397. 9 February 2016: Harry Goldsteen´s furthest position

Ten years ago, on the Dutch side of this website, I wondered how many moves a position can be removed from the initial position - and what that furthest position would look like.
    Considering Sam Loyd´s famous games where only two Kings remain after 17 moves, I was sure it wouldn´t look like a position from a game; if both sides cooperate, it must be possible to reach any ´normal´ position within 25 or 30 moves.
    This longest shortest path to get back to the pristine state made me think of Rubik´s Cube which, it was believed at the time, could never be further away than 26 turns - even if no human could solve it in anywhere near that number. In 2010 it was proved that the Cube is always theoretically solvable in 20 turns or less.
    What would this number be in chess? I couldn´t find an answer, until I thought of asking the Kuwait Wizard, Bader al-Hajiri. He showed me a then 59 years old joint composition by two retrograde grandmasters.

L. Ceriani & K. Fabel
The Furthest Position, 183 moves
Am Rande des Schachbretts, 1947

The shortest path to this position, or from this position back to the initial one, is 183 moves; they can be played over on the left. ´A record that will never be broken,´ according to Al-Hajiri.

As it turns out only now however, another legendary great in the world of retrograde analysis, my old coffeehouse chess friend Harry Goldsteen, had, upon seeing Ceriani´s and Fabel´s position on my site, already added one move to their record.

Harry Goldsteen, after L. Ceriani & K. Fabel
The Furthest Position, 184 moves
Original, 2006

This position takes 184 moves to untangle; the game can be played over, forward or backwards, on the left.

But Goldsteen never published his new record.

Two years later, he added one more move, raising the record for this task to 185 moves. Again, he did not publish it. But now he does.

Harry Goldsteen, after L. Ceriani & K. Fabel
The Furthest Position, 185 moves
Original, 2008

The game can be played over on the left; comments are based on those by Harry Goldsteen and Aran Köhler.

1.a4 g5 2.Ra3 g4 3.h3 g3 4.Rh2 gxh2 5.Nf3 h1R 6.d3 d5 7.Kd2 e5 8.Nh2 a5 9.Ke3 d4+ 10.Kf3 e4+ 11.Kg3 e3 12.Qd2 exd2 13.Rb3 d1B 14.Bf4 b5 15.Nd2 bxa4 16.Rb7 Be6 17.Ne4 Ba2 18.Nd6+ Kd7 19.Nc8 Kc6 20.Nb6 Kb5 21.Be5 Bb1 22.Bf4 a3 23.Be5 a2 24.Bf4 a1B 25.Be5 As in the 183- and 184-move games, White makes many waiting moves. Black ´leads´, as Goldsteen puts it - he must play optimally, although some changes in move-order are possible for him, too. 25...Kb4 26.b3 Kc3 27.Nd5+ Kd2 28.Nb6 a4 29.Nc8 a3 30.Nb6 a2 31.Nc8 Bb2 32.Nb6 a1B 33. Nc8 Bc1 34.Nb6 Bc3 35.Nc8 Bcb4 36.Nb6 c5 37.Na4 c4 38.Nb2 c3 39.Kf3 cxb2 40.c3 Bdc2 41.Ke4 Ne7 42.Bc7 Nd5 43.Be5 Ne3 44.Bc7 Nd1 45.e3 h5 46.Be2 Qh4+ 47.Kd5 Na6 48.Bg4 hxg4 49.Kc4 g3 50.Rb5 gxh2 51.Rb7 Re1 52.Rb5 h1R 53.Rb7 Nc5 54.Bh2 Ne4 55.Rb5 Ng3 56.Rb7 Rhf1 57.Rb5 Nh1 58.g3 Now White has created the smallest possible box in which Black´s promoted Rooks and Bishops manoeuver like tar through a sieve. 58...Ke2 59.Rb7 Kf3 The black King makes room for the white Rook so it can go to g2. In the next phase, the black Rooks must leave the box, otherwise the white Rook cannot pass. 60.Rb5 Bfc5 61.Rb7 Re2 62.Rb5 Ba2 63.Rb7 Bcb1 64.Rb5 Rc2 65.Rb7 Re1 66.Rb5 Ree2 67.Rb7 Bd2 68.Rb5 Be1 69.Rb7 Rc1 70.Rb5 Bc2 71.Rb7 Ra1 72.Rb5 Bcb1 73. Rb7 Rc2 74.Rb5 Rc1 75.Rb7 Bc2 76.Rb5 Bab1 77.Rb7 R1a7 78.Rb5 Ba2 79.Kd5 Ra1 80.Kc4 Bab1 81.Kd5 R1a6 Now the white Rook can enter via a5. 82.Ra5 f5 83.Ra1 Ba2 84.Rc1 Bcb1 85.Rc2 Qh5 Or Qg4; the Queen must be able to reach e2 in one move. 86.Re2 Bd2 87.Re1 Bc2 88.Rg1 Bab1 89.Rg2 The Rook has arrived. 89...Ra1 90.Kc4 Ba2 91.Kb5 Rc1 92.Kc4 Bab1 93.Kb5 Ra1 94.Kc4 Ba2 95.Kb5 Be1 96.Kc4 Bcb1 97.Kb5 Rc2 98.Kc4 Re2 99.Kb5 Bc2 100.Kc4 Rc1 101.Kb5 Bab1 102.Kc4 Ra1 103.Kb5 Rha8 104.Kc6 Bd2 105.Kb5 Re1 106.Kc4 Rg1 107.Kb5 Ba2 108.Kc4 Bcb1 109.Kb5 Rc2 110.Kc4 Bc1 111.Kb5 Re2 112.Kc4 Ree1 113.Kb5 Ref1 114.Kc4 Bc2 115.Kb5 Bd2 116.Kc4 Rc1 117.Kb5 Bab1 118. Kc4 Ra1 119.Kb5 Ba2 120.Kc4 Be1 121.Kb5 Bcb1 122.Kc4 Rc2 123.Kb5 Ke2 The black King must enter the box now; his first goal is a3. 124.Kc4 Kd2 125.Kb5 Kc1 126.Kc4 Re2 127.Kb5 Kd2 128.Kc4 Bc2 129.Kb5 Rc1 130.Kc4 Bcb1 131.Kb5 Rc2 132.Kc4 f4 133.Kb5 Kc1 134.Kc4 Rcd2 135.Kb5 Bc2 136.Kc4 Kb1 137.Kb5 Ka1 138.Kc4 Bab1 139.Kb5 Ka2 140.Kc4 Ka3 141.Kb5 Ba2 142.Kc4 Bcb1 143.Kb5 Rc2 144.Kc4 Rc1 145.Kb5 Bc2 146.Kc4 Ra1 147.Kb5 Bcb1 148.Kc4 Rc2 149.Kb5 Only now can the black Queen treacle into the box. 149...Qe2 150.Kc6 Qd2 151.Kd5 Qc1 152.Kc6 Re2 153.Kd7 Qd2 154.Kc6 Bc2 155.Kb5 Rc1 156.Kc6 Bcb1 157.Kd7 Rc2 158.Kc6 Qc1 159.Kd7 Rcd2 160.Kc6 Bc2 161.Kd7 Qa1 162.Kc6 Bcb1 163.Kb5 Rc2 164.Kc6 Bd2 165.Kd7 Bc1 166.Kc6 Ree1 167.Kd7 Rce2 168.Kc6 Bd2 169.Kd7 Bc2 170.Kc6 Qc1 171.Kd7 Bab1 172.Kc6 Ka2 173.Kd7 Ka1 174.Kc6 Ba2 175.Kd7 Bcb1 176.Ke8 Qc2 177.Kd7 Bc1 178.Ke8 Rd2 179.Kd7 Ree2 180.Ke8 Rfe1 181.Kd7 Rgf1 182.Bg1 f3 183.Rh2 Ba3 184.Ke8 Bcb4 185.Kd8 Ba5+ and the diagram has been reached.

Goldsteen does not have proof that 185 is unsurpassable - in fact, he thinks it is likely that it can be surpassed.

More about Goldsteen on my site:
The horse concoction or, in Dutch: Het paardenweefsel

Harry Goldsteen (left), Amsterdam, 18 December 2015. (Photo Ronit Palache)

With many thanks to Harry Goldsteen and Aran Köhler.

396. 14 September 2014: New discovery in the Rxb2 mystery: who played it, Sanz, Wojchiechowski - or Karlin? (+ PS 30 September 2015)

A recent discovery by the Czech chess historian and writer Jan Kalendovský adds an intriguing note to the mystery of Rxb2 - the famous brilliancy from a game Ortueta - Sanz, Madrid 1933, that later turned out to have been anticipated, or hoaxed, in a game Tylkowski - Wojciechowski, Poznan 1931.

On this site, I already wrote extensively about this strange case; see here and here (#327).

A brief recapitulation:

Ortueta - Sanz, Madrid 1933
Black to play

31...Rxb2 32.Nxb2 c3 33.Rxb6 c4 34.Rb4 a5 35.Nxc4 c2 and White resigned

Tylkowski - Wojciechowski, Poznan 1931
Black to play

30...Rxb2 31.Nxb2 c3 32.Rxb6 c4 33.Rb4 a5 34.Nxc4 c2 and White soon resigned.

The position on the left, first published in August 1934 in the national chess magazines of Belgium and Romania, was already famous when, beginning with a publication in the Polish magazine Szachy in 1952, the alleged predecessor on the right also started to gain fame - as an incredible coincidence or a hoax, nobody really knew.

