267. 16 November 2004: No-brainers (+ various postscripts to items 265 and 261)

Korolkov's "just-press-the-button" study (see item 265), prompted several readers to mention this well-known little joke.

Mate in 6
V. Röpke
Skakbladet, 1942

A button is not even needed here.

Richard Stanley sent me two records in the "no-brainer" class, as he and the composer, Noam Elkies, call these problems: mates in n where neither player gets around to thinking.

Mate in 7
Noam Elkies
original, 2004

White has choices, but no matter what he chooses, he mates Black at move 7. This is probably the record-nobrainer without promoted pieces.

And with promoted pieces (but in a legal position):

Mate in 10
Noam Elkies
original, 2004

White and Black have choices all the time, but the machine runs its course until it is mate at move 10.

Of course, other than in Röpke's pioneering no-brainer, one could imagine either of the players overstepping the time limit here, because they might not see that their choices make no difference.

Elkies' problems will appear in a book "Chess and Mathematics" that he and Stanley are preparing.


266. 2 November 2004: Fischer was right in 1975

- in a way.

When Kramnik retained his 'classical world title' by drawing a match 7-7 against Leko, there were discussions about the fairness of the 'Tie Clause' favouring him. But I've not seen many serious calculations of how large his odds actually were.

If we use the outcome of Kramnik - Leko for a prediction of that outcome, we have the convenient facts that the players were equally strong, each winning two games, and that White and Black were equally strong too, with two wins each. The chance of a draw being 5/7 and the match consisting of 14 games, that leads to an expectation (calculated for me by Joshua Green) of 60.1 % that Kramnik would retain the title.
    This was confirmed by a little numerical simulation program that I wrote - see below.

Surprisingly, a slightly altered version of this program shows that the much criticized odds Fischer demanded in 1975 for his 'wins-required' match against Karpov (Fischer would retain the title with 9-9; Karpov had to win 10-8), were smaller than the odds Kramnik had against Leko. Again assuming both players had equal chances to win a game, Fischer would have had a 59.2 % chance to keep his title.

Here's a table showing the tie-clause advantage for various match lengths and drawing percentages in a 'points-required' match, as shown by my simulation program. If some numbers do not seem to be completely consistent, that is because I limited my samples to 10,000 matches for each combination. (If the deviation was too great, I ordered Kramnik and Leko to play their match a million times instead of a mere ten thousand.)

24 55.256.056.157.159.163.6
22 55.456.356.757.659.464.0
20 55.956.556.958.259.664.3
18 56.156.657.558.961.166.0
16 56.357.057.860.462.066.7
14 56.957.358.959.962.968.7
12 57.158.159.160.763.370.4
10 57.758.760.061.964.572.4
 8 58.159.261.162.965.274.8
 40%50%60%70% 80%90%

Here's a list of odds from my simulator that Fischer would have had with different numbers of wins to be played for:

10 59.2 %
  9 59.8 %
  8 60.5 %
  7 61.3 %
  6 62.3 %

PS 16 November: Thomas Beuman gave me the exact calculations for 'points-required' matches:

24 55.255.856.457.459.263.6
22 55.456.056.757.859.664.2
20 55.756.257.058.260.165.0
18 56.056.657.458.660.766.0
16 56.457.057.859.261.467.2
14 56.857.458.459.862.268.6
12 57.458.059.060.663.270.3
10 58.058.859.961.664.672.5
 8 58.959.861.063.066.675.4
 40%50%60%70% 80%90%

Both Beuman and Joshua Green calculated the 'wins-required' percentages, finding just one difference with the percentages of my simulations: playing for 10 wins, Fischer would have had a 59.3 % advantage.

Some of the above simulations and calculations have been done earlier by Charles Kalme and William Hyde. See this 1999 posting by Hyde on rec.games.chess.misc.


265. 23 October 2004: Just press the button

A recent(ly found) record on my records page was a series of six consecutive mutual checks in a game Zarrouati - Brauckmann, Toulouse 1990. In a construction, a series of 37 has been achieved by G. Ponzetto, 1993. For discovered checks, the record is 11 by O. Stocchi - see item 125 in this Diary. These constructions just aim at this one effect, but such series sometimes arise in more or less normal studies. More or less, because if the series are really long, they are record attempts, too.

L.M. González
White to play and draw
Humour Tourney, 2004

1.Qf2+ Qxf2+ 2.Bxf2+ Rxf2+ 3.Nef3+ Rbxf3+ 4.Nxf3+ Rxf3+ 5.Kxf3+ Ncd4+ 6.Rdxd4+ Nxd4+ 7.Rxd4+ Be4+ 8.Rxe4+ fxe4+ 9.Kxe4 draw.

Sixteen checks in a row! It takes some lengthy analysis (in an article by Harold van der Heijden in the Dutch studies magazine EBUR of September 2004) to prove that all of them are the best move, but it seems they are.

