Thursday, April 30, 2009

Great Rivalries: Petrosian vs. Larsen (Part 5)

Petrosian and Larsen met for the third time at the Nimzowitsch Memorial in Copenhagen. Petrosian dominated the tournament, scoring 11.5/14 against the largely weak opposition. Larsen could finish only fourth, behind Petrosian, Geller, and Stahlberg.

Tigran Petrosian-Bent Larsen
Copenhagen (Nimzowitsch Memorial) 1960

I am not sure about the order of moves at the beginning of this game. Chessbase gives a completely different sequence from that in P.H. Clarke's Petrosian's Best Games of Chess 1946-1963. In any event, Petrosian gained a large advantage with a menacing pawn majority on the kingside after some passive play from Larsen in another Old Indian. The position after 22 moves was:

There followed:

23. Qh5 Rxd3 24. Rxd3 Rd8

25. Nf6+!

Friday, April 24, 2009

Great Rivalries: Petrosian vs. Larsen (Part 4)

Petrosian and Larsen next met over the board in 1960. In 1959, Petrosian had won the 26th Soviet Chess Championship and had finished third in the Candidates' Tournament, behind Tal and Keres. Larsen had been left with the somewhat humiliating task of serving as the second to the fifteen-year-old Fischer in Yugoslavia, a task that reportedly involved reading Tarzan stories to the teenager.

In the small tournament in Beverwijk, Holland, in 1960 (the predecessor of the great Wijk aan Zee and Corus tournaments of more recent years), Petrosian and Larsen tied for first place. But Petrosian prevailed in their head-to-head encounter.

Tigran Petrosian-Bent Larsen
Beverwijk 1960

1. c4 d6 2. d4 e5 3. Nf3 Nd7 4. Nc3 Ngf6 5. e4 Be7 6. Be2 O-O 7. O-O c6

This is the main line of the Old Indian. White has a small advantage, as Black's position is more passive than in the analogous lines of the King's Indian, thanks to the bishop's position on e7 instead of g7. Nevertheless, Black's position is still very solid.

8. d5

Not the most common move--in fact, Re1, Qc2, Be3, h3, and Rb1 are all seen more often in this position--but perhaps played by analogy to the Petrosian variation in the King's Indian? Petrosian closes the center, stabilizing the position and reducing Black's options. In this position, though, Black is able to establish his knight on c5 and White does not seem to gain any advantage.


Keene and Simpole, in Petrosian vs the Elite, mention that 8...c5 is a viable alternative. That move would transpose into a well-known line of the Czech Benoni, but a tempo down for Black. Of course, in such a closed position, one tempo is not particularly important. Still, White should have the edge after 9. a3 Ne8 10. b4.

9. Nd2 a5 10. b3 cxd5?

Keene and Simpole prefer 10...Ne8. That move is certainly possible, but Black's problem is not so much that he has the wrong plan but that he uses the wrong move order. Thus 10...Bd7 11. Rb1 cxd5 12. cxd5 b5! gives Black a good position. Indeed, Larsen would later use this idea when he had the position after 10.b3 a tempo down in a game against Polugaevsky in Bugojno 1982. That game went 11. a3 Bd7 12. Rb1 cxd5 13. cxd5 b5 and Black had no problems.

11. cxd5 Bd7 12. a4!

Now ...b5 is impossible and White has a powerful bind on the queenside light squares.

12...Qb6 13. Ba3 Rfc8 14. Rb1 Qa7 15. Rc1 Rc7 16. Bb5

Threatens to pave the way for an invasion of the light squares by exchanging off a key defender of them, the d7-bishop. If, for example, 16...Ne8 17. Bxd7 Nxd7 18. Nb5 Rxc1 19. Qxc1 Qb8 20. Nc4 and Black is completely tied up. Larsen avoids the exchange and also leaves the White bishop blocking the knight's access to b5.