Now, Kalendovský found an earlier publication of Rxb2 - in the 14 April 1934 Tagesbote, a German-language newspaper in Brünn, Czechoslovakia (now Brno in the Czech Republic.) And it was not Ortueta - Sanz or Tylkowski - Wojciechowski - but NN - Karlin.

NN - Karlin, Madrid 1933
Black to play

1...Rxb2 2.Nxb2 c3 3.Rxb6 c4 4.Rb4 a5 ¨and White had to resign, even if he has two extra pieces on the board. Promotion to Queen cannot be prevented anymore.¨ (But the win after 5.Nxc4 c2 6.Nxa5 c1Q+ 7.Kh2 Qc5 8.Ra4 Qb5 9.Ra3 Qb4 10.Rd3 Qxa5 11.Rd5 is not trivial, and one move earlier, 4.Nxc4! c2 5.Rc6 c1Q+ etc. leaves Black with an only slightly better position.)

About the circumstances of this game, Tagesbote wrote:

The Swedish master Karlin, who recently played in a small tournament with four participants in Brünn, has made a succesful chess journey in Spain last year. At this occasion, he played a series of serious games against a strong amateur in Madrid. In one game, he managed to execute a splendid problemlike finish (...)

Kalendovský adds:

All I can do is reconstruct how this position from Madrid 1933 surfaced in the Brünner Tagesbote in April 1934. In the months February to May 1934, master Karlin made a tour of Czechoslovakia and played some small tournaments in Bratislava, Brno and Prague. In April he stayed some time in Brno and one of his opponents was Rudolf Pitschak, chess editor of Tagesbote. That is how the world famous position found its way to a German paper in Brno! Whether this game was actually played by Karlin against an amateur, or just something he saw in Madrid in 1933 and showed to Pitschak, would be hard to find out.

Indeed, it all seems a bit too much. Within two years, the same fantastic combination happens three times? Always on the Queen´s side, and always played by Black? And twice in the same year in Madrid?

I don´t think Karlin really played this Rxb2, but he deserves the benefit of the doubt - he probably never pretended he did. He may have shown Pitschak the position (misplacing some bystanders on the King´s side) and said something like: ´Look what I saw in Madrid!´ and Pitschak may have mistakenly thought that he had played Rxb2 himself. Of course, at that moment, Rxb2 was not yet a famous move, just a little gem that Pitschak saw for the first time.

It´s a fascinating thought that Karlin and Sanz may very well have met in Madrid. (They have, see PS of January 2015 below.) Could Sanz have been the ´strong amateur´ Karlin played? Or even the one wrongly credited with the other´s brilliancy?

Both Sanz´ and Wojciechowski´s Rxb2 have nonbelievers - see the links above. And now there are three Rxb2´s, and we can´t be sure any of them was really played over the board. Kalendovský´s discovery is a wonderful addition to the mystery of this famous move.

Ored Karlin (1905 - 1968) was a strong master in the thirties, winning games against Spielmann, Böök and Krogius, and a match (+2 =1 -1) against 24-year old Najdorf in 1934.

PS 15 September: As the scan is hard to read, here is Kalendovský´s transcription.

Samstag , den 14 April 1934
Ein problemartiges Schlussspiel.

Der schwedische Meister Karlin, der soeben an einem kleinen Viererturnier in Brünn teilgenommen hat, hat im Vorjahr eine erfolgreiche Schachfahrt durch Spanien unternommen. Bei dieser Gelegenheit trug er auch in Madrid gegen einen starken Amateur eine Reihe ernster Kämpfe aus. In einem Spiel hiervon gelang ihm ein reizender problemartiger Schluss, den wir in folgender Stellung festhalten wollen:
Karlin führte die schwarzen Steine und gewann auf folgende geistvolle Art:
Ein überraschendes Turmopfer ist die Lösung der Aufgabe. Es muss natürlich schon hier die Pointe der Kombination haarscharf berechnet sein.
Etwas anderes als die Annahme des Opfers bleibt dem Weissen wirklich nicht übrig.
[2...c3 3.Rxb6!]
Wiederum die beste Verteidigung. Der Springerzug nach d3 würde an dem Abzugsschach (c5-c4) scheitern. Falls nämlich erst dann der Turm den Läufer schlägt, gewinnen die zwei verbundenen Bauern auf der sechsten Linie leicht.
Hier erst zeigt sich die Pointe der Kombination des Schwarzen! Dem weissen Springer wird das Feld d3 genommen und nun kommt auch der weisse Turm in schreckliche Verlegenheit, die drohenden Freibauern aufzuhalten.
[4.Rb4 a5!]
Und Weiss muss aufgeben, trotzdem er zwei Figuren mehr auf dem Brett hat. Die Damenverwandlung ist nicht mehr abzuwenden.
Die eines Problems würdige Idee führte hier zu einem originellen Spielschluss.

PS 15 January 2015: Alejandro Melchor from Spain informs me that Karlin played in an 8-men tournament in Madrid 1933, in which both Sanz and Ortueta also participated. (1. Almirall 6, 2-3. Cadenas, Karlin 5, 4-5, Sanz, Añón 4½, 6. Kocher 2, 7-8. Ortueta, Tramoyeres ½.) It seems safe to assume that Karlin was shown the Rxb2 position then and there by Sanz, Ortueta, or both.

PS 30 September 2015: Anders Wirgren from Sweden confirms my thoughts about this Rxb2. He writes:
¨In 1967-68 [Karlin] made a comeback to chess and played on the top board for Lunds Schackklubb. I was only a kid at that time (born in 1951), but I was the best player in the club after him. In 1968 I won the Swedish junior championship; but a few years later I took up bridge, where I was even more successful.¨
¨Since I was young and talented in the late 60's, Karlin liked to analyze with me. I enjoyed it too. From time to time, he brought scrap books to the club and showed games he had played when he was in his prime. If he did play Rxb2 as in the famous Ortueta-Sanz game, he would have shown that to me and been proud to have played something so beautiful. No doubt about that. Who wouldn't? I saw nice games he won against Spielman or Böök or other famous names, but I never saw a combination beginning with Rxb2. Therefore, I am pretty sure that he never played the combination. He was a humble man, and I don't think he ever pretended he did - my guess is that Kalendovsky misunderstood Karlin when he in fact showed something he had seen, not played. You seem to think so too, which surely is right.¨

395. 28 August 2014: Dear Sinquefield Cup,

Could this nonsense about ´the strongest chess tournament in history´ please stop? You have the numbers 1, 2, 3, 5, 8 and 9 of the most recent ranking list. The 1938 AVRO tournament had the numbers 1, 2, 3, 4, 5, 6, 7 and 8. Of your six players, four are in the top-six - a density of 4/6, or .67. The AVRO Tournament had eight of the top eight - a density of 8/8 or 1. Both tournaments are double round robins, with 30 and 56 games respectively. Multiplying these numbers, we get an index of 4/6*30=20 for Sinquefield, and 1*56=56 for AVRO.

AVRO, therefore, was almost 3 times as strong.

Should anyone see these calculations as nonsensical, they should consider that they share that with Sinquefield´s average rating of 2802. Ratings do not reflect playing strength - they reflect relative playing strength and therefore, inflation more than anything else. I could go out right now and buy the most expensive pingpong ball in the history of humankind.

With many thanks for the wonderful tournament,

Tim Krabbé

394. 2 April 2014: Fischer Random, anyone?
(PS 26 February 2015)

I analyzed this a long time ago but I couldn't really remember.
(Aronian about Mamedyarov - Aronian, Candidates 2014)

Mamedyarov claimed he forgot absolutely everything.
(Mamedyarov about Mamedyarov - Aronian, Candidates 2014)

It is in my notebook. But I can't remember.
(Kasparov about Piket - Kasparov, Amsterdam 1995)

12.Bf3 had been prepared by Kramnik for the Candidates tournament in Kazan, however, he could apply it only against Anish Giri last year. In that encounter Vladimir forgot the analysis.
(Kramnik - Giri, Netherlands 2011)

In his game against Anand, Aronian forgot his opening preparation.
(Aronian - Anand, London 2012)

I played too fast in the opening, having forgotten my own analysis.
(Karjakin about Karjakin - Naiditsch, Dortmund 2012)

Tomashevsky believed he was following his home preparation, but unfortunately couldn't remember the lines.
(Nakamura - Tomashevsky, Paris 2013)

Both players forgot their preparation.
(Nakamura - Tomashevsky, Paris 2013)

Anand confessed that he forgot his analysis.
(Nakamura - Anand, Bilbao 2011)

I forgot what I've prepared against 10...c5.
(Nakamura about Nakamura - Kamsky, Zug 2013)

Mamedyarov forgot his analysis after 9.Bh6.
(Morozevich-Mamedyarov, Zug 2013)

Cheparinov, one of the great preparation artists of contemporary chess, forgot a move in the line he had chosen to play.
(Cheparinov-Grischuk, Baku 2008)

Radjabov said he forgot important details of his preparation.
(Svidler - Radjabov, London Candidates 2013)

Leko forgot a critical bit of analysis and played a move he had known was losing.
(Leko - Kramnik, rapid 2007)

Lautier had looked at this opening with Prié, but somehow he forgot this refutation.
(Anand - Lautier, Biel 1997)

I forgot my own preparation and played the wrong move after the sacrifice.
(Kasparov about Kasparov - Lautier, Amsterdam 1995)

