29 November 2007

Opening Exchange Sac [B7x]

One of the best known opening sacrifices is the move ...Rxc3 against the Yugoslav Attack in the Dragon Variation of the Sicilian Defense. The idea has several motivations and works in many positions. Who played it first at the international level and when was it played?

After a few minutes search, the earliest example I could find is shown in the diagram. It is game no.1 in Karpov's 'Best Games' (Batsford, 1996). The opening moves were: 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 g6 6.Be3 Bg7 7.f3 O-O 8.Bc4 Nc6 9.Qd2 Qa5 10.O-O-O Bd7 11.h4 Ne5 12.Bb3 Rfc8 13.h5 Nxh5 14.Bh6 Bxh6 15.Qxh6. It is arguable whether we are still in the opening here.

The players have just exchanged Bishops on h6. Now Black played 15...Rxc3 and Karpov wrote:

The standard exchange sacrifice in the Dragon. On the one hand Black protects himself from the Knight lunging onto d5, and on the other hand he shatters the enemy King's fortress.

The note indicates that the idea was already well known at the end of the 1960s. The sacrifice is even stronger when the dark-squared Black Bishop is still on the long diagonal.

Moscow 1968-69
Gik, Evgeny

Karpov, Anatoli
(After 15.Qd2-h6(xB))
[FEN "r1r3k1/pp1bpp1p/3p2pQ/q3n2n/3NP3/1BN2P2/PPP3P1/2KR3R b - - 0 15"]

The game continued 16.bxc3 Qxc3. Karpov again:

The Black Queen here is occupying an ideal position to generate threats to the White King, and it is difficult to believe that this move can already be a decisive mistake. Either 16...Nf6 or 16...Rc8 was necessary.

Karpov is often spare with his analysis and doesn't explain why those two moves are better. We have to work it out for ourselves. He played 17.Ne2, and awarded himself a '!':

In the event of 17.Kb1 a5, Black obtains fair counterchances. However, the modest Knight retreat to e2 is extremely unpleasant for Black. The Knight deals with the problem of ousting the Queen beautifully, and simultaneously joins in the attack on the Kingside.

Now the game continued 17...Qc5 18.g4 Nf6 19.g5 Nh5 20.Rxh5 gxh5 21.Rh1 Qe3+ 22.Kb1 Qxf3 23.Rxh5 e6. Here he spent a page and a half explaining why 23...Ng6 also lost. Contrary to popular opinion, he was a wonderful tactician.

To play through the complete game see...

Anatoli Karpov vs Evgeny Gik, 05, Moscow ch-stud 1968
http://www.chessgames.com/perl/chessgame?gid=1067442

...on Chessgames.com.

27 November 2007

A Defensive Sacrifice Leads to Attack

In What to Do with Passive Rooks?, I left Petrosian's notes at the diagrammed position, where the ninth World Champion commented, 'A Rook, by no means forced, goes to a square attacked by a minor piece.' See the previous post for the link to the PGN and Java game viewer on Chessgames.com.

The game continued 32.Bxf4, which Petrosian judged inferior.

If Tal realized all the consequences he would be satisfied with a gain of a Pawn: 32.Rxf4 exf4 33.Bxf4 Bxf4 34.Qxf4 Qe7. Black would be a Pawn down, but the position quite unclear. His Knight would be able to go to e5, the Pawn d5 would be stopped. I thought this position would be better than a cramped position with a material balance.

Now after 32...exf4, White tried 33.Nd2, 'The Knight is the only White piece that can fight for e5, so Tal wants to move it to f3. Perhaps 33.Nc1 & 34.Nd3, with the same idea, would be better.'

Riga 1958
Petrosian, Tigran

Tal, Mikhail
(After 31...Rf6-f4)
[FEN "3q1rk1/3n1ppp/p2b4/P1pPp2P/1pP1PrQ1/1N2B3/1P4P1/R4RK1 w - - 0 32"]

Note how Black's moves suggest themselves, but White has decisions to make. 33...Ne5. Black committed to the sacrifice because it gave optimum play for the remaining minor pieces. 34.Qxf4.

White is not forced to capture this Pawn. He could play e.g. 34.Qe2. Then Black would have a number of possibilities: 34...g5 & 34...Qh4. It is hard to say that White's extra exchange would be tangible. Tal realized that events were taking a bad turn for him, so he tried to complicate matters.

