infiesto open 1996 and Diaz - Vega, Candas 1997, Black had always won.

PS 10 December: James Plaskett mentions having witnessed a game in a lower section of the 1984 Troon (Scotland) GM-tournament, where 1.d4 Nf6 2.Bg5 c6 3.e3 was also played, but where Black had missed Qa5+.

Thanks to Clifford Stern for posting on rec.games.chess.misc the maximal distance win in the Q+R vs. Q+R endgame, based on Ken Thompson's endgame databases.

In this position, Black is to move, and he's lost. But it takes a while.
1...Ke5 2.Qh5+ Kd4 3.Qh8+ Kc4 4.Qg8+ Kc5 5.Qf8+ Kb5 6.Rb2+ Ka4 7.Ra2+ Kb5 8.Qf5+ Kb6 9.Rb2+ Rb4 10.Qg6+ Ka5 11.Ra2+ Kb5 12.Qf5+ Kc4 13.Rc2+ Kd4 14.Qd7+ Ke3 15.Rc7 Re4 16.Qh3+ Kd2 17.Qc3+ Ke2 18.Qb3 Re3 19.Rc2+ Kf3 20.Qf7+ Kg4 21.Qg7+ Kf4 22.Qf6+ Kg4 23.Rc5 Kh3 24.Rh5+ Kg2 25.Qc6+ Kf2 26.Rh2+ Kg3 27.Qd6+ Kg4 28.Kd7 Re2 29.Qg6+ Kf4 30.Qh6+ Kf3 31.Qh5+ Ke3 32.Qe5+ Kd3 33.Qd5+ Kc2 34.Qc6+ Kd2 35.Rh3 Re7+ 36.Kd8 Re3 37.Qd5+ Kc2 38.Qf5+ Kc3 39.Qc5+ Kd3 40.Qb5+ Kd2 41.Qa5+ Kd1 42.Qa1+ Kc2 43.Rh2+ Kd3 44.Qa6+ Ke4 45.Qc4+ Kf3 46.Qd5+ Kg3 47.Rh8 Re4 48.Qg8+ Kf4 49.Qf7+ Kg3 50.Qg6+ Kf4 51.Qd6+ Kf3 52.Rh5 Kf2 53.Qf6+ Ke2 54.Rh2+ Kd3 55.Qa6+ Kd4 56.Qd6+ Kc3 57.Qc5+ Kd3 58.Qb5+ Kc3 59.Rh3+ Re3 60.Kc8 Kc2 61.Rh2+ Kc3 62.Qb2+ Kc4 63.Qa2+ Kd4 64.Qa4+ Kd3 65.Qc2+ Kd4 66.Rh4+ Re4 67.Qa4+ Ke5 68.Qe8+ Kd4 69.Qd7+ Kc4 70.Qc6+ Kd4 71.Rh8 Qe2 72.Kb8 Qg4 73.Ka8 Kd3 74.Qd6+ Ke3 75.Rh3+ Kf2 76.Qd2+ Kg1 77.Qc1+ Kf2 78.Rh2+ Kg3 79.Qg1+ Kf4 80.Qf2+ Ke5 81.Qc5+ Kf6 82.Qf8+ Ke6 83.Rh6+ Kd5 84.Qd8+ Kc5 85.Qa5+ Kd4 86.Rd6+ Kc4 87.Rc6+ Kd3 88.Rc3+ Ke2 89.Rc2+ Kf3 90.Qc3+ Re3 91.Qf6+ Qf4 92.Rf2+ Kxf2 93.Qxf4+ And the conversion is a fact, but the win against perfect play would still be very difficult, even for grandmasters. 93...Rf3 94.Qd4+ Re3 95.Kb7 Kf3 96.Kc6 Re4 97.Qh8 Rf4 98.Qh1+ Ke2 99.Kd5 Ra4 100.Qh5+ Kd2 101.Qf3 Ra8 102.Qf2+ Kd1 103.Qf6 Ke1 104.Qe5+ Kf1 105.Ke4 Ra4+ 106.Ke3 Ra3+ 107.Kd2 Ra2+ 108.Kd1 Rf2 109.Qe4 Kg1 110.Ke1 Rf1+ 111.Ke2 Rf2+ 112.Ke3 Rf1 113.Qd5 Rf2 114.Qd4 Rg2 115.Kf3+ Kh1 116.Qh8+ Kg1 117.Qh4 Rg8 118.Qd4+ Kh2 119.Qe5+ Kg1 120.Qa1+ Kh2 121.Qa2+ Kg1 122.Qxg8+ Kf1 123.Qg5 Ke1 124.Qc1 mate.