The study is a contender in Hans Böhm's Humour Tourney which still accepts studies until 31 December 2004. Original studies can be sent to HUMOUR Tourney, Harold van der Heijden, Michel de Klerkstraat 28, 7425 DG Deventer, THE NETHERLANDS.

I must say I'm surprised (but see the PS) that entries in a current tourney can be published, but Harold is the endgames studies expert, and EBUR's editor-in-chief, and he should know. As a jury member in that tourney (with Hans Böhm and Jan Timman), I would have thought I'm not supposed to know who composed what, but as a chess-loving citizen, I'm free to comment on what I see in public sources. Sixteen is quite a feat, but it's a rather cumbersome thing, and it doesn't tickle my sense of humour. Here is one with only seven consecutive checks, that does.

V. Korolkov, 1940
White to play and win

1.Nd8+ Re6+ 2.f6+ Ne5+ 3.Bxe3+ Nbd3+ 4.b4 mate.

All the moves in the solution are discovered checks, but the really funny thing is that apart from White's first move (which is practically forced), both sides always play their only legal move. The perfect chess riddle - just press the button, and sit back.

PS 24 October: I now see, maybe even remember, that these publications in EBUR were intended all along. It's an informal tourney.

PS 16 November: Noam Elkies sent me a game with 7 consecutive checks, be it only in the (published) analysis of an actual game.

Kavalek - Zuckerman, Student's Olympiad, Krakow 1964
1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 Qb6 8.Qd2 Qxb2 9.Rb1 Qa3 10.Bxf6 gxf6 11.Be2 Nc6 12.Nxc6 bxc6 13.O-O Bg7 14.Rfd1 O-O 15.Rb3 Qc5+ 16.Kh1 d5 17.f5 Ra7 18.Na4 Qd6 19.c4 Qe5 20.fxe6 dxe4 21.Qe3 Re7 22.exf7+ Rexf7 23.c5 f5 24.Rb4 a5 25.Rbd4 Be6 26.Rd6 f4 27.Qf2 Bd5 28.Nb6 f3 29.gxf3 exf3 30.Bf1 Qe2 31.Rd2 (see diagram)
Black now won comfortably with 31...Qxf2 but in his 1970 (Hebrew) book about this student's Olympiad, Shaul Hon gave this possibility: 31...Bd4 32.Qxd4 and now: 32...Qxf1+ 33.Qg1+ Qg2+ 34.Qxg2+ fxg2+ 35.Rxg2+ Bxg2+ 36.Kxg2 Rg7+ 37.Kh3 Rf3+ 38.Kh4 Rf2 and Black wins. Not a very relevant variation however, as 33...Qxg1+ 34 Kxg1 f2+ is an easier win, and so is 32...f2+ and mate next move. The game ended: 32.Rxf2 Bd4 33.Rc2 Be4 34.Rxd4 Bxc2 35.Bc4 f2 36.Rf4 Kh8 and White resigned.



264. 14 October 2004: Hensel and greedle

When the 11th match game between Kramnik and Leko was drawn in 17 moves, and there was nothing to say about the chess, ChessBase interviewed Carsten Hensel, who is the personal manager of both players. It was a remarkable interview. Here are some questions.

Q: Do Kramnik and Leko feel any pressure to play more fighting chess?
Q: Have Kramnik and Leko noticed the fans' disappointment about their short draws?
Q: Could it be that Kramnik and Leko feel intimidated by the title they are playing for?
Q: Do Kramnik and Leko feel an obligation to be creative in their games?
Q: How do Kramnik and Leko justify playing only two moves beyond their home preparation?
Q: Do Kramnik and Leko like chess?
Q: Is it difficult for Kramnik and Leko to look their sponsors in the eye after a 17-move draw?
Q: Are Kramnik and Leko afraid they will not be able to keep a straight face when they collect their fees?
Q: Are Kramnik and Leko concerned that they are turning away this new sponsor from chess?
Q: Would Kramnik and Leko admit to being cowards?
Q: Do the Dannemann representatives in Brissago know so little about chess that they don't know they're being ripped off by Kramnik and Leko?

Not one of these questions was asked. Hensel is a strong pingpong player, and he likes his wife's potato salad.


263. 12 October 2004: RIP




262. 11 October 2004: Plaskett on Major Ingram

Saturday's Daily Mail had a very interesting article by GM James Plaskett on the famous alleged coughing scam of Major Ingram in the British Who wants to be a Millionaire quiz. It is in the form of an interview with Plaskett; Plaskett's own piece can be found on the Portia website which is devoted to perceived miscarriages of justice.


261. 10 October 2004: Two small mysteries

In his simul tour in 1964, Bobby Fischer lost this well-known game.