16...Bg4 17. Qe1 Na6 18. h3 Bh5

This removes the bishop from the defense of the queenside, but 18...Bc8 19. Bxa6! bxa6 (19...Qxa6 20. Nb5 Rxc1 21. Qxc1 Qb6 22. Nc4 Qd8 23. Bxd6 wins) 20. Nc4 leaves White with a dominating position. 18...Bd7 19. Bxd7 Rxd7 20. Nb5 Qb8 21. Nc4 is the sort of thing Larsen was trying to avoid with 16...Bg4.

19. Bxa6!

At first sight, a surprising move, but it clears the way for the knights to occupy c4 and b5. If 19...bxa6, then 20. Nc4 wins the a5-pawn.

19...Qxa6 20. Nc4

Keene and Simpole don't analyze any alternatives to Larsen's exchange sacrifice. Petrosian has established such an effective bind, however, that it is hard to come up with ideas for Black. For example, after 20...Ne8 21. Nb5 Rd7 22. Qe3 Bg6 Black is badly tied up, though I don't see an immediate breakthrough for White. Nevertheless, Petrosian is the last player against whom you would want such a position, so Larsen's decision to lash out, even if his compensation for the exchange is ultimately inadequate, is understandable.

20...Rxc4 21. bxc4
Qxc4 22. f3

White could also try 22. Nb5 Qxa4 23. Bxd6 with a powerful passed d-pawn, but Petrosian's decision to keep things under control is also strong, and characteristic of his play.

22...Ne8 23. Qe2 Qd4+ 24. Kh1

Black has a pawn for the exchange, but has no real positional compensation. He has a temporary initiative on the dark squares, but once White's bishop returns to c1, he will have all of the important dark squares covered. White still has control over the queenside light squares, which he will use to organize an invasion along the b- or c-file.

24...Bg5 25. Rfd1
Qb6 26. Rb1 Qc7 27. Nb5

The establishment of a knight on b5 is a key part of White's strategy. It helps ensure his invasion on the c-file and ties Black to the defense of the d6-pawn. Incidentally, in Petrosian's King's Indian games, where he frequently used the same sort of light-square strategy, I count 26 games in which he landed a knight on b5. In those games, he scored a staggering +22 =4!

28. Bc1 Bd8 29. Be3

Petrosian has beaten back Larsen's attempt at counterplay on the dark squares. Now Larsen is helpless to prevent Petrosian's invasion down the c-file.

29...Bg6 30. g4!

Eliminating any chance of counterplay with ...f5.

30...h5 31. Rbc1 Nf6

If Black exchanges off the dominant White knight with 31...Nc7, White remains in control after 32. Nxc7 Bxc7 33. Qc4 Bd8 34. Rd2 intending Rb2.

32. Rc4 Nh7

Still angling for some counterplay on the kingside dark squares.

33. Rdc1 Bg5 34.
Bxg5 Nxg5 35. Kg2 Qd8 36. Qe3!

Another characteristic "short" move from Petrosian. 36. Rc7 was also completely winning after 36...hg 37. hg f5!? 38. gf Bh5 39. R1c3 or 38. ef Bf7 39. Qd3, but Petrosian does not allow Black even a hint of counterplay. Now if 36...hg 37. hg f5? 38. ef Bf7 39. Nxd6 and Black is crushed. White is also threatening 37. Nxd6, which forces the following weakening of the seventh rank.

36...f6 37. Rc7 Kh7

If 37...Rb8, then 38. Qa7.

38. Rxb7 Rc8 39. Rxc8 Qxc8 40. Rc7
Qf8 41. Qc1 Qd8 42. Qc6 Nxe4 43. Rc8

43. fe Bxe4+ 44. Kh2 also wins, but is not as elegant as Petrosian's choice.

43...Qe7 44. Qa8 Kh6 45. Rc7 1-0

The queen is trapped. A beautiful game by Petrosian.

A Prolegomena to a Theory of Drawing Lines (Part 1)

There are always plenty of reasons to agree to a short, agreed or semi-agreed, "grandmaster" draw. Perhaps you are tired. Or looking for an easy day after a hard defeat. Or preparing to face a difficult opponent the next day. Maybe the round is starting too early in the morning. Or maybe the Real Madrid game is on TV. There are also more sinister reasons.

