Yury Shulman vs the World

Yury Shulman

GM Yury Shulman

vs.

The World

The World wins a recent (2007) correspondence chess game against a grandmaster - replay the moves here

http://www.chessgames.com/perl/chessgame?gid=1443541

Congratulations to GM Shulman for his sportsmanship and for the enormous time it took from his busy schedule to participate in this game.

The time control was 2 days per move.

Section A: Analysis of the game

Section B: a Closer look at my Rybka analysis

Section C: Analysis performed by Jim Schwar using Fritz

Section A: Analysis of the game

White: GM Yury Shulman

Black: the World

Date: January-July 2007

Time Control: 2 days per move

Host Organization: chessgames.com

World Voting Format: hidden vote totals

1.d4 Nf6
2.c4 e6
3.Nf3 b6

(Queen's Indian defense)

after 3...b6

4.g3 Ba6
5.b3 Bb4+
6.Bd2 Be7
7.Bg2 c6
8.Bc3 d5
9.Ne5 Nfd7
10.Nxd7 Nxd7
11.Nd2 0-0
12.0-0 f5

after 12...f5

(12...Rc8 or 12...Nf6 are the more commonly played moves here. 12...f5 is rarely played but appears to be playable - exactly the kind of move you want for a correspondence game. Can white exploit the black weakness at e5? Can the analysts and computer operators find an outstanding move that nobody has played before?)

13.Rc1 Rc8
14.Bb2 Bd6
15.a3

after 15.a3

(Possibly too slow. 15.Nf3 is usually played here and 15.Rc2, 15.Re1 and 15.e3 are worth a look)

15...Qe7

(15...Nf6 was played in Zhao Xue - Kosintseva, 2007)

16.Rc2

(Mrugala-Nickel 2003 continued 16. b4 Nf6
17. Re1 Ne4 18. Nxe4 fxe4 19. c5 Bb8 20. Qd2 h5 21. Rc2 h4
22. Rf1 Bb5 23. Qe3 Rf6 24. Qg5 hxg3 25. hxg3 Qf7 26. Bc1 Bc7
27. Be3 Rf8 28. Qg4 bxc5 0-1)

16...Nf6
17.Qc1 f4

(Karpov-Almasi (1996) continued 17...Kh8 18. e3 Ne4 19. Rd1 Bb7 20. b4 Bb8 21. Bf1 Qe8 22. a4 a6 23. Nxe4 fxe4 24. b5 axb5 25. cxb5 Bd6 26. Ba3 Bxa3 27. Qxa3 Qf7 28. Rdc1 c5 29. dxc5 Rxc5 30. a5 Rxc2 31. Rxc2 bxa5 32. b6 d4 33. exd4 e5 34. Bc4 Qe8 35. d5 Qd8 36. Qxa5 e3 37. fxe3 Qf6 38. Qb4 h5 39. e4 Qf3 40. Qe1 Qf6 41. Qe3 Qg6 42. Rf2 Rc8 43. Bb5 Rd8 44. h4 Rd6 45. Rf8+ Kh7 46. Be8 Qg4 47. Qf3 Rxb6 48. Bxh5 Qh3 49. Qf5+ Qxf5 50. exf5 1-0)

after 17...f4

18.Nf3 Ne4

(According to Yrjola and Tella "The Queen's Indian", Gambit Publications 2003, 18...fxg3 19.hxg3 c5 =. This line did not look exciting for Black so we chose 18...Ne4 and achieved an unusual looking position.)

19.Ne5 Bxe5
20.dxe5 Nc5
21.cxd5 cxd5
22.b4 Nb3
23.Qd1 Rxc2

(23...Rc4 24.Bxd5! exd5 25.Qxd5+ Qf7 26.e6 was drawish while 25...Kh8 26.Rxc4 Bxc4 27.Qxc4 Nd2 white has two pawns for the exchange and Black cannot seem to mount any pressure on white's position)

24.Qxc2 Bc4

after 24...Bc4

(The knight on b3 looks trapped but there is no effective way to mount pressure on it. The knight presents an interesting puzzle - what will white do? WiIl white become obsessed with trapping and capturing the knight, or will he concentrate on the center, or on another objective?)