Viktor forgot his opening analysis in the first game.
(Stean about Karpov - Korchnoi, Merano 1981)

I completely forgot my preparation.
(Ni about Nakamura - Ni, London 2009)

Here I almost forgot my preparation.
(Wang Hao about Wang Hao - Radjabov, Norway 2013)

Here I forgot my preparation.
(Wang Hao about Wang Hao - Bacrot, Biel 2012)

I forgot my preparation and had to think for 20 minutes.
(Hess, about Shabalov - Hess, US ch semis, 2011)

In a typical fashion, I forgot my preparation.
(Nakamura about Christiansen - Nakamura, US ch 2010)

I completely forgot my preparation after 12.Qh5
(Eljanov about Anand - Eljanov, Germany 2012)

I forgot my preparation, partly because I simply did not understand it.
(Aagaard about Aagaard - Ismagambetov, Olympiad 2012)

I forgot my preparation, but people told me afterwards that what I had intended to play was even worse.
(Elizabeth Vicary, 2006)

Van Wely had annotated it for 'New In Chess', and had come to the conclusion that 12....Nd7 was a mistake! He forgot this over the board.
(Paul Lam about Nakamura - Van Wely, Wijk aan Zee 2010)

I forgot my analysis here.
(Pogonina about Pogonina - Zhao Xue, St. Petersburg 2012)

I simply forgot the variation I'd prepared for the game.
(Caruana about Aronian - Caruana, Moscow 2012)

Fabiano simply forgot all the preparation.
(Chuchelov about Grischuk - Caruana, Thessaloniki 2013)

In general I forgot the lines I had seen there.
(Radjabov about Ivanchuk - Radjabov, Candidates London 2013)

Mamedyarov pointed out that after 5.Qa4 he simply forgot all his analysis.
(Ivanchuk - Mamedyarov, Beijing 2013)

Kramnik told us afterwards that he forgot his analysis and mixed up the plans.
(Polgar - Kramnik, Geneve 2013)

Somehow I didn't fully remember my analysis.
(Kramnik about Kramnik - Anand, Candidates 2014)

PS 26 February 2015:
If I know what I forgot I maybe would remember during the game, but I don't know what I forgot.
(Grischuk about Grischuk - Dominguez, Tblisi 2015)

393. 27 January 2014: New Endgame Tables record: mate in 549 moves.

In entry 316 of this Diary, in May 2006, I gave a record 517-move win, found by Marc Bourzutschky and Yakov Konoval, in the 7-man endgame of Queen and Knight vs. Rook, Bishop and Knight, also known as KQNKRBN. Now Guy Haworth at the University of Reading, in an update of his Chess Endgame Records , publishes, among 91 sometimes very lengthy longest shortest wins in up to 7-man endgames, the deepest known mate: 549 moves, in the endgame KQPKRBN. It was found by a team of programmers at the Lomonosov Moscow State University.

Mate in 549
Lomonosov University, Moscow, August 2012

All the moves below are perfect, but not always in the same way. Some White moves have alternatives with the same depth to mate, some are only-moves to keep that depth, and some are the only moves to win at all. (Moves in that category are marked with an !) The same, vice versa, goes for Black. The whole sequence can be played over on the left; the rightmost symbol will do that automatically (in 25 minutes.)

