Greengard's ChessNinja.com

VK-DF Game 1 Revisited

| Permalink | 23 comments

Got a minute or 50? I thought this endgame from the first game of the Kramnik-Fritz match was dead and buried, but I need a better line than the direct one I have against what I now think is the best defense. The main defensive concept is 30.e3 Bc5 31.Kf3 b5 32.Ke2 e4 33.Kd2 Kg6. Now Kasparov's 34.Nc7 and directing the knight to g2 (!) looks like a clear, if slow, win. For a while I thought White could win much more directly with 34.Kc3, allowing the kingside invasion of the black king and still winning the race. Now I'm not so sure. Basically I'm still convinced 30.e3 is a win, but the ..b5, ..e4, king invasion defense is tough. There's a lot of junk in the PGN game below, but the main line after 34.Kc3 looks like a draw. Interesting stuff.

[Event "Man-Machine Duel"]
[Site "Bonn, Germany"]
[Date "2006.11.25"]
[Round "1"]
[White "Kramnik, V."]
[Black "Deep Fritz 10"]
[Result "1/2-1/2"]
[ECO "E03"]
[WhiteElo "2750"]
[PlyCount "93"]
[EventDate "2006.??.??"]

1. d4 Nf6 2. c4 e6 3. g3 d5 4. Bg2 dxc4 5. Qa4+ Nbd7 6. Qxc4 a6 7. Qd3 (7. Qd3
Rb8 8. Nc3 b5 9. Bg5 Bb7 10. Bxb7 Rxb7 11. Qf3 Ra7 12. Rd1 b4 13. Na4 Qa8 14.
Bxf6 gxf6 15. d5 Bd6 16. dxe6 fxe6 17. Qh5+ Ke7 18. Nf3 Qc6 19. b3 Ne5 20. O-O
Nxf3+ 21. exf3 Raa8 22. Rfe1 Rae8 23. Rc1 Qb5 24. Nc5 Kd8 25. Rxe6 Bxc5 26.
Qd5+ Bd6 27. Rxd6+ cxd6 28. Qxd6+ Qd7 29. Qxf6+ Qe7 30. Qb6+ {
1-0 Vaulin,A (2540)-Anjuhin,A (2460)/St Petersburg 2001/CBM 082 ext}) (7. Nd2
c5 8. dxc5 Bxc5 9. Nb3 Be7 10. Nf3 b5 11. Qd4 Bb7 12. O-O O-O 13. Rd1 a5 14.
Bg5 a4 15. Nc1 h6 16. Bxf6 Nxf6 17. Nd3 Qb8 18. Qe5 Nd5 19. Qxb8 Raxb8 20. a3
Nb6 21. Nfe5 Bxg2 22. Kxg2 Rfc8 23. Rac1 Nc4 24. e3 {
1/2-1/2 Smyslov,V (2620)-Karpov,A (2540)/Moscow 1971/MCL}) 7... c5 8. dxc5 Bxc5
9. Nf3 O-O (9... Ra7 10. Be3 O-O 11. O-O b6 12. Rd1 Bxe3 13. Qxe3 Rc7 14. Ne5
Qe7 15. Nxd7 Rxd7 16. Rxd7 Nxd7 17. Nc3 Qb4 18. b3 a5 19. Rd1 Nf6 20. Bf3 Bd7
21. Rd4 Qe7 22. Rd3 Rb8 23. Kg2 h6 24. h4 {
1-0 Vaulin,A (2540)-Yakovich,Y (2577)/Elista 2001/CBM 083 (98)}) 10. O-O Qe7
11. Nc3 b6 12. Ne4 Nxe4 13. Qxe4 Nf6 14. Qh4 (14. Qxa8 Bb7 15. Qxf8+ (15. Qa7
b5)) 14... Bb7 15. Bg5 Rfd8 (15... h6) 16. Bxf6 Qxf6 17. Qxf6 (17. Ng5 Bxg2 18.
Qxh7+ Kf8 19. Qh5 g6 20. Qh4 Qd4) 17... gxf6 18. Rfd1 Kf8 (18... Bxf3 $1 19.