25.Rd1 Qf7

(25...a5; 25...Qe8)

26.gxf4 Qxf4

after 26...Qxf4

27.e4

(27.e3 might be better)

27...Qf7
28.exd5 exd5
29.a4

(there is -0.08/24 29.Bh3 b5 [29...Qh5] 30.e6 Qh5 [30...Qe7] 31.Bg2 h6 32.f4 d4 33.Bxd4 Nxd4 34.Rxd4 Bxe6 )

29...Qe6

(29...a6 30.b5 axb5 31.axb5 -0.13/32 Re8 32.h3 Nc5)

(-0.14/33 29...a6 30.b5 axb5 [31.axb5 Re8 32.h3 Nc5 and now sliding to here we have: (-0.16)/29 33.Qc3 Ne6 34.Qe3 Nf4 35.Kh2 Nxg2 36.Kxg2 ]

after 29...Qe6

30.Re1?

(perhaps losing - 30.Qxc4 dxc4 31.Bd5 Qxd5 32.Rxd5 Rc8 33.e6 Kf8 showed to be unclear after deep analysis, with white drawing under absolutely best play. The World team decided to risk this line as White would be under pressure to find the draw.

Comments made on the World forum:

TheDestruktor: I don't know about you. But even if this endgame proves to be a dead draw, I will enjoy playing it. At the very least, I will learn a lot about endgames.

This, of course, if he plays 30.Qxc4, which we still don't know.

Skylark: I don't know about anyone else here, but no matter who I'm playing, if I think I have even an inkling of an advantage I will play out a position to the death. And considering that at the worst it looks like we may have to go into the teft ending in which White will have to try very hard to hold the draw, I think it would be silly to forgo the possible mass amount of endgame knowledge we could all consume by pondering such a position. The World doesn't relinquish a draw without a fight; or at least I'd hope not.

a) 30.a5 Kh8 31.h3 h6 -0.43/28;

b) 30.h3 Kh8 30.a5 h6 -0.43/28; 30.b5 Kh8 31.Re1 h6 -0.56/28; and c) 30.h4 Kh8 31.Qc3 d4 -0.26/28 are alternatives that are not appealing.

After 30.Re1: -0.66/29 30...d4 31.Re4 Rc8 32.Qd1 Qf5 33.e6 Bxe6 34.Qd3 Rd8 35.Qg3 h6 36.h4 Kh8 37.a5 )

30...d4
31.Re4

(2. (-0.90/29): 31.Rd1 Rf4 32.h3 a6 33.b5 axb5 34.axb5 h6 35.Be4 Kh8 36.f3 Qxh3 37.e6 Bxe6

3. (-0.93/29): 31.Qe4 d3 32.Rd1 d2 33.f4 Kh8 34.h3 Qf7 35.Bc3 Qxf4 )

31...Rc8
32.Qd1 Qf5

after 32...Qf5

(-0.75/30: 32...Qf5 33.e6 Bxe6)

33.a5

(33.Bxd4 Bd5 34.Re3 Rc1 is bad for white. Because of various pins and threats the black pawn on d4 is now immune from capture. 33.h3 Be6 and 33.e6 Bxe6 34.Qd3 are no better for white in terms of improving the drawing chances.)

33...Be6
34.axb6 axb6
35.h3

after 35.h3

35...h6!

(35...h6 was not in the top 12 moves offered by Rybka. At one point <TefthePersian> suggested I analyze 35.h4 h6 and it was only a curiosity that made me try 35.h3 h6. White appears to be in zugzwang. At this point GM Shulman took one of his 3 allowed 1-week extensions due to his presence at a tournament. After Yury's 1 week delay he plays the following )

36.Qf3

(36.Qd3 Rd8 might have been the last chance to hold the draw, but this is unlikely as the score is now -1.53/31.

What follows is a strange looking series of moves that results in white having no choice but to weaken his position.)

36...Qg6
37.Qd1 Kh8
38.Kh2 Rf8
39.f4 Qf5
40.Kg1 Rd8

after 40...Rd8

41.Re3

(Perhaps desperation. Black has to show the correct path)

41...Qxf4
42.Rf3 Qxe5
43.Rxb3 Bxb3
44.Qxb3 Qe1+
45.Kh2 Qd2
46.b5 d3

0-1

after 46...d3

A White bishop last moved on move 14.

SectionB: A Closer look at Rybka analysis

A look at the World's computer analysis: I regret that I did not keep complete records of this game but here I will be assembling some of the final scores by the machine for each position in this game. On one particular move the machine ran for 340 hours, which is my record.