1.Kf5! Rb5+ 2.Kg4! Rb4+ 3.Kf3! Nd7 4.Qh8+! Ke7 5.g7! Bf6 6.g8N+! Ke6 7.Qh3+! Kd6 8.Kg2! Rb2+ 9.Kf1 Rb1+ 10.Ke2 Rb2+ 11.Kd1 Rb1+ 12.Kc2 Rb2+ 13.Kc1 Rb5 14.Qh6 Rc5+ 15.Kb1 Rb5+ 16.Ka2 Ra5+ 17.Kb3 Rf5 18.Ka3 Rf3+ 19.Ka2! Ke6 20.Qh7! Ne5 21.Nh6! Rf2+ 22.Kb1! Rf4 23.Qg8+! Kd6 24.Ka2 Nc6 25.Ng4! Bd4 26.Qg6+! Kd5 27.Qg5+ Ke4 28.Kb1 Nb4 29.Qe7+! Kd3 30.Qh7+! Ke2 31.Qh2+ Kf3 32.Qh3+ Ke4 33.Qg2+ Kd3 34.Qg3+! Ke4 35.Qe1+ Kd3 36.Qd1+ Ke4 37.Qe2+ Kd5 38.Kc1 Kd6 39.Qh2 Nd5 40.Kd2! Bc3+ 41.Ke2! Bd4 42.Qh6+ Kd7 43.Qh7+ Kc6 44.Qg8 Kc5 45.Kd3 Nb4+ 46.Kd2 Nd5 47.Qc8+ Kd6 48.Nh6 Bc3+ 49.Ke2 Re4+ 50.Kf2 Rf4+ 51.Kg1 Bd4+ 52.Kh2! Ne3 53.Qd8+! Kc6 54.Qe8+ Kc5 55.Nf7! Rf2+ 56.Kh3! Rf3+ 57.Kh4 Rf5 58.Nh6! Rf1 59.Qa8 Kb4 60.Kh5 Kc3 61.Qc8+ Kd3 62.Kg6 Rf2 63.Qa8 Rf1 64.Qb8 Kd2 65.Qb3 Ke2 66.Qb5+ Kd2 67.Qg5 Kc2 68.Kh7 Kd3 69.Qb5+ Kc3 70.Qa5+ Kd3 71.Qg5 Rf6 72.Qb5+ Ke4 73.Qb7+ Kd3 74.Qd7 Kc3 75.Qe8 Kd2 76.Qb5 Kc2 77.Qe2+ Kc3 78.Qh5 Rf1 79.Qg6 Rf6 80.Qe4 Kc4 81.Qe8 Kd3 82.Ng8 Rf1 83.Ne7 Rh1+ 84.Kg6! Rg1+ 85.Kf7! Rf1+ 86.Ke6 Rf6+ 87.Kd7 Bc5 88.Nc6! Nd5 89.Nd8 Rd6+ 90.Kc8 Ne7+ 91.Kc7! Nd5+ 92.Kb7 Rb6+ 93.Kc8! Ne7+ 94.Kd7 Nd5 95.Qe1 Rd6+ 96.Kc8! Be3 97.Qd1+ Ke4 98.Qc2+ Kd4 99.Qa4+ Ke5 100.Qh4 Kf5 101.Qh3+ Ke4 102.Nf7 Rc6+ 103.Kd8 Kd4 104.Ke8 Rc7 105.Kf8 Bf4 106.Qe6 Ra7 107.Qe2 Be3 108.Qg4+ Kc5 109.Qc8+ Kd4 110.Qb8 Kd3 111.Kg8 Rc7 112.Qb5+ Kd4 113.Qa4+ Kc5 114.Qa6 Kd4 115.Kf8 Bd2 116.Qf1 Bc3 117.Kg7 Rc6 118.Kg8 Nf6+ 119.Kf8 Rc8+ 120.Kg7! Ne4 121.Qb5 Ke3+ 122.Kh7! Rc7 123.Qb6+ Rc5 124.Qe6 Kf4 125.Qh6+ Kf3 126.Qh3+ Ke2 127.Kg8 Rc6 128.Kf8 Bb4+ 129.Kg7! Bc3+ 130.Kg8 Rf6 131.Qh4 Kd3 132.Nd8 Kd4 133.Qg4 Ke5 134.Qe2 Kf4 135.Qc4 Kf3 136.Qb3 Kf2 137.Ne6 Ke3 138.Qa2 Rf3 139.Qc2 Be5 140.Qd1 Bc3 141.Qc1+ Kd3 142.Nf4+ Kc4 143.Ng2 Rf2 144.Nh4 Nf6+ 145.Kg7 Ne4+ 146.Kg6 Rf6+ 147.Kh5 Kd4 148.Qd1+ Ke3 149.Qb1 Bd2 150.Ng6 Ng3+ 151.Kh6 Bc3 152.Qc1+ Kd3 153.Qg1 Ne4 154.Kh7! Rd6 155.Kg8 Rd8+ 156.Kf7! Rd7+ 157.Ke6! Rd6+ 158.Ke7 Bf6+ 159.Ke8 Rd8+ 160.Kf7 Rd7+ 161.Ke6 Rd6+ 162.Kf5! Rd5+ 163.Kf4 Bg5+ 164.Kg4! Ke2 165.Qa7 Bd2 166.Qe7 Rg5+ 167.Kh3! Ke3 168.Nh4! Rg3+ 169.Kh2 Rg4 170.Qa3+ Bc3 171.Qa7+! Ke2 172.Ng2 Be5+ 173.Kg1! Nd2 174.Qf2+! Kd1 175.Kh1! Bd4 176.Qe1+! Kc2 177.Qe2! Re4 178.Qh5 Nf1 179.Qd5 Nd2 180.Qf5 Kd1 181.Nf4 Be5 182.Qh5+ Kc2 183.Ng6 Bf6 184.Kg2 Bg5 185.Qh7 Kd1 186.Qh8 Be3 187.Qh5+ Kc2 188.Nh4 Bd4 189.Nf5 Be5 190.Qe8 Rg4+ 191.Kh3! Re4 192.Qc8+ Kd1 193.Qc5 Nc4 194.Nh6 Nd2 195.Ng4 Bd4 196.Qb5 Rf4 197.Kh4 Be3 198.Qd3 Bd4 199.Kh5 Re4 200.Qa3 Rf4 201.Qd6 Re4 202.Nh6 Bf2 203.Kg6 Rc4 204.Qe5 Bb6 205.Qg3 Rc6+ 206.Kg7 Rc4 207.Nf7 Bc7 208.Qg1+ Kc2 209.Nh6 Be5+ 210.Kg8 Re4 211.Nf5 Bf4 212.Qh1 Be5 213.Qh3 Kd1 214.Kf8 Bf4 215.Qg4+ Ke1 216.Qg2 Be5 217.Qh1+ Ke2 218.Ke8 Bg3+ 219.Kd7 Bf2 220.Ne7 Rd4+ 221.Nd5 Kd3 222.Kc6 Rh4 223.Qg2 Rc4+ 224.Kb5 Be3 225.Qh1 Bc5 226.Qh7+ Ke2 227.Qh5+ Ke1 228.Kc6 Bd4+ 229.Kd6 Bf2 230.Kd7 Rd4 231.Qh1+ Ke2 232.Ke6 Rh4 233.Qg2 Rc4 234.Ke7 Nf3 235.Qh3 Re4+ 236.Kd8 Rd4 237.Qh5 Bg3 238.Qe8+ Be5 239.Qb5+ Kf2 240.Kc8 Re4 241.Kb7 Bd4 242.Qc6 Rg4 243.Ka6 Rg7 244.Nb4 Re7 245.Nd3+ Ke3 246.Qc1+ Ke4 247.Nb4 Re5 248.Qb1+ Ke3 249.Qc2 Rc5 250.Qd3+ Kf4 251.Qe2 Kg3 252.Qe7 Rh5 253.Nd3 Rh6+ 254.Ka5 Rh4 255.Qe6 Bc3+ 256.Ka6 Bd4 257.Qc6 Be3 258.Qe8 Bd4 259.Qb8+ Kg2 260.Qg8+ Kh3 261.Qg6 Rg4 262.Qc6 Kg3 263.Qc7+ Kh3 264.Nf4+ Kh4 265.Qd6 60 Kg3 266.Ng6+ Kf2 267.Qe6 Rg1 268.Nf4 Ra1+ 269.Kb7! Ra7+ 270.Kb8 Ra3 271.Qe2+ Kg3 272.Nd3 Rb3+ 273.Kc8 Rb6 274.Qe4 Rf6 275.Kd7 Bc3 276.Nc5 Bd4 277.Ne6 Be5 278.Ng5 Bb2 279.Qe3 Kg4 280.Ne4 Rf7+ 281.Kc6 Rf5 282.Nf2+ Kg3 283.Nd3 Be5 284.Kd7 Rf7+ 285.Ke8 Rf5 286.Qc5 Kg4 287.Nf2+ Kg5 288.Qd5 Kg6 289.Nd3 Bg7 290.Qe4 Kg5 291.Ke7 Bd4 292.Kd7 Rf6 293.Kc8 Rf8+ 294.Kb7 Rf7+ 295.Ka6 Rf6+ 296.Ka5 Bc3+ 297.Ka4 Ra6+ 298.Kb5! Ra5+ 299.Kb6 Rf5 300.Qe7+ Kg6 301.Qe8+ Kg5 302.Kb7 Bd4 303.Qe7+ Kh5 304.Qe4 Rf7+ 305.Ka6 Rf6+ 306.Kb5 Kg5 307.Nb4 Be5 308.Nd5 Rf5 309.Qe3+ Kh5 310.Ka6 Rf7 311.Qf2 Ng5 312.Qc2 Nf3 313.Qc8 Kg5 314.Qe6 Rf5 315.Kb7 Bd4 316.Qe7+ Bf6 317.Qe3+ Kh5 318.Qe4 Kg5 319.Nc7 Nd4 320.Qg2+ Kh5 321.Nd5 Nf3 322.Ne3 Rf4 323.Qh3+ Nh4 324.Qe6 Kg5 325.Qd5+ Kg6 326.Qd6 Kg5 327.Qc5+ Kg6 328.Qc7 Kg5 329.Nd5 Rf3 330.Qc1+ Kg4 331.Ne3+ Kg5 332.Nc4+ Rf4 333.Nb6 Kg4 334.Qd1+ Kg3 335.Nd5 Rf3 336.Qd2 Kg4 337.Ne3+ Kh5 338.Nc4 Bg5 339.Qd1 Bf4 340.Nb6 Bg3 341.Nd7 Kg5 342.Qd5+ Rf5 343.Qe4 Rf4 344.Qe7+ Kg4 345.Qe6+ Kf3 346.Nf6 Kg2 347.Qd5+ Kh2 348.Ne4 Nf5 349.Ng5 Nd6+ 350.Kb6 Bf2+ 351.Kc7 Bg3 352.Kd7 Nf5 353.Qd2+ Rf2 354.Qd1 Rf4 355.Qe2 + Rf2 356.Qh5+ Kg2 357.Kc6 Rc2+ 358.Kd5 Rd2+ 359.Kc4 Rf2 360.Qh3+ Kg1 361.Kb4 Rf1 362.Ka5 Rf4 363.Ka6 Ra4+ 364.Kb5 Rf4 365.Ka5 Rf2 366.Qh8 Kg2 367.Qc3 Rf4 368.Qd2+ Kf1 369.Qd5 Nd4 370.Kb6 Bf2 371.Kc7 Rh4 372.Qe5 Kg2 373.Ne4 Kf3 374.Nd2+ Kg2 375.Qd5+ Kh2 376.Kb7 Rh7+ 377.Ka6 Rh6+ 378.Ka5 Rh4 379.Qd6+ Kh3 380.Qd7+ Kg2 381.Qb7+ Kg3 382.Qg7+ Rg4 383.Nf1+ Kh3 384.Qd7 Kh4 385.Qh7+ Kg5 386.Qg7+ Kf4 387.Qc7+ Kg5 388.Nd2 Rf4 389.Qg7+ Kf5 390.Nc4 Nf3 391.Nd6+ Ke6 392.Ne8 Kf5 393.Ka6 Bd4 394.Nd6+ Ke6 395.Qg8+ Kf6 396.Qh7 Ke6 397.Ne4 Kd5 398.Qb7+ Ke6 399.Ng3 Be5 400.Qb3+ Kd6 401.Qb6+ Ke7 402.Ne4 Kf7 403.Qh6 Kg8 404.Qe6+ Kh8 405.Qe7 Rf5 406.Kb6 Bg7 407.Ng3 Rf6+ 408.Kb5 Nd4+ 409.Kc4! Rc6+ 410.Kd3 Rc3+ 411.Ke4 Rc6 412.Qe8+ Kh7 4 13.Kd3 Rd6 414.Qh5+ Kg8 415.Ne4 Rd8 416.Ke3 Bh6+ 417.Kf2 Rf8+ 418.Kg2 Kh7 419.Ng3 Ne6 420.Qh4 Ng7 421.Qe4+ Kg8 422.Qd5+ Kh7 423.Qd6 Rf7 424.Qd3+ Kg8 425.Qd5 Bc1 426.Kh3 Ba3 427.Kg4 Be7 428.Qc4 Kf8 429.Qc8+! Ne8 430.Nf5 Rf6 431.Qd7 Rf7 432.Qd3 Bf6 433.Qe3 Be7 434.Qh6+ Kg8 435.Qh1 Rh7 436.Qd5+ Kf8 437.Kf3 Nf6 438.Qe5 Ne8 439.Ke2 Bf6 440.Qd5 Rf7 441.Kd2 Bg5+ 442.Kc2 Rc7+ 443.Kb2 Bf6+ 444.Kb1 Rh7 445.Qf3 Ra7 446.Qb3 Rc7 447.Qb4+ Kf7 448.Qb6 Rd7 449.Qb3+ Kf8 450.Nh6 Rg7 451.Qa3+ Be7 452.Qf3+ Bf6 453.Ng4 Ke7 454.Kc2 Rg5 455.Qe3+ Kf7 456.Qb3+ Kf8 457.Qe6 B g7 458.Ne3 Re5 459.Qb6 Bf6 460.Ng4 Rf5 461.Qe6 Rf4 462.Qc8 Bd4 463.Qd7 Bg7 464.Ne3 Rf2+ 465.Kb1 Rf7 466.Qc8 Rc7 467.Qa6 Rd7 468.Qa3+ Kf7 469.Qb3+ Kf8 470.Qb4+ Nd6 471.Qf4+ Kg8 472.Nd5 Nf7 473.Qc4 Ra7 474.Nf4 Rb7+ 475.Kc2 Rb2+ 476.Kd1 Rb6 477.Qc8+ Kh7 478.Qc7 Rd6+ 479.Ke2 Kg8 480.Qe7 Bf8 481.Qe8 Ng5 482.Qc8 Rf6 483.Qg4 Bh6 484.Nh5 Re6+ 485.Kf1 Re5 486.Qc8+ Bf8 487.Qc2 Re6 488.Qa2 Kh7 489.Qg2 Kh6 490.Qh1 Rb6 491.Nf4+ Kg7 492.Qh5 Nh7 493.Qg4+ Kf7 494.Qd7+ Be7 495.Qd5+ Kf8 496.Ne6+ Ke8 497.Qh5+ Kd7 498.Ng7 Nf6 499.Qf5+ Kc7 500.Qe5+ Kd7 501.Qd4+ Kc7 502.Qe3 Nd5 503.Qe5+ Kc6 504.Nf5 Rb7 505.Qe4 Kc5 506.Qc2+ Kb6 507.Qb3+ Kc6 508.Nd4+ Kc5 509.Qxb7 Kxd4 510.Kf2 Ke4 511.Qb1+ Kd4 512.Qg6 Bf6 513.Kf3 Be7 514.Qg1+ Kc4 515.Qg4+ Kc5 516.Ke2 Nf6 517.Qf5+ Kd4 518.Qf4+ Kd5 519.Kf3 Ke6 520.Qc4+ Kf5 521.Qa6 Nd7 522.Qd3+ Ke6 523.Qe3+ Kd6 524.Kf4 Nc5 525.Kf5 Bf8 526.Qf3 Nd7 527.Qc3 Ke7 528.Qd4 Ke8 529.Ke6 Nc5+ 530.Kd5 Nd7 531.Kc6 Nb8+ 532.Kb6 Nd7+ 533.Kc7 Nc5 534.Qc4 Be7 535.Qg8+ Bf8 536.Kc6 Na6 537.Qe6+ Be7 538.Qc8+ Kf7 539.Qxa6 Ke6 540.Qd3 Bg5 541.Qe4+ Kf6 542.Kd6 Bd2 543.Qe5+ Kg6 544.Ke6 Bg5 545.Qf5+ Kh5 546.Qf3+ Kh6 547.Qf7 Bh4 548.Kf5 Bg5 549.Qg6 mate.