In any event, the "whys" of grandmaster draws are almost infinite. But that still leaves the question of "how." Even if you and your opponent both know that you are going to agree to a draw, you still need to go out there and play some moves. If you try to agree to a draw on move one, the organizers might not like it.

What follows is my attempt at a survey of a little-studied branch of chess opening theory: the theory of drawing lines. It turns out that there are a great variety of standard lines that lead to short draws, either by producing a perpetual check or a sterile position. Some players even specialize in certain lines.

I propose that it doesn't make sense to classify these openings according to the usual ECO code or the common names of openings (Queen's Gambit, King's Indian, etc.). Instead, I suggest a subjective classification based on the style of "game" that the opening produces. After all, given that the objective result of all of these games is the same--a draw--only the subjective, aesthetic aspects of the openings are relevant criteria for classification.

Symmetrical Exchanging Lines

These lines are perhaps the simplest way to agree to a draw: produce a symmetrical position with no imbalances, trade off some pieces, and sign the score sheets--what could be easier?

  1. The Exchange Slav. Perhaps the most popular drawing line of all. As with so many of these drawing lines, as we shall see in a later post, White can actually play for the win in the Exchange Slav, but if he wants a draw, then he can choose the following line: 1.d4 d5 2.c4 c6 cd 4.Nc3 Nf6 5.Nf3 Nc6 6.Bf4 Bf5 7.e3 e6 8.Bd3 Bxd3 9.Qxd3 Bd6 10.Bxd6 Qxd6. In my database, 190 games ended as draws in this position. After 11.O-O you get another 105 draws; 11...O-O 274 more 12.Rac1 84 more (alternatively, 12.Rfc1 113 more draws) and so on. In total, I count at least 1335 draws in twelve moves or fewer in this line. Some players are connoisseurs of this variation. GM Ognjen Cvitan, for example, has "played" 36 draws of no more than 16 moves in the Exchange Slav. Other avid practitioners include Ionescu (36 draws), Panchenko (27), Milorad Knezevic (23), Vadasz (23), Naumkin (22), and Kirov (20).
  2. The Symmetrical English (1). Another symmetrical drawing line goes: 1.c4 c5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7 5.Nf3 Nf6 6.O-O O-O 7.d3 d6 8.a3 a6 9.Rb1 Rb8 10.b4 cb 11.ab b5 12.cb ab. This line is not as popular as the Exchange Slav--there is actually a lot of play left in the position--but nevertheless I count at least 170 early draws in this line. Hulak has 9 draws in this line. Our friend Cvitan has 3. (Indeed, if you want to learn the theory of drawing lines, you can do worse than to study Cvitan's games.)
  3. The Symmetrical English (2). 1.c4 c5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7 5.e3 e6 6.Nge2 Nge7 7.O-O O-O 8.d4 cd 9.Nxd4 Nxd4 10.ed d5 Nxd5 12.Nxd5 ed. I find 62 draws in this position, with dozens more on the moves immediately before or after it, including the classic Hulak-Cvitan, Sarajevo 1988, which continued 13. Be3 Be6 14. Qd2 Qb6 15. Rac1 Rfc8 16. h3 h5 17. Rfd1 1/2-1/2 Hulak is again a frequent player, with at least 12 draws, although he does not match Csom's 19 games in this line. Even Tal played 4 short draws (as Black) from this position.
Incidentally, Csom might be the all-time champion of grandmaster draws. My database has 643 Csom draws of fewer than 20 moves. Cvitan is close behind with 588. By comparison, such supposed "draw masters" as Kramnik and Leko have only 166 and 103 such draws, respectively.

In the next installment, I will turn to the always exciting draw by perpetual check.

Tuesday, April 21, 2009

Great Rivalries: Petrosian vs. Larsen (Part 3)

Let's look at some of the Petrosian-Larsen games.

