Friday, June 03, 2005

Re-evaluating a famous game

[Event "Moscow cf"]
[Site "?"]
[Date "1974.??.??"]
[Round "?"]
[White "Karpov, Anatoly"]
[Black "Kortschnoj, Viktor"]
[Result "1-0"]
[ECO "B78"]
[WhiteElo "2700"]
[BlackElo "2670"]
[PlyCount "53"]
[EventDate "1974.??.??"]

As a preface to my study of another of my old Dragon games, the context provided by this game loomed large. This is one of the most famous "prime time" outings for the Dragon, and the best known game from what I call Karpov's Reign of Terror. Anatoly Karpov amassed an almost unbelievable record of victories against the Dragon, by my count on Chessbase it was something like 14 wins, one draw, and one loss (in a practice match against Kortschnoj). His victories include two wins over Miles and three over Sosonko, the leading Dragonists of the time. Despite this incredible pattern of Dragon-slaying,
none of the variations Karpov used to defeat the Dragon is now considered critical. In the Karpov-Gik game Black subsequently learned not to play .... Qxc3. The classical lines with Bg5, although troublesome back in 1977, have been adequately met and are no longer considered to offer any advantage. 9 g4 in the Yugoslav is considered the weakest of White's three main choices: 9 Bc4, 9 O-O-O and 9 g4, but it defeated the likes of Miles and Mestel in the early
1980s. This game is no exception. After this game it was discovered that 16...Re8, to preserve the Dragon bishop, yields no advantage for White and maybe a disadvantage, really. 16...Qa5, however, was considered inferior and good for advantage for White. White's chances are considered so good that White players looked for creative ways to transpose their games to the position after White's 19th move.
I have my doubts, however. I'm prepared to contend that Black found the fourth best move on move 19, and that three other lines could have kept Black in the game and given him a chance for
equality.

1. e4 1... c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 g6 6. Be3 Bg7
7. f3 Nc6 8. Qd2 O-O 9. Bc4 Bd7 10. h4 Rc8 11. Bb3 Ne5 12. O-O-O Nc4 13. Bxc4 Rxc4 14. h5 Nxh5 15. g4 Nf6 16. Nde2

This is one of something like ten or twelve possibilities here. Thanks to 16....Re8, it is no longer considered critical. 16. Nb3 was recommended in the second edition of Beating the Sicilian, 16. Bh6 was played a lot prior to this game, 16. b3 is a tricky line
that beat me in a postal game once, 16. e5 is sharp but answerable, 16. Kb1 may be dangerous.

16... Qa5 17. Bh6 Bxh6 18. Qxh6 Rfc8 19. Rd3 19... R4c5

Here is the fourth-best move that loses for Black.

A) Almost completely unnoticed in the literature on the game was the move 19...R8c5, keeping the rook in an aggressive position on c4.} 19... R8c5 20. g5 (20. Nf4 Be6 21. g5 Rxg5 22. Nfd5 Rxd5 23. Rxd5 23... Qc7 {looks fine for Black}) 20... Nh5 21. Nf4 Rxc3 (21... Rxg5 22. Rd5 wins for White) 22. Rxc3 Rxc3 (22... Rxg5 23. Nd5 Rxd5 24.
exd5 Qxa2 25. Ra3 Qxd5 26. Rxa7 Qc6 27. Ra8+ Be8 28. Rd8 Nf6 29. Kb1 29... b5 looks unclear to me.) 23. Nxh5 gxh5 24. Qxh5 Kf8 25. Qh6+ Ke8 26. Qxh7 Rxc2+27. Kxc2 Be6 28. a3 Qxg5 29. Qh5 29... Qg2+ with compensation).

19... Qd8 20.
B1. 20. g5 Nh5 21. Nf4 (21. Rxd6 Rxc3 22. Nxc3 Rxc3 23. Rxd7 Qxd7 24. bxc3 e6) 21... Qf8 22. Qxf8+ (22. Nxh5 Qxh6 23. gxh6 gxh5 24. Rxh5 R8c5 25. Rxc5 Rxc5 26. Nd5 Kf8 27. Rd1 Bc6 28. b4 Rc4 29. Ne3 Rxb4 30. Rg1 Bxe4 31. fxe4 Rxe4 32. Kd2 Rh4 33. Nf5 Rh534. Rf1 = {Sapi and Schneider}) 22... Kxf8 (22... Rxf8 23. Nxh5 gxh5 24. Nd5 Re8 25. Rdd1 Be6 26. Nf4 Rec8 27. Rd2 R4c5 28. Nxe6 fxe6 29. Rdh2 29... Rf8 { Ding Yulin-Yu Hanling Torch Real Estate 2005 1/2 34}) 23. Nxh5 gxh5 24. Rxh5 Kg7 25. Rd2 Rxc3 26. bxc3 Rxc3 27. Rd3 Rxd3 28. cxd3 Kg6 29. Rh6+ Kxg5 30. Rxh7 Be6 looks fine for Black.