- The record position shares honours with two near-identical positions where the black Rook is on b2 and b1 - as the first moves are 1.Kf5 Rb5+ in all three cases, there is not much difference.

- Bourzutschky's and Konoval's record of 517 was of a different kind: both sides aimed not for the quickest/slowest mate, but for the quickest/slowest conversion, through capture or promotion, to a different winnable endgame. Or, in jargon, that struggle was about DTC (Depth to Conversion) instead of DTM (Depth to Mate.) This 517-move DTC record later turned out to have a DTM of 545 moves. As the 6-man position after the conversion at move 517 has a DTM of 8, and White can therefore mate at move 525, that is 20 moves quicker than when DTM-strategy is followed from scratch. It might seem strange that a detour via conversion would lead to a faster mate, but both sides' DTC-strategies are different from their DTM-strategies (the lines diverge at move 226) and apparently, in terms of mate, weaker. If a parent has to take a struggling child to the dentist, the struggling might be less, and the dentist reached sooner, if the child is taken to a tram that will take him to the dentist.

- This 545-move mate (one of 24 KQNKRBN positions with that maximum DTM) also means that the 549-move mate above is not the DTM-record for KQNKRBN. That endgame only arises after White's 6th move, and has a DTM of 543. However, the 549-move KQPKRBN mate is the longest 7-man win and, for now, the longest Endgame Tables win.

- As in all longer Endgame Tables sequences, the moves are incomprehensible. Haworth writes. "These extreme positions are the outposts, the Everests or Mariana Trenches of chess’ state space: they should be hailed, visited and contemplated not only because they are there but because the lines from them can perhaps be analysed and explained in terms of some chessic principles."
    Very perhaps, I'm afraid. In the 1097 moves above, there are at least 1000 that I could never understand. If White is following an infallible path to mate, shouldn't it at least be possible to put the positions below (all with White to play) that are reached on this path, in the right order?

To me, all five seem equally distant from any win. But they represent huge leaps of progress - from left to right, they arise after Black's 100th, 200th, 300th, 400th and 500th move. It is unfathomable that in the 200-move eternity between the 200- and 400-move diagrams, White should have improved his position - if anything, Black seems freeer after 400 moves than after 200.

It is hard to see a shred of conventional strategy. There is no forcing Black's King to the edge or the corner - it is chased (or just goes) to corners, edges and the center in seemingly random fashion. In fact, the fatal position after move 508, where Black cannot avoid the loss of the exchange, occurs quite suddenly when his King is on c5. White's King too, marches all over the board - it only leaves 12 squares unvisited.

A scary part of Haworth's article is where he extrapolates the findings in sub-8-man Endgame Tables to predictions for 8- to 10-man endgames.

It is "pretty awesome that maxDTM seems to be doubling with each man," Haworth writes - in this logarithmic graph, we see what that might mean: an 80% probability that a 10-man endgame exists where a forced mate takes more than 5,000 moves.

With Thanks to Guy Haworth.
His publications, many of them about Endgame Tables, can be found here.

Earlier articles about Tablebase Endgames on this site:
Stiller's monsters or perfection in chess
In Open Chess Diary:
#60: Play chess with God
147: Chess at particle level
189: The Palview dream
282: First 7-men endgame database
283: A smaller board
294: The mindboggling perfection of 7-man endgames
298: New endgame record - a 290 move win
311: White plays and wins in 330 moves
316: A 517-move win
324: A cooked, correct study
369: Shirov in the face of Total Truth
370: Spotting progress in a tablebase endgame
In Dutch:
Het vijfpaarden eindspel
De driehonderd zetten barrière
Het fruitvliegje van de K.I.
Schaken met God

392. 22 January 2014: An obvious, hard to see breakthrough

In entry 322 of this Diary, I wrote about an elementary breakthrough in pawn endgames, which had been missed many times - often by both players, and sometimes more than once in one game. In all 27 examples I gave, a blunder enabled it; in 20 it was missed - and in 12 of those, that cost a full point.

I wrote that in August 2006 - in the meantime, there have been five new examples - in all of which this breakthrough was overlooked by both players.

Medina - Pham, Ho Chi Minh City, 2012
Black to play

White, a pawn up and totally winning, had just blundered with g3-g4. With 46...Ke5, Black missed her chance: 46...g5 is a simpe win, e.g. 47.Kd3 gxh4 48.Ke3 hxg4 49.fxg4 Kd5 etc.
After Ke5, White is again winning. There followed: 47.g5 Kf4 48.Kd4 Kg3 49.Ke5 Kxh4 50.Kf6 Better was 50.f4 Kg3 51.f5 h4 52.hxg6 etc. 50...Kg3 51.Kxg6 h4 52.Kf5 h3 53.g6 h2 54.g7 h1Q 55.g8Q+ Kxf3 56.Qb3+ Ke2 57.Qc2+ Ke3 58.Qc3+ Ke2 59.Qc4+ Ke3 60.Ke6 Qd1 61.Qc3+ Kf2 62. Qf6+ Kg3 63.Qg5+ Kf2 and although this is still a tablebase mate in 56, a draw was agreed ½-½.

Sandor - Boguslavsky, Budapest, 2012
Black to play

In a drawn position, White had just blundered with 37.g2-g4, but with 37...g6? Black allowed him to escape - which he didn't. 37...g5! was winning: 38.hxg5 h4 39.Ke3 fxg5 or 38.gxh5 f5+ 39.Ke3 gxh4, followed soon by e4. 38.gxh5 gxh5 39.c4 f5+ 40.Ke3 a5 41.b3 Kd6 42.a3 Kc5 43.Kd3 b5 44.Kc3? (44.cxb5 is still a draw) 44...bxc4 45.bxc4 a4 46.Kd3 e4+ 47.fxe4 fxe4+ 48.Kxe4 Kxc4 49.Kf5 Kb3 and a few moves later, White resigned 0-1.

Jiang - Kelly, Stillwater 2011
White to play

Again, the same two blunders. Black had thrown away a draw with 36...b7-b5, upon which White threw away a win and lost with 37.b3? 37.b4 wins, e.g. axb4 38.a5 Kd6 39.e5+ etc.) There followed: 37...b4 38.Ke3 c5 39.Kd3 Kf4 40.e5 Kxe5 41.Kc4 Ke4 42.Kb5 Kd3 43.Kxa5 c4 and 20 moves later, White resigned 0-1.

Skawinski - Tomczak, European ch rapid, Warsaw 2011
Black to play

Black is winning, which he quickly did with a blunder: 40...b5? (40...f4!) 41.Kc3? bxa4 42.bxa4 c5 43.dxc5 Kxc5 44.g4 fxg4 and White resigned 0-1.
White could have won with 41.b4!, but there is an interesting finesse: 41...axb4 42.a5 Kd6 43.g4 (White must at some point neutralize Black's breakthrough g4) hxg4 44.hxg4 fxg4 45.fxg4 Kc7 46.Kc2 Kb7 47.Kb3 Ka7 48.Kxb4 Ka6 and now White must allow Black to promote first, but he still wins: 49.Kc5 Kxa5 50.Kxc6 b4 51.d5 b3 52.d6 b2 53.d7 b1Q 54.d8Q+ Ka4 55.Qa8+ Kb3 56.Qb7+ Kc2 57.Qxb1+ Kxb1 58.Kd5 etc.

PS 22 October 2015: Ricky Demer shows that after [40...b5?] 41.b4, Black can use his Kingside breakthrough to draw: 41...bxa4! 42.bxa5 a3 43.Kc2 a2 44.Kb2 a1Q+ 45.Kxa1 g4 46.hxg4 f4 47.gxf4 h4 48.a6 h3 49.a7 h2 50.a8Q h1Q+ and although White is three pawns up, it´s a draw: perpetual, tablebase or otherwise, e.g. 51.Kb2 Qg2+ 52.Kb3 Qxf3+ 53.Kb4 Qxf4 54.Qd8+ Ke4. Same kind of thing after 46.fxg4 h4.
See Demer´s detailed analysis.