Their first encounter was in the 16th round of the Portoroz Interzonal in 1958, an event whose outcome Larsen described as "the greatest failure of my chess career." He ended up finishing 16th, with 8.5/20. Petrosian lost only one game in the whole tournament and tied for third with Benko, with 12.5/20, thereby qualifying for the Candidates' Tournament. But his one defeat came in his game against Larsen.

Before this game, in the 16th round, Petrosian was leading the tournament. Larsen was near the bottom of the table.

Larsen opened with the Bird (1.f4), a bit of a provocation against Petrosian, who was already establishing a fearsome record against the Dutch Defense as white. According to my (surely incomplete) database, his record at the time was +5, =5, and would eventually grow to +15, =7, including wins over Korchnoi, Bondarevsky, Tolush, Bronstein, Nikolic, and Larsen himself. the Korchnoi game in particular, played when they both were juniors, is a classic demolition of black's old-fashioned Stonewall strategy (7...O-O? instead of 7...Qe7).

Petrosian's disdain for the Dutch is captured by remarks such as "If they want to play the Dutch, there is no reason to prevent it!" and "What a delight! I love playing against the Dutch." (The latter is from Petrosian's annotations to his game against Larsen (!) from the wonderfully-named Church's Fried Chicken First International Chess Tournament, in San Antonio, in 1972. I will come to this game later.

Anyway, to the game:

Bent Larsen - Tigran Petrosian
Portoroz Interzonal (16) 1958

1. f4 Nf6 2. Nf3 d5 3. e3 Bg4

Larsen later wrote that he had intended to play 3...g6 4.b4!? in this game.

He got his chance to play this novelty against Simagin at the Alekhine Memorial in 1959, a 76-move draw. He subsequently employed it four more times, most prominently in his win over Spassky at the 1964 Interzonal.

4. Be2 Nbd7 5. Ne5 Bxe2 6. Qxe2 e6 7. O-O Bd6 8. d4 O-O 9. Nd2 c5 10. c3 Rc8

It is hard to believe that White has any advantage here. It is a standard Stonewall Dutch position in reverse, in which Black has succeeded in exchanging off his bad bishop. Larsen, ever the optimist, nevertheless decides to start pushing his kingside pawns.

11. g4 A committal move. White would like to develop the dark-squared bishop, but 11.b3? cxd4 leaves White with problems on the c-file, while 11.Ndf3 Ne4 looks slightly better for Black.

11...Ne8 12. Ndf3 Ndf6
Perhaps inviting a repetition after 13.Nfd2 Nd7.

13. Bd2 Ne4?!
This move leaves Black's other knight without prospects. Better seems to be 13...Bxe5 14.Nxe5 Ne4 15.Be1 f6 16.Nd3 N8d6, with an advantage for Black.

14. Be1 f6 15. Nd3 Be7 16. f5!? Continuing Larsen's risky play Qd7 17. Nh4 Nc7 18. Bg3 Bd6 19. Bxd6 Qxd6 20. Rad1 Rce8 21. dxc5 Nxc5 22. Nf4 b5? Black is fine after 22...e5 23.Nxd5 Nxd5 24.c4 Nc3 25.bc Qc6, but now White gets a strong initiative.

23. Qg2 Rd8 24. b4!
Removing a defender of e6 Ne4 25. fxe6 g6 26. Nxd5 Nxd5 27. Qxe4 Nxc3 28. Rxd6 Nxe4 29. Ra6 Rd3 30. Rxa7 White has a clean extra pawn. Re8 31. Ng2 g5 32. Rc1 Rd6? This allows the other rook to invade the seventh rank, which will prove fatal for Black. 32...Nc3 was better. 33. h4! Rdxe6 34. Rcc7 Nd6? 34...h6 was necessary, not only to save the pawn, but to prevent what follows 35. Rxh7 Rc8 36. Rhg7+ Kh8 37. h5 Ne8 38. Rgf7 Nd6 39. Rh7+ Kg8 40. h6 Ne8 41. Rhf7 1-0

Black has to give up a piece to avoid mate. (41...Re4? 42.h7+ Kh8 32.Rf8#) Active play by Larsen, but uncharacteristically weak defense from Petrosian.