1.d4 Nf6

2.c4 e6

3.Nf3 b6

4.g3 Ba6

5.b3 Bb4+

6.Bd2 Be7

7.Bg2 c6

8.Bc3 d5

9.Ne5 Nfd7

10.Nxd7 Nxd7

11.Nd2 O-O

12.O-O f5
13.Rc1 Rc8

14.Bb2 Bd6

15.a3 Qe7

16.Rc2 Nf6

17.Qc1 f4

18.Nf3 Ne4 We are now completely out of book

19.Ne5 Bxe5

20.dxe5 Nc5 -0.07/26

21.cxd5 (0.00/25) cxd5

22.b4 Nb3

23.Qd1 Rxc2
24.Qxc2 Bc4 -0.12/29

25.Rd1 -0.04/30 Qf7

26.gxf4 Qxf4

27.e4 Qf7 -0.09/31

28.exd5 exd5
29.a4 Qe6 (29...a6 -0.13/32)

30.Re1 -0.69/28 (30.a5 -0.29/23) d4

31.Re4 -0.70/28 (31.Rd1 -0.64/28) Rc8

32.Qd1 Qf5 -0.60/31

33.a5 -0.80/28 (33.e6 -0.60/31) Be6

34.axb6 axb6 -1.01/30

35.h3 -1.56/30 h6 -1.53/31, 346 hours computer time

36.Qf3 -1.59/29 Qg6 -1.86/30

37.Qd1 -1.90/31 Kh8 -1.84/32 (31...Rf8 -1.90/31)

38.Kh2 -1.84/32 Rf8 -2.24/33

39.f4 -2.50/32 (39.Qe1 -2.24/33) Qf5 -2.42/29
40.Kg1 -2.43/28 Rd8 -2.43/28

41.Re3 -2.53/30 Qxf4 -2.73/29

42.Rf3 -2.73/27 Qxe5 -2.99/31

43.Rxb3 -4.21/32 (43.Qd3 -2.79/29) Bxb3 -4.21/32

44.Qxb3 -4.21/32 Qe1+ -4.21/32

45.Kh2 -4.21/32 Qd2 -4.21/32

46.b5 -4.83/31 (46.Ba1 -4.21/32) d3 -5.12/28

0-1

Section C: Jim Schwar's analysis using Fritz

Jim Schwar (username: Kutztown46) presents his analysis of the game using Fritz (reproduced with permission)

After the conclusion of the GMYS game, I replayed the game using Fritz 9, starting at move 10. For each move, I put Fritz on infinite analysis and let it run overnight. In each case I have shown the move actually played along with its rating after the overnight run. If the actual move was not the 1st choice of Fritz, I have shown what the 1st choice was with its corresponding rating. In a few cases, the move was obvious and I cut the analysis short.

Please note this was Fritz 9, not Rybka. Also note that most of the ratings are 19- or 20-ply. This is how deep my computer got overnight. Deeper ply results from a different engine would no doubt give different results.

10. Nxd7 (.04/19) – 1st choice

10…Nxd7 (.02/19) – 1st choice

11. Nd2 (.01/19) – 1st choice

11…O-O (-.01/18) – 1st choice

12. O-O (.01/18) – 1st choice

12…f5 (.03/18) – 1st choice

13. Rc1 (-.09/15) – 7th choice, 1st choice was a4 (.06)

13…Rc8 (.02/18) – 4th choice, 1st choice was Ba3 (-.11)

14. Bb2 (.00/17) – 1st choice

14…Bd6 (.07/18) – 3rd choice, 1st choice was Bb7 (.03)

15. a3 (-.05/18) – 4th choice, 1st choice was e3 (.06)

15…Qe7 (-.02/19) – 2nd choice, 1st choice was Nf6 (-.04)

16. Rc2 (-.25/18) – 7th choice, 1st choice was Nf3 (-.01)

16…Nf6 (-.20/18) – 2nd choice, 1st choice was Bb7 (-.21)

17. Qc1 (-.30/18) – 4th choice, 1st choice was Re1 (-.25)

17…f4 (-.39/19) – 1st choice

18. Nf3 (-.38/18) – 1st choice

18…Ne4 (-.30/18) – 2nd choice, 1st choice was fxg3 (-.37)

19. Ne5 (-.18/18) – 1st choice

19…Bxe5 (-.19/18) – 2nd choice, 1st choice was fxg3 (-.24)