Zengeni - Olatunji, Ol. w Istanbul, 2012
White to play

A special case.
That same Queen's ending trick could have been used here, Black had just made the mistake 69...g6-g5, but after 70.Kd3 gxh4 71.gxh4 Kf4 she won without much ado: 72.Kd4 Kg4 73.e5 fxe5+ 74.Kxe5 Kxh4 75.Kf4 Kh3 76.Kf3 h4 77.Kf2 Kg4 78.Kg2 h3+ 79.Kh2 Kh4 80.Kg1 and White resigned 0-1.
White could have used the same double trick and won: 70.g4! hxg4 71.h5 Ke6 72.Kf2 Kf7 73.Kg3 Kg7 74.Kxg4 Kh6 75.Kf5 Kxh5 76.Kxf6 g4 77.e5 g3 78.e6 g2 79.e7 g1Q and again 80.e8Q+ Kh4 81.Qh8+ Kg3 82.Qg8+ Kf2 83.Qxg1+ Kxg1 84.Ke5 Kf2 85.Kd5 Ke3 86.Kc5 Kd3 87.Kb6 Kc3 88.Kxa6 Kb3 89.Kxb5 and wins.

What is so special about that? The fact that the exact position of the diagram had already been shown by the Italian endgame theoretician Salvioli in 1888, in his Trattato Completo dei Finali di Partita - that was the premiere of both the breakthrough and this Queen's ending. In this entry 322, I also showed two games in which the Salvioli position occurred, but that was with colors reversed. (In one of those games, the win was found by a 10-year old, Dov Zifroni, who became a grandmaster with a top rating of 2550 in 2003.)

The Salvioli study can be played over on the left.

391. 18 December 2013: Carlsen castles

The least spectacular sort of move was the most spectacular move in the London tournament.

Black to play
Gelfand - Adams
London rapid, 2013

This can't even be called a blunder - it has nothing to do with tactics. Adams probably just forgot that castling is still legal at this stage of the game. With the double attack 25.O-O-O White won an exchange and, some 30 moves later, the game.

Of course this castling cheapo has a long beard - it is even theory: 1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 d6 5.Bxc6+ bxc6 6.d4 f6 7.Nc3 Rb8 8.Be3 Rxb2? 9.dxe5 fxe5 10.Nxe5! and now 10...dxe5 is not possible because of 11.Qxd8+ Kxd8 12.O-O-O+ and White wins.

Art preceded life - in its most elementary form:

White to play and win
A. Selesniev
Tidskrift for Schack, 1921

1.d7 Kc7 2.d8Q+ Kxd8 3.0-0-0 and wins.
White must not play his trump too soon: 1.0-0-0? Ra2! or 2.0-0-0? Rb8!

This is the earliest known example - in games or studies.

The question arises whether this castling double attack could also happen on the King's side. That would seem unlikely, as it calls for the thematical Rooks facing each other, when the one on the second/seventh rank could already be captured. The tactics must therefore be a little more complicated - but in item #295 in this Diary, I gave an example where it almost happened:

Black to play
Kupreichik - Kapengut
Minsk 1978

White has just taken on h7, and after 16...Rxh7 17.Qxg6+ Kf8 18.Qxh7, Black would be a pawn down. But he found the witty 16...Bxf2+! Now Kxf2 loses an exchange to 0-0+, and after 17.Ke2 Rxh7 18.Qxg6+ Kf8 19.Qxh7 Qh4 20.Qxh4 Bxh4 Black had saved his pawn - the game ended in a draw ten moves later.

Searching again, I found a later game in which a double attack with 0-0 did appear on the board. Of all these castlings, it is the most interesting, but even if the castler was Magnus Carlsen, the game seems to have gone unnoticed. (Doubtlessly, it will now soon turn up, presented as his own research, in the writings of Professor Christian Hesse.)

Karjakin - Carlsen, Blindfold World Cup Rapid, Bilbao 2007
1.e4 c5 2.Nf3 Nc6 3.Bb5 e6 4.Bxc6 bxc6 5.d3 Ne7 6.Qe2 Qc7 7.Ng5 e5 8.f4 exf4 9.O-O Ng6 10.Qh5 d6 11.Bxf4 Nxf4 12.Rxf4 g6 13.Qf3 Bg7 It seems incredible that this is possible. 14.Rxf7 Bd4+ 15.Kh1 Qd8 16.c3 Be5 17.Rg7 King's side castling does not seem imminent, but it is the refutation of this brilliancy. 17.h4 was still a draw. 17...Qf6 And not 17...Rf8 18.Rg8 and White remains a pawn up. 18.Qxf6 Bxf6 19.Rxh7
see diagram
Attacking the still guarded Rook and the not yet accessible f1 - the latter is essential, because otherwise White could escape with 20.Rh6 Bxg5 (Bg7 21.Rxg6) 21.Rxg6+ and 22.Rxg5. Now, that would lead to mate by Rf1. So: Black wins a piece.
After 20.Na3 Bxg5 21.Rc7 Rf7 22.Rxc6 Bf4 23.Nc4 Bd7 24.Ra6 Bb5 25.Ra5 Bxc4 26.dxc4 Be5 27.Rd1 Raf8 28.g3 Rf2 29.b4 Bxc3 30.Rxa7 Bd4 31.Rd7 Rxa2 32.bxc5 Rff2 33.Rd8+ Kg7 34.Rd7+ Kh6, White resigned.

This is the only game in a 3 million game database with a double attack by 0-0.

PS 19 December: Knowing Chess Today's healthy predelection for unusual tactics, I double-checked my separate CT-database, and found that back then, they had already singled out this game (it was funny to see the ratings, 2694 for White and 2714 for Black) and Mikhail Golubev had extensively analysed it.

A few highlights:

- Up to 15.Kh1, the players followed an analysis by Maxim Notkin in CT-2200.
- 16.Nxh7 Bf5 17.Rxf5 gxf5 18.Qh5+ Kd7 19.Qxf5+ Kc7 20.Ng5 seems to be equal.
- 17.Rg7 was, as Golubev says, "the wrong execution of a very good idea." White should have played 17.h4 h6 and only now 18.Rg7 Rf8 Then, 19.Rf7 is a possible draw by repetition and 19.Rg8 Qe7 20.Rxf8+ Qxf8 21.Qxf8+ Kxf8 22.Nf3 Ba6 leads to "approximately equal chances."
- 18.Rxh7 is better (than Qxf6) but after Rxh7 19.Nxh7 Qh4 20.Qf8+ Kd7 21.Qf7+ Kd8 22.g3 Bxg3 23.Qf6+ Qxf6 24.Nxf6 Be5 25.Nh7 Rb8 "White is in trouble, despite his extra pawn."

390. 13 December 2013: Emmser Hesse

It's about time Mr. Christian Hesse (ChessBase, books) wrote something he couldn't first have read in my work.

389. 25 November 2013: The 9th matchgame riddle

After the 9th matchgame in Chennai, which Carlsen won with Black without a move by his still present Queen, there was some speculation that this might have been unique. It isn't really; excluding games of under 25 moves, I found such queenmoveless wins in around 1 in 900 games.
    But there was also the riddle-like circumstance that Carlsen still had his Queen, had never moved it, yet won with (that is, his opponent resigned immediately after) a Queen move. That was a promoted Queen of course, and that was special - I only found four earlier games (in a 3 million game database) that are also solutions to this riddle.

Alzate - Handoko, Luzern, Olympiad 1982
1.e4 e5 2.Nf3 Nc6 3.Bb5 f5 4.Nc3 Nd4 5.O-O Nxb5 6.Nxb5 fxe4 7.Nxe5 Nf6 8.Ng4 Be7 9.Nxf6+ Bxf6 10.Qh5+ g6 11.Qc5 d6 12.Qxc7 O-O 13.Qxd6 Bg4 14.Qg3 Be2 15.Re1 Bxb5 16.Qb3+ Kh8 17.Qxb5 Bd4 18.Kh1 Rxf2 19.c3 Bg7 20. d4 exd3 21.Be3 Rf5 22.Qxb7 Rb8 23.Qxa7 Rxb2 24.Rad1 d2 25.Re2 Rb1 26.Rxb1 d1Q+ 27.Re1 Q1d6 28.Rb7
see diagram
28...Qxh2+ and White resigned; it's mate in 3.

388. 22 November 2013: Hull er tull

Information about Magnus(IM) (Last disconnected Wed Jan 28 2004 09:01):

          rating [need] win  loss  draw total   best
Wild        1670  [6]     0     2     0     2
Loser's     1632  [6]     0     7     0     7
Crazyhouse  1869  [6]     1     4     0     5
Bullet      2256        370   424    71   865   2256 (28-Jan-2004)
Blitz       3166       1746  1684   388  3818   3199 (26-Jan-2004)
Standard    2180  [6]     7     5     1    13
5-minute    2357  [8]   145    97    25   267   2357 (03-Jan-2004)
1-minute    2225        710   599    83  1392   2273 (31-Dec-2003)

 1: Magnus Carlsen (13) from Norway.
 2: I'm better than JonLudvig
 3: Current FIDE 2484  5. in Norway !
 4: 1-2 in B12 World Youth Champ 2002   3 in European Youth Champ B14 2003
 5: Her stod noe før, men siden det ville meningsløst at det ikke stod noe
her står det noe meningsløst i stedet
 6: 1-minute er dataspill
 7: jeg er den tregeste spilleren med 3000
 8: hull er tull
 9: 1 GM-norm Corus 2004 10,5 of 13 needed 9,5 !!!