34...Nxc4 35.e5 Nxe5. Petrosian: 'By means of counter sacrifices White has opened files for his Rooks. However Black has plenty of counter chances.' 36.Ne4 h6 37.Rae1 Bb8 38.Rd1. 'Those who want to to practise calculating are advised to study this game starting from the 38th or 39th move. There are a lot of interesting variations.' 38...c4 39.d6.

White is already faced with great difficulties: the threat is 39...Ba7+ followed by 40...Nd3 with attack against his King. Moreover, when the Knight comes to d3, the White Rook is cut off and the Pawn d5 is in danger. Tal seeks defending resources.

39...Nd3 40.Qg4 Ba7+ 41.Kh1 f5. The sealed move; if 42.Rxf5, then 42...Rxf5 43.Qxf5 Qh4+ 44.Qh3 Qxe4. 42.Nf6+. Now after a number of exchanges, Petrosian noted, 'Black has good winning chances, but I failed to exploit them, and the game ended in a draw.'

25 November 2007

Soviet Women Players

Women chess players often present special difficulties in any survey of historical chess players. The techniques used to select male players don't always work for female players. Why not? Because women frequently play in events for women only. For example, in this project on Soviet players, one of my criteria for inclusion is participation in the finals of a Soviet Championship. Women never played in any of these events.

My other start point, the lists included in Kotov & Yudovich's 'Soviet School of Chess', resulted in eight women being included. As the book was published in the 1950s, later players were not mentioned. To fill this gap I first turned to 'Women in Chess' by John Graham. Starting with Nona Gaprindashvili, this gave me the names of six Soviet players (many Georgian) who competed in international events.

Another reference, which I've just started to study, is 'Soviet Women in Chess' by Elizaveta Bikova (Fizkultura i Sport, 1957).


This book, a treasure trove of info about early Soviet women's events, has crosstables for

  • Soviet Championships
  • Russian (RSFSR) Championships,
  • Moscow Championships,
  • Leningrad Championships, and
  • early FIDE Championships.

The book also has photos and short biographies of the best known players. The biggest problem is understanding; struggling with the Russian language gives me excellent insight into what it must be like going through life as a semi-literate.

23 November 2007

A Game of Chess with Marcel Duchamp

This clip is French language with English subtitles. "There is no solution, because there is no problem."


A Game of Chess with Marcel Duchamp - Part 1 (6:22) • Jeu d'echecs avec Marcel Duchamp; French TV documentary (Interview, 1963)

All parts: Part 1Part 2Part 3Part 4Part 5Part 6Part 7

21 November 2007

Most Sacrifices Are in the Notes

While working on my latest Every Move Explained, 2007 Barcelona - Krasenkow vs. Nakamura, I encountered the position shown in the diagram. Nakamura played 11...c5, an excellent move.

The tactical justification for the move was 12.dxc5 d4 13.Na4 bxc5. Now if 14.e5, then 14...Nxe5 15.Bxa8 Qxa8.

Barcelona 2007
Nakamura, Hikaru

Krasenkow, Mikhail
(After 11.Nf3-d2)
[FEN "r2q1rk1/3nbppp/bpp1pn2/p2p4/2PPP3/1PN3P1/P1QN1PBP/R1B2RK1 b - - 0 11"]

As I explained in my own notes to the game,

The Black Queen and light-squared Bishop would then operate unopposed on the a8-h1 diagonal, putting the White King at considerable risk, while the White Rooks would lack an open file to break into the Black position.

The same sacrificial theme played an important role over the next few moves, although it was never played in the game. It reminded me of the old saying that most sacrifices are in the notes to the games. Good players don't let their opponents sacrifice too easily.

To play through the complete game see...

Mikhail Krasenkow vs Hikaru Nakamura, Casino de Barcelona 2007
http://www.chessgames.com/perl/chessgame?gid=1477101

...on Chessgames.com.

19 November 2007

What to Do with Passive Rooks?

Continuing with Petrosian's Exchange Sacrifice, what opening led to the diagrammed position? I assumed it was some sort of 1.d4 opening and was surprised to see that the game started as the Chigorin Variation of a Closed Lopez. Petrosian introduced the game with

Every chess player has memorable games which are especially precious for him. My game with Mikhail Tal (the 25th USSR Championship, 1958) is memorable for me as a creative achievement rather than a sporting success. Some chess players are proud of almost every game they have played; some have enough self-criticism. I must say that as a rule, I am seldom satidified with my own play. The game with Tal is one of those which have brought me pleasure, due to a successfully performed idea.