Bxf3 Rab8) 19. Ne1 Bxg2 20. Kxg2 f5 21. Rxd8+ Rxd8 22. Nd3 Bd4 23. Rc1 e5 24.
Rc2 (24. Rc6 e4 25. Nf4) (24. e3) 24... Rd5 (24... a5 25. e3) (24... e4 25. Nb4
a5 26. Nc6 Rd6 27. Nxd4 Rxd4 28. Rc6 Rd2 (28... Rb4)) 25. Nb4 Rb5 26. Nxa6 (26.
Nc6 Bxb2 27. Na7) 26... Rxb2 27. Rxb2 Bxb2 28. Nb4 (28. Nc7 Bd4) 28... Kg7 (
28... Bd4 29. Nd5 b5) 29. Nd5 Bd4 30. a4 (30. f3 b5 31. e4 fxe4 32. fxe4 Kg6
33. Kf3 (33. Ne7+ Kf6 34. Nf5 h5 35. Kf3 Bc5 36. g4 Kg6 37. h4) 33... Bc5) (30.
Kf3 f6 31. Ne7) (30. e3 $1 {GK} Bc5 31. Kf3 b5 (31... Kg6 32. a4 Kg5 33. Ke2) (
31... f6 $6 {Seirawan} 32. Ke2 e4 33. Kd2 b5 34. Kc3 Kf7 35. Nc7 Bb6 36. Nxb5
Ba5+ 37. Kd4 Be1 38. a4) 32. Ke2 e4 33. Kd2 Kg6 34. Kc3 {
The obvious direct route but does it win?} (34. h3 Kg5 35. Nc7 b4 36. Kc2 h5
37. Na6 Bd6 38. Kb3 h4 39. gxh4+ Kxh4 40. Nxb4 Kxh3) (34. Nc7 {
This circumspect line from Kasparov seems to win.} b4 35. Nd5 Kg5 36. h3 h6 (
36... Bd6) (36... h5 37. h4+ Kg4 38. Ke2) 37. Nf4 $1 (37. Kc2 h5 38. h4+ Kg4
39. Kd2 Kf3 40. Ke1 Bd6 41. Kf1 f4 42. gxf4 Bc5 43. Nf6 Be7) 37... Bd6 38. Kc2
h5 (38... Be5 39. Kb3 Bd6 40. Nd5 h5 41. Nxb4 h4) 39. Ng2 $1 {Kasparov} Be5 40.
Kb3 Bd6 41. Kc4 Be7 42. Kb5 Bd6 (42... Bf8 43. Kc6 Bg7 44. Kd7 Kg6 45. Nf4+ Kh6
46. Nd5 Be5 47. h4 Kg7 48. Kc6 Bc3 49. a3 Be1 50. axb4 Bxf2 51. b5 {
[%eval 156,25]}) 43. Kc6 Be5 (43... Bf8 44. Kd7 Bc5 45. Ke8 f6 46. Kf7 Bd6 47.
Kg7 Bc5 48. Nh4) 44. Kc5 (44. Kd7 Kf6 45. h4 Bc3 46. Kd6 Kg6 47. Kc5 Kg7 (47...
Bd2 48. Kc4 Kf6 49. Kb3 Ke7 50. Kc2 Bc3 51. a3 (51. a4 Kd6 52. Nf4 Be1) (51.
Kd1 Be5) 51... Kd6 52. axb4 Bxb4 53. Kd1 Ke5 54. Nf4 Ba5 55. Nxh5 Bb4 56. Nf4
Bd6 57. Ng2 Kf6 58. Kc2 Bb4 59. Kb3 Ba5 (59... Be7 60. Kc4) 60. Ka4 Bc3 61. Kb5
Ke6 62. Kc6 Bb4 63. Nf4+ Ke5 64. Kd7 Ba3 (64... Be1 65. Ke7) 65. Ke8 Kf6 66. h5
Kg7) 48. Kc4) 44... Bc3 45. Kc4 Bd2 46. Kb3 Kf6 47. Kc2 Bc3 48. a3 Ke6 49. Kb3
$18) 34... Kg5 (34... Bd6 35. Kb3 Kg5 36. Nc3 b4 37. Nb5 Bf8 38. h3 (38. Nd4
Kg4 $11) 38... h5 39. Nd4 Bc5 (39... h4 40. f4+ exf3 41. Nxf3+ Kf6 42. gxh4 Kg6
43. Nd4 (43. Kc4 {gk} Kh5 44. Kb5 Bd6 45. Kc6 Bf8 46. Kd5 Bh6) 43... Be7 44.
h5+ (44. Kc4 Bxh4) 44... Kxh5 45. Nxf5 Bf8 46. Kc4 (46. Nd4 Kh4 47. Kc4 Kxh3
48. Nc2 f5 49. Nxb4 Be7 50. a4 Bd8 51. Nc6 Bc7 52. Kd4 Kg3 53. Ne7 Kg4 54. Nd5
Ba5) 46... Kg5 47. e4 Kf4 48. Kd5 Kg5 49. Nd6) 40. Kc4 Bd6 41. Kd5 Bf8 42. Ke5