Fischer - Burger, simul San Francisco, 13 April 1964
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 Nd4 6.c3 b5 7.Bf1 Nxd5 8.cxd4 Qxg5 9.Bxb5+ Kd8 10.Qf3 Bb7 11.O-O exd4 (see diagram) 12.Qxf7?? (after d3, Re1 or even Bd3 White's position is perfectly playable) 12...Nf6 and White resigned. Everything hangs; his King, his Bishop and his Queen (Bd5).

A typical simul disaster, but a reader from Brazil, Romario, pointed out that this game was duplicated between grandmasters in a strong tournament: Bronstein - Lengyel, Sarajevo 1971. In my large database however, which is a compilation of several games collections, I found two copies of this Bronstein - Lengyel, both with the same moves, but one tagged "0-1", and one "1/2". Two large online databases, ChessLab and ChessBase's Chesslive, as well as ChessBase's "Mega 2000" all give this game as 0-1. On the other hand, there is no such game in the early Informators, and Informator 11 has a tournament table of Sarajevo 1971, where Bronstein - Lengyel is a draw.

Who knows more?

PS 2 November: Jovan Petronic reminds me that there is another version of this game, ending with 11...e4 12.Qxe4 Bd6 13.d3 Bxh2+ 14.Kxh2 Nf4 (0-1) In 1992, during the rematch with Spassky, Jovan Petronic had a chance to ask Fischer about this game. Fischer said he did not remember playing it, but that he knew Burger, "a very friendly fat man, who liked drinking wine."

PS1, 16 November: Rik van der Heiden points out that according to Donaldson's A Legend On the Road, the 14-move game is not a game, but a variation in:
Fischer - NN, clock simul Montreal, 24 Februray 1964
1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 d5 5.exd5 Nd4 6.c3 b5 7.Bf1 Nxd5 8.cxd4 Qxg5 9.Bxb5+ Kd8 10.Qf3 e4 11.Qxe4 Bd6 12.O-O Bb7 13.d3 Nf4? Here, Bxh2+ wins 14.Bxf4 Qxb5 15.d5 Qxb2 16.Bxd6 cxd6 17.Re1 Qf6 18.Nc3 Rc8 19.Qb4 Re8 20.Qa5+ Kd7 21.Qa4+ and Black resigned.

PS2, 16 November: Thanks to Andy Ansel for unearthing the following game, from the original bulletins.
Bronstein - Lengyel, Sarajevo 1971
1.e4 e5 2.f4 exf4 3.Bc4 d5 4.exd5 Nf6 5.d4 Nxd5 6.Nf3 Be7 7.0-0 Be6 8.Bxd5 Bxd5 9.Bxf4 0-0 10.Nc3 c5 11.Nxd5 Qxd5 12.dxc5 Qxc5+ 13.Kh1 Nc6 14.c3 Rad8 15.Qe2 Bd6 draw agreed.
So it was just a normal draw - the mystery is how the Fischer - Burger moves got associated with that game.


Recently, I stumbled upon another remarkable simul game.

Reshevsky - NN, simultaneous Tel Aviv 1950
1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 c5 5.Nge2 d5 6.Bd2 Qa5 7.a3 (see diagram)
7...Nc6!? Could that really be possible? 8.axb4 After cxd5 or dxc5 White retains a small advantage. 8...Nxb4 Threatening both Nd3 mate and Qxa1. 9.Qa4+ Qxa4 10.Rxa4 Nd3+ 11.Kd1 Nxf2+ 12.Ke1 Nd3+ and a draw was agreed. Black should have played for a win with 12...Nxh1.

But when I checked whether that funny 7...Nc6 had occurred in other games, I only found this:

Reshevsky - Z. Margolits, simultaneous Haifa 1958
1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 c5 5.Nge2 Qa5 6.Bd2 d5 7.a3 Nc6 8.axb4 Nxb4 9.Rxa5 Nd3 mate.

Which of the two versions is true? Could Reshevsky really have overlooked that mate? The first version is more believable, but Reshevsky did not play a tournament in Tel Aviv in 1950 and he did play in Haifa in 1958. ChessLab has Reshevsky - Z. Margolits, Haifa 1958 and Reshevsky - Margolit, Haifa 1968, both ending with 9...Nd3 mate; ChessLive has the Haifa 1958 version; the online database of New in Chess has 'Haifa 1968', 'R.B. Margolit' and 9...Nd3 mate, and Mega 2000 doesn't have this game at all.

Which version was really played? And who was Z. Margolits / R.B. Margolit?

PS, 16 November: Norm Antokol remembers seeing the game in Chess Life in 1958; it ended with 9...Nd3 mate.


© Tim Krabbé, 2004

1-20 | 21-40 | 41-60 | 61-80 | 81-100 | 101-120 | 121-140 | 141-160 | 161-180 | 181-200 | 201-220 | 221-240 | 241-260
Top | Main chess page | Main page | Contact me |