As with the other Petrosian games looked at so far in this series, the motivation for the exchange sacrifice was a poor position. The players have just exchanged Bishops on a4. Petrosian commented:

White has a great positional advantage. He practically has an extra passed Pawn d5. Right now, it is not so important because it can be blockaded at d6, d7, or even d8, therefore it is not directly dangerous. But when the game will simplify to an endgame, the passed, well protected Pawn can be decisive. How should Black defend his position?

He then warned that passive play wouldn't work because of the inferior activity of the Black Rooks. After Tal played 25.Qf3, Petrosian answered 25...Rd6, and noted

This move seems strange. According to strategical principles, the stronger the blockading piece, the less it fits this role. It it is a Queen, it must move away if attacked by any other piece. A Rook is uncomfortable being attacked by a minor piece, but my idea was somewhat different.

Petrosian has already given two important positional ideas: 1) the danger in a cramped position is reduced activity for the Rooks, and 2) the Rook is a poor blockading piece.

Riga 1958
Petrosian, Tigran

Tal, Mikhail
(After 24...Rbd8)
[FEN "3r1rk1/2q1bppp/p4n2/P1pPp3/RpP1P3/4B2P/1P1N2P1/3QR1K1 w - - 0 25"]

The game continued 26.Nb3 Nd7 27.Raa1 Rg6.

This is the idea invented and beloved by me. Black foresees that his Rooks, being left 'at home' would be too inactive and 'drags out' one of them to supply it with active functions. I think the Rook stands well enough on g6.

28.Rf1 Bd6 29.h4 Qd8 (Petrosian: 'I could have played 29...Rf6, exchanging the Rook, but this was not my idea.') 30.h5 Rf6 31.Qg4 Rf4 (Petrosian: 'Today I would take on f1, and the result would be a draw or a loss. In 1958, my mind worked some other way.')

Having reached the sacrifice, I'll walk through the continuation on another post. To play through the complete game see...

Mikhail Tal vs Tigran Vartanovich Petrosian, Riga 1958
http://www.chessgames.com/perl/chessgame?gid=1106392

...on Chessgames.com.

17 November 2007

Soviet Players, Reality Check

As documented on Soviet Players on Chessgames.com, I determined that nine out of 359 players are missing from my two initial sources of biographical data. Their names are: Chersakov, A.; Kliavin, P.; Kozlov, Vadim; Mikliaev, I.; Mund, A.; Pavlov Pianov, Nikolay; Selezniev, V.; Shamis, Alexey; and Skotorenko, N.

How did these names end up on my list? Three players are mentioned in an appendix to Kotov & Yudovich titled 'Soviet Masters': Chersakov,A. (Leningrad); Kliavin,P. (Riga); and Skotorenko,N. (Kemerovo). Double checking Gaige, I confirm that all are missing. There is, however an entry for 'Skotorenko, Vasily Grigorievich, b.1927 at Kremenchug', which might be the same as 'Skotorenko,N. (Kemerovo)'. Confirmation is needed.

The other six players are listed in my PGN file of Soviet Championships. I received this file in 2001 from another fan of historical chess games. The file is well prepared except for two thousand duplicate games. These need to be eliminated to make the file even more useful.

Two of the names on my list played in the early Soviet Championships: Mund, A. (URS-ch01 1920), and Pavlov Pianov, Nikolay (URS-ch01 1920, URS-ch05 1927, and URS-ch06 qf2 1929). Cafferty & Taimanov (C&T) list Mund and Pavlov-Pyanov for the 1920 and 1927 events, with no additional info. They don't list Pavlov-Pyanov in 1929, where the semifinal and final events are mentioned without the quarterfinal ('qf') events.

The other four players all played in URS-ch35 1967. Why so many from one event? C&T explain

Karkov in the Ukraine was the venue for the 1967 event proper, which was dedicated to the 50th anniversary of the October Revolution. It was decided that a return to mass activity was called for and the experiment of a Swiss system for 130 players over 13 rounds was tried.