20. dxe5 (-.18/15) – 1st choice

20…Nc5 (-.05/20) – 1st choice

21. cxd5 (.00/20) – 1st choice

21…cxd5 (.00/21) – 1st choice

22. b4 (-.18/20) – 2nd choice, 1st choice was Qd1 (.00)

22…Nb3 (-.23/21) – 1st choice

23. Qd1 (-.16/21) – 1st choice

23…Rxc2 (-.12/21) – 1st choice

24. Qxc2 (-.12/18) – 1st choice

24…Bc4 (-.18/17) – 1st choice

25. Rd1 (-.12/20) – 2nd choice, 1st choice was gxf4 (-.07)

25…Qf7 (.00/20) – 3rd choice, 1st choice was a5 (-.05)

26. gxf4 (.00/20) – 1st choice

26…Qxf4 (.06/18) – 1st choice

27. e4 (-.22/21) – 2nd choice, 1st choice was e3 (.00)

27…Qf7 (-.13/20) – 1st choice

28. exd5 (-.22/19) – 1st choice

28…exd5 (-.13/18) – 1st choice

29. a4 (-.53/21) – 2nd choice, 1st choice was Bh3 (-.50)

29…Qe6 (.25/20) – 10th choice, 1st choice was a6 or Kh8 (both -.45). This was the largest discrepancy of all black moves. This is very strange why Fritz rated Qe6 so poorly. The line it gave had 30. Qxc4. As shown on the next move below, Fritz rated that as -.42. I do not know how Rybka rated 29…Qe6.

30. Re1 (-.54/21) – 2nd choice, 1st choice was Qxc4 (-.42)

30…d4 (-.65/20) – 1st choice

31. Re4 (-.76/21) – 1st choice

31…Rc8 (-.64/20) – 1st choice

32. Qd1 (-.76/20) – 1st choice

32…Qf5 (-.86/19) – 1st choice

33. a5 (-.95/20) – 1st choice

33…Be6 (-1.04/20) – 1st choice

34. axb6 (-.94/20) – 1st choice

34…axb6 (-1.05/19) – 1st choice

35. h3 (-.91/19) – 1st choice

35…h6 (-.61/19) – 10th choice, 1st choice was Rc7 (-1.04). This was the 2nd largest discrepancy of all black moves. I recall discussion at the time about how the computers did not see the value of …h6.

36. Qf3 (-1.20/20) – 2nd choice, 1st choice was Kh2 (-.97)

36…Qg6 (-1.11/20) – 2nd choice, 1st choice was Rd8 (-1.20)

37. Qd1 (-1.25/21) – 1st choice

37…Kh8 (-1.08/20) – 2nd choice, 1st choice was Rf8 (-1.25)

38. Kh2 (-1.12/19) – 1st choice

38…Rf8 (-1.51/21) – 1st choice

39. f4 (-1.81/21) – 3rd choice, 1st choice was Qe1 (-1.70)

39…Qf5 (-1.77/20) – 2nd choice, 1st choice was Rc8 (-1.86)

40. Kg1 (-1.63/20) – 1st choice

40…Rd8 (-1.59/20) – 1st choice

41. Re3 (-2.48/21) – 2nd choice, 1st choice was Qd3 (-2.12). This was the 2nd largest discrepancy of all white moves.

41…Qxf4 (-2.33/20) – 1st choice

42. Rf3 (-2.77/19) – 2nd choice, 1st choice was Rd3 (-2.48)

42…Qxe5 (-2.72/20) – 2nd choice, 1st choice was Qg5 (-2.80)

43. Rxb3 (-2.77/20) – 1st choice

43…Bxb3 (-2.82/19) – 1st choice

44. Qxb3 (-2.89/20) – 1st choice

44…Qe1+ (-2.96/21) – 1st choice

45. Kh2 (-2.89/19) – 1st choice

45…Qd2 (-3.07/20) – 1st choice

46. b5 (-3.63/20) – 2nd choice, 1st choice was Ba1 (-3.20). This was the largest discrepancy of all white moves.

46…d3 (-4.29/20) – 1st choice

Summary:

White played Fritz 9’s 1st choice 23/37 times (62%).

Black played Fritz 9’s 1st choice 24/37 times (65%).

Looking at it differently:

White played Fritz 9’s 1st choice or another move rated within 0.10 of the 1st choice 26/37 times (70%).

Black played Fritz 9’s 1st choice or another move rated within 0.10 of the 1st choice 33/37 times (89%).

Enter supporting content here