 Groups : Norway Ims

387. 23 August 2013: 54 consecutive checks by Alexey Khanyan

Alexey Khanyan did it again: he took back the record for consecutive checks from Sampsa Lahtonen, now with a stunning series of 54 checks.

54 consecutive checks (legal position, promoted pieces allowed)
Alexey Khanyan (after Lahtonen)
Original, 2013

1.Qb6+ Rc5+ 2.Qd6+ Nxd6+ 3.Bf6+ Nxf6+ 4.Nfg4+ Qf2+ 5.Nhf3+ Kd5+ 6.e4+ Nfxe4+ 7.Nf6+ Nxf6+ 8.Be4+ Nfxe4+ 9.Qf5+ Nxf5+ 10.Rad6+ Nexd6+ 11.Bc4+ Nxc4+ 12.Re5+ Nxe5+ 13.Nc4+ Qfd2+ 14.Nfxd2+ Rf3+ 15.Nxf3+ Qd2+ 16.Ncxd2+ Rc4+ 17.Bxc4+ Qxc4+ 18.Ne4+ Bd2+ 19.Qxd2+ Nd3+ 20.Qxd3+ Nd4+ 21.Nf6+ Qxf6+ 22.Qf5+ e5+ 23.Nxe5+ Rf3+ 24.Nxf3+ Qe5+ 25.Nxe5+ Nf3+ 26.Qd4+ Qxd4+ 27.Qe4+ Qxe4+

A pgn-file can be downloaded for watching this on your home screen. And the position is legal; Khanyan sent me a proof game.

As Khanyan used Lahtonen's matrix of his 53-checks record (see item 386 below), and he did not want to hide that in any way, he at first felt an obligation to publish it without reversing the colors, and therefore with a "null-move" as his modification made it necessary for Black to start the checks:

54 consecutive checks (legal position, promoted pieces allowed)
Alexey Khanyan (after Lahtonen)
Black to play
Original, 2013

0...Qg3+ 1.Rf4+ Qe3+ 2.Nxe3+ Bc3+ 3.Nxc3+ Ncb5+ 4.Qc7+ Nac6+ 5.Ke4+ d5+ 6.Ncxd5+ Nc3+ 7.Nxc3+ Bc5+ 8.Ncxd5+ Qc4+ 9.Nxc4+ Rhe3+ 10.Ndxe3+ Bf5+ 11.Nxf5+ Rd4+ 12.Nxd4+ Nf5+ 13.Qce7+ Ncxe7+ 14.Rc6+ Nxc6+ 15.Qe7+ Nfxe7+ 16.Rf5+ Bxf5+ 17.Qxf5+ Nd5+ 18.Be7+ Qxe7+ 19.Ne6+ Qxe6+ 20.Ne5+ Nc3+ 21.Qxc3+ Qc4+ 22.d4+ Nxd4+ 23.Rc6+ Nxc6+ 24.Qd4+ Nxd4+ 25.Nc6+ Qe5+ 26.Qxe5+ Qd5+ 27.Qxd5+

This is probably not the end to this task. The fact that in Lahtonen's 53 checks only 5 pieces remain in the final position and in the new 54-record there are 6, leads Khanyan to believe that if he could get this down to 5 pieces, too, he could add two checks and make it to 56. "56 checks in a row do not seem to me as being higher than the clouds!"

386. 15 December 2010: 53 consecutive checks by Sampsa Lahtonen

When I published Alexey Khanyan's record of 51 consecutive checks (legal position, promoted pieces allowed) almost three years ago (see item 378), I concluded with his remark: "52 checks in a row is not a completely impossible result. I am going to look for them."

But now I received a position with 53 consecutive checks by Sampsa Lahtonen, who had held this record before.

53 consecutive checks (legal position, promoted pieces allowed)
Sampsa Lahtonen
Original, 2010

1.Qc1+ Bc3+ 2.Nxc3+ Ndb5+ 3.Qc7+ Nac6+ 4.Ke4+ d5+ 5.Ncxd5+ Nc3+ 6.Nxc3+ Qgd5+ 7.Ncxd5+ Qc4+ 8.Nxc4+ Rge3+ 9.Ndxe3+ Bf5+ 10.Nxf5+ Rd4+ 11.Nxd4+ Nf5+ 12.Qe5+ Nxe5+ 13.Rc6+ Nxc6+ 14.Qe7+ Nfxe7+ 15.Qf5+ Nd5+ 16.Be7+ Qxe7+ 17.Ne6+ Qxe6+ 18.Ne5+ Nc3+ 19.Qxc3+ Qc4+ 20.Qd4+ Nxd4+ 21.Rc6+ Qxc6+ 22.Nxc6+ Nxf5+ 23.d4+ Nxd4+ 24.R1f5+ Bxf5+ 25.Rxf5+ Qe5+ 26.Rxe5+ Qd5+ 27.Rxd5+

The history of the record is:
45 - Werner Frangen, 1974 (see entry 308 in this Diary)
46 - Sampsa Lahtonen, 2007 (entry 351)
47 - Sampsa Lahtonen, 2007 (entry 351)
49 - Unto Heinonen, 2007 (entry 374)
50 - Sampsa Lahtonen, 2008 (entry 375)
51 - Alexey Khanyan, 2008 (entry 378)
And now:
53 - Sampsa Lahtonen, 2010

Lahtonen sent me a proof game for the legality of the above position.

385. 10 April 2009: Karpov - Krabbé

I have to correct an important error Hans Ree makes in his interesting article about blindfold chess in the latest issue of "New in Chess". Ree (who lately has never passed an opportunity to mention that Fischer died at the chess age of 64, although always in the form of a sneer at others who have also mentioned this), says that the game Karpov - Krabbé in Karpov's 5-board blind clock simul in Amsterdam 1992, was a draw. But I won that game - an interesting game, too.

Karpov (blind clock (40/2) simul, 5) - Krabbé
Schaken op het Spui, Amsterdam, 6 September 1992

1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Ba4 Nf6 5.O-O Be7 6.Bxc6 dxc6 7.d3 Nd7 8.Nbd2 O-O 9.Nc4 f6 10.Nh4 Nc5 11.Nf5 Be6 12.b3 Bxc4 13.bxc4 Ne6 14.Qg4 Kh8 15.f4 Bc5+ 16.Kh1 g6 17.Ng3 Nxf4 18.Bxf4 exf4 19.Rxf4 b5 20.e5 f5 21.Qf3 Bd4 22.Re1 c5 23.e6 I had a small advantage, which I could have kept with Qe7 or Bf6. After 23...Qd6, it was White who was better. 24.e7 Rfe8
see diagram left
As the Pe7 would otherwise be lost, Karpov now had to choose between two sacrifices on f5.
The wrong one. I remember that Ree himself published this game in his newspaper column, and that a reader later pointed out that with 25.Rxf5! gxf5 26.Qxa8 Rxa8 27.e8Q+ Qf8 (Rxe8 28.Rxe8+ Kg7 28.Nxf5+) 28.Qh5 White could have kept a sizeable advantage. 25...gxf5 26.Rxf5 Now, Black would be better after Rab8, but: 26...Rg8
see diagram left
Chessplayers will be chessplayers - somebody like Karpov plays such exhibitions for the money, but while the closing ceremony was still going on, and people were still applauding him for his 3½-1½ win, he turned his back to them, stepped over to me, and said: "I could have won with Rg5." He was right: 27.Rg5 Qd7 28.Qf7 Rxg5 29.e8=Q+ Qxe8 30.Rxe8+ Rxe8 31.Qxe8+ and Black can resign. After 27.Rd5 Qg6 28.g3 Rae8 29.Rd7 c6 30.Re2 Qg4 31.Qxg4 Rxg4 32.Rd8 Rgg8 33.Rd6 Rg6 34.Rd8 Rgg8 35.Rd6 Rc8 36.Ree6 Rge8 37.Kg2 Kg7 38.Kh3 Rc7 39.Rxc6 Rxc6 40.Rxc6 Rxe7 41.c3 Be3 42.Rxa6 bxc4 43.dxc4 Re4 (see diagram right) the game was adjourned, and adjudicated (by Timman, I believe) as a win for me.

I also remember this game for a missed quip. That same evening I was in a TV-program for something else, but the interviewer knew about the simul and said: "That must have been a boyhood dream, to beat Karpov."
    I dutifully - and rightfully - played down my achievement, saying that Karpov's double handicap was enormous, and that chances would have been about even. It still nags me that I didn't say: "No, my boyhood dream was to sit there and be beaten by Tim Krabbé."

Ree writes about the wonderful, but sadly discontinued chess Sunday of "Schaken op het Spui", a street event where 4-man teams competed in a blitz tournament for the right to play a top grandmaster in a blind simul - a "Mystery Guest" (a chess-playing celebrity) being the 5th opponent. My five participations in that simul offer a clue as to how big the handicap of a blind 5-board clock simul is. My score, between 1991 and 1998, was +1, with wins against Karpov and Van Wely, draws against Anand (see A love story with a diagram, elsewhere on this site) and Yusupov, and a loss against Sokolov. So I guess I was right when I said chances had been about even against Karpov. As my rating must have been around 2150-2200 in those years, and the GMs I played would have been 2650-2750, that suggests their task made them weaker by some 500 rating points.

384. 14 February 2009: I'm fine, but...