A quick calculation gives

65 games per round * 13 rounds = 845 games

but the PGN file has only 152 games. These games mention only 72 players, so there is a real problem with my data here. Since URS-ch34 (in 1966) and URS-ch36 (1968) had only 21 and 20 players respectively, it makes no sense to have 72 (or 130) players from a single year. Furthermore, C&T continue

One reason to let in so many players was that 1967 had seen rare international tournaments in Leningrad and Moscow (a star-studded event, won by Stein) as well as other events internally and the Sousse Interzonal, so many of the big names would be taking a rest.

C&T list the final places and scores for the four players as: 41-57 places 7.0 points I.Miklyaev; 71-88 6.0 A.Shamis, V.Kozlov; and 89-101 5.5 V.Seleznev. Looking at the other players listed by C&T, 17 finished with 8.5 points or better, while 26 finished 8.0 or better. I should exclude players that finished lower in the cross table.

That gives me three actions for subsequent steps:

  • Locate one or more source that include Chersakov, Kliavin, Skotorenko, Mund, and Pavlov-Pyanov.
  • Remove the duplicate games from the file of the Soviet Championships.
  • Exclude certain players from URS-ch35 (1967).

It's always useful to do periodic reality checks when working with data from disparate sources.

15 November 2007

Opening Exchange Sac [D89]

The position in the diagram arises after the moves 1.d4 Nf6 2.c4 g6 3.Nc3 d5 4.cxd5 Nxd5 5.e4 Nxc3 6.bxc3 Bg7 7.Bc4 c5 8.Ne2 O-O 9.O-O Nc6 10.Be3 cxd4 11.cxd4 Bg4 12.f3 Na5 13.Bd3 Be6. The first time I encountered it was as White in a pre-Internet correspondence game where I had opened 1.d4 instead of my usual 1.e4.

Gruenfeld D89
Black

White
(After 13.Bg4-e6)
[FEN "r2q1rk1/pp2ppbp/4b1p1/n7/3PP3/3BBP2/P3N1PP/R2Q1RK1 w - - 0 14"]

White has the choice between 14.d5, sacrificing the exchange, and 14.Rc1. I spent hours analyzing the sacrifice, couldn't get a grip on which side was better, finally chickened out, and played 14.Rc1. The game was eventually drawn.

I sometimes wondered what verdict had finally been delivered by theory on the variation. I was surprised to find that it was the subject of a recent article on Chess.com, An exchange sacrifice to beat the Grunfeld.

You probably know the most recent games where this moves order was played: Topalov - Shirov (Wijk aan Zee 2007), Aronian - Shirov (Elista WCM 2007). But what about the oldest ones? Some are really brilliant and I'm going to show you one of them.

The stem game for the 14.d5 variation appears to have been played in 1950. How is it possible that after almost 60 years and hundreds of games, the sacrifice is still being explored? The first of the games mentioned by the Chess.com article is at...

Veselin Topalov vs Alexey Shirov, Corus 2007
http://www.chessgames.com/perl/chessgame?gid=1444665

...on Chessgames.com.

13 November 2007

Watson's Example Exchange Sacs

A resource complementing Petrosian's Exchange Sacrifice, is a chapter in John Watson's Secrets of Modern Chess Strategy, titled 'The Exchange Sacrifice'.

It is instructive to search pre-1930 databases for successful exchange sacrifice. Almost without exception, we find compensation only from direct mating attacks and/or the immediate acquisition of at least two Pawns for the exchange. Long-term sacrifices are seldom seen; one feels that this simply must reflect a pre-modern attachment to static material evaluations. There are nevertheless a few precursors of the modern attitude. Tarrasch himself, in annotating a game between Janowski and Lasker from 1909 (won by Janowski, the exchange down), commented that a Knight in the middle of the board, protected by a Pawn and out of the range of any enemy Pawn, is nearly as strong as a Rook.

The Lasker - Janowski game was played in the May 1909 non-title match to determine if Janowski was a worthy challenger for a title match. The four game match ended tied with two wins each and the two players met again in a longer match in October, although still not for the title. The exchange sacrifice starts from the position shown in the diagram.

Janowski [Janowsky] played 47...Qh6, which looks like a blunder. White wins the exchange with 48.Ng4 Qh7 49.Ngxf6+ Rxf6 50.Nxf6+ Rxf6.