Bg7+ 43. Kd6 Bf8+ 44. Kd5 Bg7 45. Kc5 Bf8+ 46. Kb5 Bd6 47. Kc4 (47. Nc2 h4 48.
gxh4+ Kxh4 49. Nxb4 Kxh3 50. a4 Bc7 51. Nd5 Bd8 52. Nb6 (52. a5 Bxa5) 52...
Bxb6 53. Kxb6 Kg2 54. a5 f4 55. a6 Kxf2 56. a7 fxe3 57. a8=Q e2) 47... Bf8 48.
Nc6 h4 49. gxh4+ Kxh4 50. Nxb4 f4 51. Nd5 f3 52. a4 Bd6 53. a5 Kxh3 54. Kb5 Kg2
55. a6 Kxf2 56. Kc6 Bg3 57. Nc3 Bb8 58. Kb7 Ke1 59. Kxb8 f2 60. a7) 35. Nc7 (
35. h3 h5 36. Nc7 (36. h4+ Kg4 37. Kd2 Kf3) 36... h4 $11) 35... Kg4 (35... b4+
36. Kc4 Bd6 37. Nd5 Kg4 38. Nxb4 Kf3 39. a4 Kxf2 40. a5 Bc7 41. a6 Bb6 42. Nd5
Ba7 43. Kb5) 36. Nxb5 Kf3 (36... h5 37. Kc4 Bb6 38. Nd4 Kh3 39. Nxf5 Kxh2 40.
Kd5 Kg2 41. Kxe4 Kxf2 42. a4 Bd8 43. Ng7 Kxg3 44. Nxh5+ Kg4 45. Ng7 f6 46. Nf5)
37. Kc4 Bb6 38. Nd4+ Kxf2 39. Nxf5 Bxe3 (39... Ke2 40. Kd5 Kd3 41. Nd6 Bxe3 42.
Nxe4 f5 43. Nf6 $18) 40. Kd5 Kf3 41. Nh4+ Kg4 42. Kxe4 Bb6 (42... Bd2 43. a4
Bc3 44. Nf5 Kh3 45. Nh6 Kxh2 46. g4 Kg3 47. Kf5 Kf3 48. Nxf7 Ke3 49. Ng5 h6 50.
Nf7 Kd4 51. Nxh6) (42... Bg1 43. Nf5 Bxh2 $4 44. a4) 43. a4 (43. a3 $5 {
It's hard to prove this is actually better than a4 or whether it just delays
the forced drawing sequences that come after a4.} h6 (43... h5 $6 44. a4 Bc7 (
44... Bd8 45. Nf5) 45. Ng2 f5+ 46. Ke3 Bb6+ 47. Ke2 Bc7 48. Ne3+ Kh3 49. Nd5
Ba5 50. Kf3 Kxh2) 44. Nf5 h5 45. Ne3+ (45. a4) 45... Kh3 46. Nd5 Bd8 47. Nf4+
Kg4 (47... Kxh2 48. Nxh5 Kh3 49. Kf5) 48. Nd3 h4 49. Nf2+ Kg5 50. Kf3 hxg3 51.
Ne4+ Kg6 52. Nxg3) 43... h5 $1 {Draw?!} (43... Kh3 44. Nf5 Kxh2 45. g4 Kh3 46.
Nh6 Kh4 47. Kf5 $18) 44. Nf5 (44. Ng2 f5+ 45. Ke5 Bc7+ (45... Bg1 46. Ke6 Bxh2)
) 44... Bd8 (44... Bc7) (44... Kh3 45. Kf3 Kxh2 46. Ng7) (44... Ba5 45. Ne3+
Kh3 46. Kf3 Kxh2 47. Nd5 Kg1 48. Nf4 Kf1 49. Nd3 (49. Nxh5 Ke1 50. Ke3 Bb6+ 51.
Kd3 (51. Ke4 Kd2) 51... Kf2 52. Ke4 Bc7 53. g4) 49... Kg1 50. Ne5 Kh2 51. Nxf7
Be1 52. Ng5 Bxg3 53. a5) 45. Ne3+ Kh3 46. Kf3 Kxh2 47. Nd5 Kg1 48. Nf4 h4)
30... Bc5 31. h3 (31. e3 f6) 31... f6 (31... h5) 32. f3 Kg6 (32... e4 33. g4)
33. e4 h5 (33... fxe4 34. fxe4 f5 35. Kf3) 34. g4 (34. exf5+ Kxf5 35. g4+ hxg4
36. fxg4+) 34... hxg4 35. hxg4 fxe4 (35... fxg4 36. fxg4 Kg5 37. Kf3) 36. fxe4
Kg5 37. Kf3 Kg6 (37... Bd4 38. Ke2 Kxg4 39. Nxf6+ Kf4 40. Kd3 Bf2) 38. Ke2 Kg5
39. Kd3 Bg1 40. Kc4 Bf2 (40... Kxg4 41. Nxf6+) 41. Kb5 Kxg4 42. Nxf6+ Kf3 43.
Kc6 Bh4 44. Nd7 Kxe4 45. Kxb6 Bf2+ 46. Kc6 Be1 47. Nxe5 1/2-1/2