Mostly, a series of incomprehensible moves, chess at particle level. I especially like the silence of moves 15, 60, 72 and 73. According to Lewis Stiller in his thesis, this endgame is won in 83 % of all possible positions, which is surprising, even if it is obvious that the first move is of great value with so much unrestrained force on the board. The moves come from Ken Thompson's endgame databases. Also, see item 60 in this Diary, and Stiller's Monsters, elsewhere on this site.

Mate in 8
W. Shinkman
St. Louis Globe Democrat 1887

Mate in 6
ms. 791, Casanatense Library, Rome
15th Century

Some time ago (see The Kuwait immortal), I wrote about the Shinkman classic in the diagram on the left. The solution is: 1.O-O-O Kxa7 2.Rd8 Kxa6 3.Rd7 Kxa5 4.Rd6 Kxa4 5.Rd5 Kxa3 6.Rd4 Kxa2 7.Rd3 Ka1 8.Ra3 mate After 1.Kd2, it is also mate in 8.

Later, Harold van der Heijden and Bader Al-Hajiri pointed me to a book 100 Classics of the Chessboard by Dickins and Ebert (Cadogan, 1995) which gives an amazing 15th century anticipation of Shinkman's problem. See the diagram on the right. (Note: the piece on h2 is the old Arabian Alfil.) The Alfil is a 2,2 leaper, and the Ah2 therefore only guards f4. The solution is: 1.Nb3 Kxe7 2.Ra8 Kxe6 3.Ra7 Kxe5 4.Ra6 Kxe4 5.Ra5 Kxe3 6.Re5 mate.

PS: As Rik van der Heiden and Bader al-Hajiri remarked, White can also play 1.Ra8+ and 2.Nb3. Van der Heiden also points to 1.Nb3 Kxe7 2.Nd4 and mate in time. But in those days, dual solutions didn't matter. Also, obviously, the Alfil could never have reached h2 from its starting square c1. Al-Hajiri suggests replacing it by a Pg3.

In my story Master Jacobson, Jacobson plays a correspondence game for which it is easy to find the model in reality. He also tries to compose a problem, a nineteenmover he calls "The Bottomless Pit" - a very difficult construction which must "couple the brute force of a harvesting machine with the precision of a ladies’ watch." He imagines himself demonstrating it: "It’s not correct yet, but this is the idea. Mate in nineteen. The mate itself is easy, but Black is going to delay it. He can interpose a pawn, and this one too and one more; a knight, and this knight too. Now a rook, and another one, the queen, and even the bishop: nine sacrifices, all on the same square. And now, finally, White can mate, you see? Beautiful, isn’t it?"
A reader asked me if this problem is based on real life too, and it is.

Mate in 19
P. Drumare & R. Le Pontois
Thèmes-64, 1962 1.Bd3 White must always threaten a discovered mate by the Nd2, and Black can only defend by putting something on e4. 1...e4 2.Bxe4 e5 3.Bd3 e4 4.Bxe4 e5 5.Bd3 e4 6.Bxe4 Ngf6 7.Bd3 Ne4 8.Bxe4 Nf6 9.Bd3 Ne4 10.Bxe4 Re8 11.Bd3 Re4 12.Bxe4 Re8 13.Bd3 Re4 14.Bxe4 Qc6 15.Bd3 Qe4 16.Bxe4 Bb7 17.Bd3 Be4 18.Bxe4 d3 19.Bxd3 mate

It's nowhere near the record (see my records page) but I still liked the way Kasparov left his opponent's Bishop en prise for 11 moves in this very recent game:

Kasparov - Movsesyan, Eurotel Kasparov Trophy, Prague 18 October

1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 a6 5.Nc3 Qc7 6.Bd3 Nc6 7.Nxc6 bxc6 8.O-O Nf6 9.Qe2 Bd6 10.f4 e5 11.Kh1 h6 12.Bd2 O-O 13.f5 Re8 14.Bc4 a5 15.Rf3 Bb4 16.Rg3 Kf8 17.Qe3 Ng8 18.f6 gxf6 19.Rf1 Bxc3 (see diagram) 20.Rxg8+ Kxg8 21.Qg3+ Kf8 22.Rxf6 d5 23.Bxh6+ Ke7 24.Rxf7+ Kd6 25.Qg6+ Be6 26.Rxc7 Kxc7 27.exd5 cxd5 28.Bb5 Reb8 29.a4 Bd7 30.bxc3 Bxb5 31.axb5 Rxb5 32.h4 a4 33.Bg7 d4 34.Qf7+ Kb6 35.Qe6+ Kb7 36.Bxe5 Rxe5 37.Qxe5 a3 38.Qd5+ Kb8 39.cxd4 a2 40.Qb3+ Kc7 41.Qc3+ Kd7 42.Qa1 Kd6 43.c4 and Black resigned.

The game was played on board 1 of a 2-round clock simul that Kasparov plays against the Czech team. For the first 50 moves, Kasparov has 3 hours; the Czechs have 2,5. Both have 30 minutes for the rest.
The winner gets / winners get $ 40,000. The return match is today. Kasparov leads 2½-1½. (PS 21 October: He won the second round 3-1.)

Famous endgame studies get busted all the time, but once in a while, it happens to a famous game. Here is one that I already showed in this Diary, as item 47.

White Rook - Tal, correspondence 1968/69
White were the readers of the Pionerskaya Pravda; (a Soviet youth magazine), calling themselves "White Rook" for the occasion.
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 Bc5 5.Nxf7 Bxf2+ 6.Kxf2 Nxe4+ 7.Kg1 Qh4 8.g3 Nxg3 9.hxg3 Qxg3+ 10.Kf1 Rf8 11.Qh5 d5 12.Bxd5 Nd4 13.Qh2 Qg4 14.Qxe5+ Be6 15.Bxe6 Qf3+ 16.Kg1 Ne2+ 17.Kh2 Qf2+ 18.Kh3 Qf3+ 19.Kh4 Qf2+ 20.Kh5 Rxf7 21.Bxf7+ Kxf7 22.Rh2 Qf3+ 23.Kh4 g5+ 24.Qxg5 Rg8 25.Qh5+ Qxh5+ 26.Kxh5 (see diagram) The game now ended in a draw quickly with 26...Ng3+ 27.Kh6 Nf5+ 28.Kxh7 Rg7+ and a draw was agreed.
A few months later however, one of the pioneers,

Vadim Brodsky, discovered that Tal had missed an easy win with (from the diagram) 26...Nf4+! 27.Kh6 (27.Kh4 h5 and Rg4 mate) Rg6+ 28.Kxh7 Rg7+ 29.Kh6 (29.Kh8 Ng6 mate) Kg8! and mate with Rg6.
    Recently, this game was copied by two, apparently curiosity-loving, Hungarians for a prearranged draw, securing a first place for Black. In Horvath - Bokros, "Young Masters" tournament, Groningen 1999, they inserted the repetition 19.Kh2 Qf2+ 20.Kh3 Qf3+ and chose a different draw with, from the diagram: 28...Ng3+ 29.Kh4 Nf5+ 30.Kh5 Ng3+ (28...h6 29.d4) 31.Kh4 draw, not only copying Tal's game, but also his blunder (28...)Ng3+.
    Some time ago however, in his "Rotterdam" club magazine, Maarten de Zeeuw showed something that had gone undetected for over 30 years: Tal had missed mate twice in two moves. After his blunder 26...Ng3+, the White Rook's answer 27.Kh6 was also a blunder (one that Bokros avoided) - Black could have won with 27...Rg6+! 27.Kxh7 Ne4! and mate in a few moves.
    So in the White Rook - Tal sequence 26...Ng3+?? 27.Kh6?? Nf5+??, a quick mate was overlooked and half a point blundered on each move - a little bit much for a correspondence game.

From a Bombo Radyo (Philippines) interview with Bobby Fischer, given a few hours after the attacks on the US.