Several readers have expressed their concerns about my health, as I had not updated this Diary in over 4 months. In fact, I had only added one new item in the 4½ months before that. Thanks for worrying, but don't - my health is perhaps too good for this Diary, as I'm devoting more and more time to my other sport, cycling. And to my real work, writing fiction.
    This is not a final goodbye - there will be something new here sometime, but don't expect anything too soon.

383. 7 October 2008: Quantum leaps

In entry 334 of this Diary, I introduced an idea for a construction task: the Worst Possible Move. White has n moves, of which (n-1) are mate in 1 and the nth, the Bad Move, is selfmate in 1 - a higher n meaning a worse move.
    This challenge was immediately taken up by Noam Elkies with a position with n=28, based on a 1936 position with 29 forced mates (White has 29 moves, all of which are mate in 1) by H.H. Cross. Both records still stand.

But in this kind of task, there is always a separate record for (legal) positions with promoted men, either with or without promotion moves. And especially in the latter category, various composers outbid each other with record upon record, until in January of this year, over a year after the thread had started, Alexey Khanyan came up with n=48, based on an old position by Wolfgang Dittmann with 50 forced mates.

White to play
50 forced mates (legal, promoted men, no promotion moves)
Wolfgang Dittmann, Die Schwalbe 1967

White has 50 moves, 14 mates by Rb2; 1 by Pe4; 8 by Qg4; 6 by Qh5; 5 by Qh7; 4 by Qg8; 3 by Qf8; 4 by Qe8 and 5 by Nf7.

White to play
Worst possible move, n=48 (legal position, promoted men, no promotion moves)
Alexey Khanyan (after Dittmann), January 2008

White has 48 moves, 47 of which are mate, the 48th (Qe6+) forcing Black to mate him in one.

And now Sampsa Lahtonen from Finland, a regular guest to these pages who has set (and lost) quite a few task records, comes with quantum leaps in both domains, adding no less than 5 mates to Dittmann's 50, and adding no less than 4 to Khanyan's n=48 in the Worst Possible Move task.

White to play
55 forced mates (legal; promoted men, no promotion moves)
Sampsa Lahtonen (after Dittmann), 2008

White has 55 moves, all of which are mate in one - 12 by Rg2; 1 by Pb4; 5 by Qa5; 1 by Qe5; 6 by Qa7; 7 by Qe7; 7 by Qb8; 7 by Qd8; 1 by Rf7 and 8 by Nc5.

White to play
Worst possible move, n=52 (legal position, promoted men, no promotion moves)
Sampsa Lahtonen (after Dittmann), 2008

White has 52 moves, 51 of which are mate in one (12 by Rg2; 8 by Qe7; 6 by Qd8; 6 by Nb6; 5 by Qb8; 4 by Qa7; 3 by Qa6; 4 by Qa5; 1 by Qe5; 1 by Pb4; 1 by Rf7) and the 52nd (Qxc4+) forcing Black to mate him in one.

Both positions seem to be easily legal.

PS 15 December 2010: Joaquim Crusats informs me that Lahtonen's 55 forced mates were anticipated in 1972 by Ludwig Zagler. See entry 386 above for a nice answer by Lahtonen.

55 forced mates (legal; promoted men, no promotion moves)
Ludwig Zagler
Feenschach 1972

All of White's 55 moves are mate in one - 12 by Rb2; 1 by Pg4; 1 by Qd5; 7 by Qd7; 7 by Qe8; 7 by Qg8; 6 by Qh7; 5 by Qh5; 1 by Rc7 and 8 by Nf5.

382. 14 July 2008: Willy Hendriks' most beautiful move (+ PS 30 September 2015)

In this week's Schaaknieuws, Dutch IM Willy Hendriks shows "the most beautiful move of my career." Two things about it, however, are regrettable: "I didn't think of it myself (Fritz did) and I didn't play it (but I was close.)"

Hendriks - Spanton, Hastings 2006
1.e4 e5 2.Bc4 Nf6 3.d3 c6 4.Nf3 d5 5.Bb3 Bb4+ 6.Bd2 Bxd2+ 7.Qxd2 dxe4 8.Nxe5 O-O 9.dxe4 Qe7 10.Qf4 Nh5 Hendriks had been looking at this variation, thinking it was nothing for White, because he seems to lose his Ne5. Then he noticed Fritz liked White, because of the Move. White only has 11.Bxf7+ "My heart started pounding, as I could already see how the spectators would flock around my board, wondering how White would be able to extricate himself, only to marvel in ecstasy and disbelief at my move." Alas, Spanton did not play Kh8, but smashed Hendriks' hopes by losing quickly with 11...Rxf7 12.Qxf7+ Qxf7 13.Nxf7 Kxf7 14.Nd2 Be6 15.f4 g6 16.O-O Nf6 17.f5 gxf5 18.exf5 Bd5 19.c4 Be4 20.Rf4 Bd3 21.Rd4 Bxf5 22.Rf1 Bg6 23.Rd6 and Black resigned.

After 11...Kh8 (see diagram), White still seems to lose the Ne5. What should he play? Hendriks showed the position to various strong players, "and some of them couldn't find it, even after a long think."

Solution in a few days - or any moment you decide to switch on your chess engine.

PS 19 July 2008: The move is of course 12.Qg3, when Nxg3? leads to a symmetrical clearance of the h-file: 13.Ng6+ hxg6 14.hxg3 and mate.
    As one reader said: "It took me - I hate being honest about this - about 7 minutes, and of course I would NEVER have found it without being told there was something worth looking for."
    After 12...Rxf7 etc. White has the same tiny advantage as in the game, but it's worth to note that instead of 13.Nxf7+, White can play another typical computer move, 13.Qg5, keeping that same advantage.

PS 30 September 2015: In the 8th Finals of the still ongoing World Cup in Baku, Wei Yi actually played Hendriks´ (or his own? - his computer´s?) brilliant 12.Qg3 against Ding Liren. The game ended in a draw after 12...Rxf7 13.Nxf7+ Qxf7 14.Qd6 Be6 15.Nc3 Nd7 16.O-O-O Re8 17.Rhf1 Bc4 18.Rfe1 Ne5 19.b3 Ba6 20.Kb1 h6 21.f3 Nf4 22.Rd2 Kh7 23.Red1 Re6 24.Qb8 Qf6 25.Na4 Be2 26.Rc1 b6 27.Nc3 Ba6 28.Rcd1 Nc4 29.bxc4 Qxc3 30.Qxf4 Qb4+ 31.Ka1 Qc3+ draw agreed.

When I searched my database for possible earlier examples, I discovered something heartbreaking: Hendriks did get to play his Move after all, two years before Wei - only: he lost that game.
In Hendriks - Kerigan, Univé Open, Hoogeveen 2013, there followed, after 12.Qg3: 12...Rxf7 13.Nxf7+ Qxf7 14.Qd6 Be6 15.Nd2 (Wei played Nc3 here) 15...Nd7 16.O-O-O Re8 17.f3 Nf4 18.g3 Ne2+ 19.Kb1 Bxa2+ 20.Ka1 Re5 21.b3 Re6 22.Qa3 c5 23.Nc4 Ra6 24.Qxa6 bxa6 25.Kxa2 Nc3+ 26.Kb2 Nxd1+ 27.Rxd1 h6 28.f4 Nb6 29.Ne5 Qb7 30.f5 Na4+ 31.Kc1 Qxe4 32.Ng6+ Kh7 33.bxa4 Qxf5 34.Nf4 Qe4 35.Rd5 c4 36.Kb2 Qe1 37.c3 Qf2+ 38.Ka3 Qb6 and White resigned.

That does not change the fact that 12.Qg3 is Hendriks´ move: he saw it first and he played it first.

381. 21 May 2008: Earliest King move to cause resignation

A reader, Robin Fishbein, asks "what the earliest King move was that caused an opponent to resign".

The answer is one of my great disappointments in chess - a game from a simul I once gave in the south of the country.

Krabbé - NN, Bergen op Zoom 1986
1.e4 e6 2.d4 Nf6 3.e5 Nd5 4.c4 Bb4+? 5.Ke2! (see diagram) and after five moves, four of them with pawns and the last one with the King, and with all the pieces still on the board, White is winning a piece - both after 5...Nb6 6.c5, or 5...Ne7 6.a3 the black Bishop's retreat is cut off, and it will be lost.

To be honest, I had already seen this trick, but in that predecessor, Smekalin - Yegorov, Chelyabinsk 1983, Black had not resigned immediately, but had continued with 5...Nb6 6.c5 Qh4 7.cxb6, resigning only then.

Of course I hoped that my NN would resign now, making me the answer to Fishbein's future question. But when I came back to that board, I was in for a disappointment. Not only did NN not resign, playing 5...Nf4+ instead (see diagram on the right) - but when I wanted to capture with 6.Bxf4, I noticed this was not possible, because the board had been set up wrongly. I thought of playing 6.Bb1-c1xf4 in one go, but, deciding this would not be in the spirit of the evening, I accepted to play a new game with my Bishops and Knights in the correct positions.

Later this pain was softened when I discovered that a game with 5.Ke2, followed by an immediate resignation, had been played nine years earlier - in the Netherlands, of all places.

Den Broeder - Van der Ent, correspondence, Netherlands 1977
1.e4 Nf6 2.e5 Nd5 3.d4 e6 4.c4 Bb4+ 5.Ke2 and Black resigned.

© Tim Krabbé 2008 - 2014

Top | Main chess page | Main page | Contact me |