Here's a link to Karsten Mueller's analysis at Chessvibes (previously Doggers).

J.A. Topfke

Thanks. Yah, the main line is the same but the 43..Be8 he credits to Feist seems weaker than playing 43..h5 immediately, threatening ..h4 in many lines. The king sits on g4 with pawns on f5 and h5 and the bishop can stay on the a4-d8 diagonal. The white king can go all the way to g6 without effect and Black then has the threat of ..Bg1. It looks wrong to lock the pawns like that but it looks unbreakable.

I analysed the 34.Kc3 line with Rybka immediately after game 1 and I posted analysis in game 1 Daily Dirt comments. My conclusion was that 34.Kc3 does NOT win.

You can find my analysis by searching for "34.Kc3"

Thanks, I think I remember some of those notes but I left the Bahamas a few days later and never put much together at the time. I see some others came to similar conclusions. Interesting if the ..b4+ line you gave also draws against 34.Kc3. So is Garry's line the only winner? In the next issue of New In Chess Kramnik's second Lutz gives several winning ideas for White but both against inferior defense (not the ..e4, ..b5, king invasion line we are talking about). He never plays ..e4. All of Seirawan's stuff at ChessBase includes ..f6 and king to the center. Interesting.

Well, it looks like White is still close to winning after 30.a4?! Bc5 31.Kf3 (not Kramnik's 31.h3), so it's really not all that critical, I suppose.

Hi Mig, I also did extensive analysis at the time and concluded that Kasparov's Ng2 plan was the only one that seemed to be winning.

Is there any way you can post your analysis as a downloadable pgn. This is a horror getting into Chessbase.

I'm still fiddling with it. Will put up a cbv later. You can copy and paste that into a text file and rename the extension .pgn, btw.

Finding out if White is still winning by force after 31.Kf3 is probably more important. I'm doubting it right now. It does seem clear to me now that 34.Kc3 doesn't win in the 30.e3 line, which is particularly intriguing if 31.Kf3 does win in the game line.


I like to use Arena Chess to play thru downloaded games. The look is not quite as good as Fritz/Chessbase but for playing thru games it's much nicer. You can paste games directly from the clipboard and such.

Something I'm not following here: if Kasparov's 34.Nc7! wins, then why fuss about 34.Kc3---?

And why does Black have to play 32...e4?

Mousecopy Mig's PGN, open Notepad, paste, save as mathproof.pgn, already I get a red ChessBase icon and it works perfectly, every subline... Just to confirm, on my Win XP laptop.

1 day to Corus, and still no betting on whether or not Kramnik and Topalov will shake hands?

Re: pgn/cbv etc.

Well, I hate to give away one of my favourite secrets, however....

1. Highlight text of game
2. Right click ->copy
3. At board window in Chessbase/Fritz Edit->Paste->Paste Game or simply ctrl-v

Voila. Job done, including variations and annotation.

Chessbase programmers are clever...

very beautiful line: Ng2, the King walk, a3

Yes, it's remarkable. Garry joked that he was turning into Ulf Andersson in his old age. I'd like to see a computer come up with that one. And if the bishop stays on the a3-f8 diagonal the white king infiltrates to the kingside. If it makes it to g6, then Nh4 mates with f4 exf3 Nxf3#!

Ken, I was trying to debunk 34.Kc3, basically. As for ..e4, everything without it seems to lose fairly quickly and I consider it the best defensive plan. It creates the f3 hole for the black king to infiltrate and appears to force the need for the long Kasparov maneuver of the knight to g2. The main reason to leave the pawn on e4 is to try ..f4 at some point, but it doesn't seem to be sufficient. Lutz spends his time on those lines in the next issue of New In Chess and White does well.