Pablo Mercado: We would like your commentary on what happened at the World Trade Center just about a few hours ago including that of the White House and the Pentagon too, it’s evening right here.
Fischer: This is all wonderful news. It’s times for the fucking US to get their heads kicked in. It’s time to finish off the US once and for all. (...) This just shows you that what goes around comes around, even for the United States (...)
Mercado: You’re saying you're happy at what happened?
Fischer: Yes, I applaud the act. (...) The US and Israel have been slaughtering the Palestinians, just slaughtering them for years, robbing them and slaughtering them, nobody gave a shit. Now it is coming back to the US. Fuck the US. I want to see the US wiped out.
Mercado: Ha, ha, ha.
Fischer: What I’m really hoping for, Pablo, did you ever see that movie "Seven Days In May". That’s a movie about a general who tries to take over the USA. I think it was with Burt Lancaster and Kirk Douglas, it was based on a book. I saw that years ago. I was really for the generals, you know, but in the end the president of the so-called democracy won. But I’m hoping for some kind of a Seven Days In May scenario, where the country will be taken over by the military, to close down all the synagogues, arrest all the Jews, execute hundreds of thousands of Jewish ringleaders, and you know, apologize to the Arabs for the killing (...) for all the Jews over there of that bandit state, you know, Israel.
    I’m hoping for a total new world. Nobody gives a shit about the Japanese. How many hundreds of thousand people did the US kill with the atom bombs? Justifying it with the most ridiculous excuse that it saved millions American soldiers, when Japan would gonna surrender in a few weeks or month or so anyway. Right? The United State is based on lies, is based on theft.
    Look what I have done for the US. Nobody has single-handedly done more for the US them me, I really believe in this. When I won the World Championship in 1972, the United States had an image of you know a football country, baseball country, but nobody thought of it as an intellectual country. I turned all that around single-handedly, right? But I was useful then because it was the cold war, right? But now I’m not useful anymore, you see, the cold war is over and now they want to wipe me out, get everything I have, put me into prison.
    You have to go back to the root of history of the country, look at the history of the country. Get something for nothing. Take and kill. Rob the country, they don’t come in a civilized manner and say we like to marry your women, and so on. No, they take your land and they kill you off. That’s the history of the US. Why did the white man not come to America, like in a civilized manner, preaching freedom of religion, say we like to come here. We like to assimilate, we like to marry your women. But no, we take your land and kill you off, right? Bring over slaves from Africa. That’s the history of the United States. A despicable country, you know. Even as a boy I never had the slightest interest in the history of the US, I knew there was something rotten in Denmark.
(...)
The US just will not do what they have to do. The US has to say we’re sorry, our whole foreign policy has been wrong for the last several hundred years, we are going to pull back all our troops from all over the world, we stop supporting Israel and so on. But they only will say that this cowardly act will be punished.
(...)
Democracy is just a load of bullshit, it is just a cover for the criminal nature of the United States of America. But I’m hoping for the Seven Days In May scenario, where sane people will take over the US, military people. They will imprison the Jews, they will execute several hundred thousand of them, at least. And they will bring home all the troops to the US.
    And ultimately the white man should leave the US, the black man should go back to Africa, the white back to Europe, and the country should be returned to the American Indians who lived there for, who knows how many, ten of thousands of years. They kept the land crystal clean. It was a beautiful country when the white man came. This is the future I would like to see for the so-called United States.
(...)
Death to the US. They are the worst liars and bastard. This is a wonderful day.

Thanks to Peter van der Hoog for this partial transcription of the latest (19th) Philippine radio interview with Fischer. He put it on rec.games.chess.misc, but some news servers seem to have refused it. MP3's for all 19 interviews can be found at this Japanese Fischer site.

Thanks to Joost de Heer for sending me this problem, one of the wittiest I've seen in years.

a) diagram; mate in 2
b) Ng7; mate in 3
c) Bg7; mate in 4
d) Rg7; mate in 5
e) Qg7; mate in 6

a) 1.g8N d2 2.Nh6 mate.
b) 1.Ne6 d2 2.Nd4 Kxf4 3.Nxf3 mate.
c) 1.f7 d2 2.Ld4 Kxf4 3.Lf6+ Nd4 4.Rxd4 mate.
d) 1.Re7 d2 2.Re2 Nd4 3.Rxg2+ Kxf4 4.Rxd4+ Ke5 5.Re2 mate.
e) 1.Qe7 d2 2.Re4 fxe4 3.Qxe4 Nf5 4.Qxf3+ Kh4 5.Qf2+ Ng3 6.Qxg3 mate.