After this game, Petrosian closed the tournament by drawing four games in a row, thereby conceding first place to Tal.

Friday, April 17, 2009

Great Rivalries: Petrosian vs. Larsen (Part 2)

Back to Petrosian and Larsen.

After playing over Larsen's fine victories over Petrosian in Santa Monica (which I will discuss below), a few questions occurred to me:

1. Had any other sitting world champion ever lost two games to the same opponent in a tournament?

2. Had Petrosian, who was so notoriously difficult to defeat, on any other occasion lost two games to the same opponent in a tournament?

3. What was Larsen's lifetime score against Petrosian? Did he always displace such dominance?

1. I found the answer to the first question quickly enough, having a vague memory that Alekhine had lost twice to somebody in AVRO 1938. Sure enough, Fine won both of his games against Alekhine there. Are there any other examples?

2. It also turned out that Larsen was not the only person to win two games against Petrosian in a single tournament. Taimanov had done it in the 1953 Candidates' Tournament, as had Olafsson in the 1959 Candidates' Tournament.

3. Finally, looking up the lifetime score between Petrosian and Larsen revealed an extraordinarily interesting and combative series of contests. Petrosian won the series, +9, -4, =9. These were hard-fought games. They did not draw a game until their tenth encounter. Only 4 of their first 17 games were draws.

To provide some perspective: in the period between 1958 and 1979, when they played those 13 decisive games out of 17 (only 24% draws), Larsen overall had 30% draws and Petrosian had 56% draws.

Great Rivalries: Petrosian vs. Larsen (Part 1)

I have been reading Larsen's game collection and recently looked at his two victories over Petrosian from the great Second Piatigorsky Cup tournament in 1966.

Incidentally, my copy of this book, the 2003 reprint by Hardinge Simpole, is called "Bent Larsen's Best Games of Chess: Master of Counter Attack." This book was originally published, in Danish, as "50 Udvalgte partier, 1948-1969," in 1969. I believe that a literal translation of this title is "50 Selected Games, 1948-1969." An English translation, called "Larsen's Selected Games of Chess, 1948-1969," appeared the next year. Batsford then reissued it as "Bent Larsen: Master of Counter-Attack: Larsen's Selected Games of Chess 1948-1969," in 1992.

I mention all of this because on page 99, Larsen writes the following: "It is not by chance that the title of this book makes no reference to "best games"--for where do you find two chess players who agree which games are best? Which do you rate highest, courage or foresight? Imagination or precision? Poetry or prose? Music or mathematics?" All of which sounds ridiculous when the book you are holding in your hands in fact does make reference to "best games" in the title. Does a publisher really think that it will sell more copies of a collection of "best games" than of "selected games?"

While I'm on the topic, what is with these "Master of ..." books? The Batsford Larsen book was the predecessor of the recent Everyman Chess series, which includes:

Alexander Alekhine: Master of Attack
Boris Spassky: Master of Initiative
Rudolf Spielmann: Master of Invention

Actually, now that I look at the Everyman Chess web site, I see that the full titles of these books are in fact "The Masters: Alexander Alekhine Master of Attack," etc.

Can a great chess player be summed up in one word? If so, are these words the correct choices?

Having played through the games in the Larsen book, I submit that "Master of Counter-Attack" is a serious misnomer. I have always thought that counter-attacking in chess involved luring your opponent into over-extending himself in an attack and then responding with an attack of your own. In other words, you cannot counter-attack until your opponent has first attacked you. My favorite example of such play is Geller-Euwe, Zurich Candidates' Tournament 1953. Of the all-time greats, Lasker and Korchnoi seem to have the styles best described as "counter-attacking."

Larsen, at least in the games in this collection, is all about fighting for the initiative from the start. I'll post some examples later.

And I will also get back to the alleged topic of this post, the games between Petrosian and Larsen, later.