Paris, May 1909, Game 2
Janowski, David

Lasker, Emanuel
(After 47.f4-f5)
[FEN "5rk1/2p2r2/p1Pp1bq1/1p1PpP1p/1P1nN3/3RN1PQ/P7/3R3K b - - 0 47"]

The game continued 51.Rf1 Qf7 52.Qg2 Rxf5 53.Rxf5 Qxf5, when Black had recovered one Pawn for the exchange. A few moves later Black also won the b-Pawn, but had to sacrifice the h-Pawn to keep the files closed on the Kingside. Black eventually forced a passed Pawn on the Queenside by trading a pair of Pawns on the a-/b-files. After the Queens were swapped off, the Rook was unable to cope with the Knight and passed Pawn. To play through the complete game see...

Emanuel Lasker vs David Janowski, Paris 1909
http://www.chessgames.com/perl/chessgame?gid=1064704

...on Chessgames.com. Watson used five of Petrosian's examples and added 12 of his own:-

1921 Triberg, Selesniev - Alekhine
1922 Teplitz-Schonau, Treybal - Spielmann
1945 USSR Chp, Tolush - Botvinnik
1943 Moscow, Ljublinsky - Botvinnik
1943 Moscow, Panov - Simagin
1950 Moscow, Bondarevsky - Mikenas
[Petrosian games]
1991 Wijk aan Zee, Seirawan - Kozul
1960 Leipzig, Gligoric - Tal
1981 Moscow, Beliavsky - Kasparov
1993 Linares, Karpov - Gelfand
1983 Barcelona, Martin Gonzalez - Dolmatov
1996 Dos Hermanas, Ivanchuk - Kramnik

11 November 2007

Soviet Players on Chessgames.com

Just as I did for Soviet Players in Gaige, I checked all 359 names on my list of Soviet players against the Chessgames.com Chess Player Directory. I found 53 names without corresponding player pages. This is better than I expected to find.

Matching these with names not found in Gaige, I have nine unknown names. I'll tackle these next time.

More useless information: of players on the database with patronyms, the most common, with 10 or more each, are Nikolaevich, Mikhailovich, Alexandrovich, Ivanovich, and Yakovlevich. If I were Russian, would I be Mark Bobovich, Mark Robertevich, or something else?

09 November 2007

How to Make a Chess Video

The link to this video was posted on the About Chess forum. It's worth repeating.


Letsplaychess.com presents 'How to make a chess video' (4:44) • From: Majnu2006

I honestly don't remember where I found the following link...

Produce Special Interest Videos on DVD
http://www.studio1productions.com/sivkit.htm

...'Free PDF book on Special Interest Videos Production', no strings attached, which is unusual for a marketing freebie.

07 November 2007

Chess Consulting

Chess is frequently used for business metaphors. Here is a project I helped with. Recognize the position?


Teradata Magazine | Archives [Vol.4, No.1; Jan 2004]
http://www.teradata.com/t/page/146551/index.html

***

'The history of the gold coins move (*)

'Chess enthusiasts still talk about the "Gold Coins Move." It was so bold, so unexpected and so effective that it's considered the greatest move in chess history.

'In 1912, at the 18th German Chess Congress in Breslau, U.S. champion Frank Marshall faced Russian master Stepan Levitzky. In move 23, Marshall made the most unlikely move on the board, leaving his queen vulnerable to Levitzky. Marshall sacrificed the piece so he could penetrate to the hostile king and win. Levitsky knew his best possible move would only result in a hopelessly lost endgame, and he gave up. Legend has it spectators showered the final position with gold coins, hence the name.

'Your next business move could be routine, safe, expected and altogether ineffective. Or you could surprise the competition and win the game. The choice is yours.'

05 November 2007

Opening Exchange Sac [E38]

I find the exchange sacrifice particularly attractive when it happens in the opening, by definition almost always for long term positional compensation. I found the following example in the Soviet School of Chess. Verlinsky was one of ten players that Kotov and Yudovich included in a section 'Illustrious Names', all Soviet players who had died before the book was published.

In the diagrammed position, Black was undoubtedly expecting 11.Bd2 Nxd2 12.Qxd2 Qxd2+ 13.Nxd2, with a playable game. White played instead 11.Nd2. Now if 11...Nxd2 then 12.Bxd2, and the Black Queen must retreat leaving White with the advantage.

Black played the obvious 11...Nb4, perhaps expecting the White Queen to retreat. White again found the better move 12.axb4, forcing Black to win the exchange with 12...Qxa1.