Hi Mig (and Bill---fabulous paper you posted in Jan 2 thread, will comment there when I get time!),

I'd be curious to see Lutz's analysis---I wonder if chess people do "peer-review" like we have to in math before things go to press. Especially since an endgame analysis is much more like a math proof than a game, or even than a middlegame analysis... Then again, I've only had time to do 1 hour of game analysis since November!

Yes it is a sure win and the only one imo. Although the lines may be cleaned up a bit for a 'definite' version. Anyone still believing there is another win?

Briliant copy paste by the way Babson!

KWRegan said:
"I'd be curious to see Lutz's analysis---I wonder if chess people do "peer-review" like we have to in math before things go to press. Especially since an endgame analysis is much more like a math proof than a game, or even than a middlegame analysis... Then again, I've only had time to do 1 hour of game analysis since November!"

Chess analysis is more like political opinion than a math proof, especially in positions that are difficult to assess or are considered equal, unclear, or a slight advantage for either side.

Dear Dr. Regan:

Any regrets about not making a career of chess, or at least spending a few years to become a GM?

I remember a game you played in Atlantic City, NJ, beating World Junior Champ Mark Diesen with your patented opening as white, fianchettoing both bishops and I think it was e3 and Ne2 and d3 and c3 and Na3-c2 and d4, although I could be wrong about where your queen knight was.

Most respectfully,


Dear R,
Sorry, I missed your followup comment way back 2 years ago. I found it now on searching for Mark Diesen, who passed away suddenly and much too untimely last month. I just saw the note on the Buffalo Niagara Chess Corner and Susan's blog.
I do have some regrets about not being a GM, though the IM title has sufficed to open the doors that matter to me now. Operationally what happened is I made 2 norms in 1977 right off the bat, but they totaled only 23 of the required 24 games. Then followed a host of near-misses, until I finally got the third in late 1980 or early 1981. By then I had won a scholarship for graduate study at Oxford and was solidly into my mathematical research---and my play in some 1983-4 events was rusty/stale. But I did have the pleasure of being at Merton College (one of 30+ colleges that comprise Oxford) with now-GM William Watson and FM-strength Dugald Macpherson (also a mathematician). If Merton had been able to declare itself a UK sporting entity (http://www.chessbase.com/newsdetail.asp?newsid=4727), we might have finished solidly in the middle of the table at some Olympiads.

---Ken R.

Yes, Diesen's untimely passing. Someone I barely know (from online blitz) emailed me about that last weekend.

I remember Diesen as a cute (but strong) little kid playing in Washington DC events circa 1970. One time he brought a pet locust with him, and placed it on an empty board between rounds and let it wander around the squares. Definitely an offbeat sight at a chess event. I guess locusts don't have much interest in flying, unless there's grain or other food around; either that or Mark's pet was domesticated for indoors.

That was not a locust. That was Fritz 1.

I was so sad to hear of Mark Diesen's death. Actually, he passed away very unexpectedly last December 9. I didn't hear about it until a couple of weeks later. Believe me, I shed real tears. Mark was a great guy.

If you guys also like to hang out at chessgames.com, Mark has a game collection there. It includes many of his games from the mid-1970s. He beat Walter Brown, Michael Rhodes and Arthur Bisguier, among others. If you were his friends, play out some of his games. That's what Mark would have wanted. He loved the game, and he was a tremendous player.

- Alicia

Thanks, Ken, for taking the time to reply to my question posted a while back.

That is so sad about Mark Diesen. I remember a Chess Life article, by Kavalek, about Diesen winning the World Junior. Kavalek was his second. What an achievement by Diesen. I think now one automatically becomes a GM for winning the World Junior, but not back then.

Not too long ago it seems I was chatting with Diesen on ICC (he didn't know me or anything--I just had the temerity to strike up a conversation). I mentioned to him the same game I mentioned above, his loss to Ken Regan in a US Amateur Team Championship way back. Diesen told me, yes he remembered that game, but then he said that he got revenge soon afterward, and he named the subsequent tournament in which he beat Regan. I thought that was great, Diesen expressing his competitiveness that way.

Here's to great American chessplayers, GMs or not.

That's cool that people can take the loan and it opens new opportunities.

Twitter Updates

    Follow me on Twitter



    About this Entry

    This page contains a single entry by Mig published on January 10, 2007 9:12 AM.

    Cheating Hearts Hats was the previous entry in this blog.

    Melting Chess is the next entry in this blog.

    Find recent content on the main index or look in the archives to find all content.