7th USSR Championship 1931
Kirillov, Vladimir

Verlinsky, Boris
(After 10...Qd8-a5+)
[FEN "r1b1k2r/pp3ppp/2n1p3/q2p4/2PPn3/P2BPN2/2Q2PPP/R1B1K2R w KQkq - 0 11"]

The game continued 13.Nxe4 dxe4 14.Bxe4 f5 15.Bd3 Bd7 16.O-O Qa4 17.Qb2 O-O and White won in the middle game. What is White's compensation for the exchange? The game's annotator wrote, 'White has only an extra Pawn, but his overwhelming superiority in the center, plus his two Bishops, give him every reason to count on victory.' To this I would add that the mobility of Black's Bishop is restricted by the e-Pawn and the Queenside Pawns will become an object of attack.

To play through the complete game see...

Boris Verlinsky vs Vladimir Grigorevich Kirillov, URS-ch07 1931
http://www.chessgames.com/perl/chessgame?gid=1272376

...on Chessgames.com. I know of a few more opening exchange sacrifices that I'll mention if I can locate the games.

03 November 2007

Soviet Players in Gaige

Continuing with Biographical Data for Soviet Players, I checked all 359 names on my list of Soviet players and found that 46 were not in Gaige. Wahrheit commented on the previous post that finding four out of 50 names not in Gaige was evidence of success. After overcoming my initial astonishment that anyone could be the least bit interested in this project, I cautioned that it was probably because Gaige's work ended before the Soviet Championships did. That seems to be the case. There are many names that I recognize -- Bologan and Kramnik are two good examples-- but who are not listed in my 1987 copy of Gaige. There are also a few names that I don't recognize. If I don't find them on Chessgames.com, then they are either obscure players or evidence of a mistake somewhere.

At this time there's not much to say about the data. The most common first names in the list, around two dozen players each, are Alexander and Vladimir. These are followed by Yury, Mikhail, Sergey, and Nikolay, with a dozen or so examples each. Boris and Anatoly also rank fairly well. This undoubtedly says more about popular Russian names than about any magic to help babies become strong chess players.

01 November 2007

Understanding Petrosian

Continuing An Obvious Positional Advantage, the diagram shows the position after the final move in the previous post. Petrosian considered that 'Black undoubtedly had the edge'. It doesn't matter how long I play chess or how much I study, I will never understand Petrosian's approach to the game. Black is an exchange down in the diagram, and Black is still better?

White continued 27.Be2. After 27...Bh6, Petrosian noted, 'Not the best. Black should have played 27...h5 first.' This is typical Petrosian; there is no explanation why ...h5 is better. Because the Bishop is sheltered against an attack on the h-file? Because the Black King gets a new escape square on h7? Because an eventual g2-g4 is prevented? All of the above or something else?

San Antonio 1972
Petrosian, Tigran

Portisch, Lajos
(After 26...Ne7-f8(xB))
[FEN "4rnk1/1bq3bp/1pNp2p1/pPnPpp2/2P5/5B2/P4PPP/1NRQ1RK1 w - - 0 27"]

The game continued 28.Rc2 Bc8 29.Nc3 Nfd7 30.Re1 Nf6 31.Bf1 f4. Of his last move, Petrosian wrote:

Here I was a bit hasty. Such moves require great caution. The e5-f5 Pawn pair has become less mobile, and the Pawn e5 can be blockaded. Naturally I had taken into consideration that my pieces (Nc5, Nf6, eventually Bf5) kept the square e4 under control, so I hoped to play e5-e4 safely. Of course, I should have taken some prophylaxis like 31...Kh8.

Considering the further course of the game, it is not clear why ...f4 is worse here or ...Kh8 is better here than in a move or two. What difference does it make? 'Of course, I should have taken some prophylaxis'; why 'of course'?

There followed 32.Rce2 Rf8 33.Na4 Nxa4 34.Qxa4 Nd7 35.Ne7+. Petrosian: 'I had overlooked this simple move.' Now the game was drawn after 35...Kh8 36.Nxc8 Qxc8 37.Qa3 Nc5 38.Qf3 Qf5 39.h3 1/2-1/2.

From this I understand that 34...Nd7, because it allowed White to exchange one of the dangerous Bishops, was an error. What move was better? 34...Bf5, or something else? Petrosian doesn't say. Three times the former World Champion criticizes his own play and three times I am left to wonder why. What a difference between his level and mine!