B2. 20.Nd5 (Shamkovich) Rxc2+ (20... e6 is a key alternative not considered in the books} 21. Nxf6+ Qxf6 22. Qxh7+ Kf8 23. Nc3 Ke7 (23...Qf4+ 24. Kb1) 24. Qh2 Rd4 25. Rxd4 Qxd4 26. Qh4+ Qf6 27. Qxf6+ Kxf6 28. Kd22 g5 {And White's advantage is hard to find.})
21. Kb1 e6 22. Ndc3 Rxe2 23.Nxe2 Bb5 24. Rd2 Bxe2 25. Rxe2 {Shamkovich says White's exchange advantage might be decisive, but} 25... Qb6 26. Reh2 Qd4 27. Rd2 Qe5 28. Qe3 a6 29. g5 Nh5 30. Qb6 Qxg5 31. Qxd6 Qc5 32.Qxc5 Rxc5 33. Rd7 b5 34. Rc1 Rxc1+ 35. Kxc1 35... Nf4 {looks more than OK for Black. I think Shamkovich underestimated Black's active queen.})

C. 19... Be6 {The best-known of the altermatives.} 20. g5 Nh5 21. Ng3 Qe5 22. Nxh5 gxh5 23. Qxh5 (23. Rxh5 Qg7 24. Rd2 Qxh6 25. Rxh6 R8c5 26. f4 Bg4 27. Rdh2 Bf3 28. Rxh7 Bxe4 29. Nxe4 Rxe4 30. R7h4 Kg7 31. c3 a6 32. Rh7+ Kf8 33. Rh8+ Kg7 34. R2h4 (34. R2h7+) 34... Rcc4 35. R8h7+ Kf8 36. R4h6 36... Rxf4 and White has nothing better than a draw.) 23... Kf8 24. Qh2 Qxg5+ 25. f4 Qf6 26. f5 Rxc3 27. bxc3 Bxa2 28. Qxh7 Ke8 29. Qh8+ 29... Kd7 is an old Nesis postal game which looks satisfactory for Black. 29... Qxh8 {
is also possible},

The actual game continued:
20. g5 Rxg5 21. Rd5 Rxd5 22. Nxd5 Re8 23. Nef4 Bc6 24. e5
Bxd5 25. exf6 exf6 26. Qxh7+ Kf8 27. Qh8+ 1-0

9 comments:

Dennis Monokroussos said...

Hi Victor:

I'm glad you've resumed your Dragon series - the earlier entries were outstanding!

Four initial comments on this one.

First, I hope you'll spend a little time making the format a bit more user-friendly - the content is very interesting, but all the parentheses will probably scare people away.

Second, part of what made the earlier entries so good was your taking the time to explain many of the key ideas and finesses. There's very little of that here, though, and I had to stop for a moment or two on more than one occasion to figure things out.

(Two examples: (1) The move 21.Rxd6 after 19...Qd8 20.g5 Nh5. (2) In the line 19...R8c5 20.g5 Nh5 21.Nf4 Rxc3 22.Rxc3 Rxc3 23.Nxh5 gxh5 24.Qxh5 Kf8 25.Qh6+ Ke8 26.Qxh7 Rxc2+ 27.Kxc2 Be6 28.a3 Qxg5 29.Qh5 Qg2+ with compensation, I had to figure out (a) why Black doesn't play 25...Rxc2+, (b) why White threw in 25.Qh6+ before taking on h7, (c) 28.a3 and (d) 29.Qh5.)

Third, it would be helpful to your readers if you put in links to your previous Dragon posts, especially because they were so good!

Finally, an analytical comment. I wonder about the looooooong variation 19...Qd8 20.Nd5 (labeled "Shamkovich's supposed refutation") 20...Rxc2+ 21.Kb1 e6 22.Ndc3 Rxe2 23.Nxe2 Bb5 24.Rd2 Bxe2 25.Rxe2 (Here, I guess Shamkovich claims White is winning, but you demur, continuing as follows:) 25...Qb6 26.Reh2 Qd4 27.Rd2 Qe5 28.Qe3 a6 29.g5 Nh5 30.Qb6 Qxg5 31.Qxd6 Qc5 32.Qxc5 Rxc5 33.Rd7 b5 34.Rc1 Rxc1+ 35.Kxc1 Nf4.

You stop here, claiming that this "looks more than OK for Black". In my opinion, Black is simply lost: White brings the king to the kingside to deal with the h-pawn, the rook picks up at least one of Black's queenside pawns, and then White wins with the resulting queenside pawn majority.

Of course, there may be an improvement earlier in that line (between moves 25 and 35), and even if 19...Qd8 is as bad as Shamkovich claims, White must still dispose of 19...R8c5 and 19...Be6 before this line regains its theoretical significance for the first player.

Victor Reppert said...

Dennis: I did some reformatting to make this look at lot nicer. And I'll get to a couple of the other issues as soon as I can.

Victor

Victor Reppert said...

Maybe in the Shamkovich line 29...b6 is an improvement. Though this is a GM's evaluation of the overall chances in and endgame--and I suspect he may be right in the end. Also look at the immediate 20...e6 against 20. Nd5. That seems to have been overlooked.

Dennis Monokroussos said...

Assuming you mean 28...b6, that does seem like an improvement, but I think White still has a significant advantage after 29.Qh6 (29...d5 30.exd5 exd5 31.Rdd1; 29...Rc5 30.Qh2 Qxh2 31.Rhxh2 Ne8 32.Rc2 - I'm a bit less sure about this line, but I still wouldn't want to have Black here).

As for 20...e6, 21.Ne3 R4c5 22.Rxd6 is clearly better for White.

Victor Reppert said...

Dennis: This is not one of these tactical situations where computer numbers are going to be very helpful; this is a complex position where it may very well turn out that Black is simply being ground down. So at this point I have some doubts about White's chances in the ...Qd8 variation, leaving the other two as preferable, at this point in my thinking.

Dennis Monokroussos said...

I think you mean to express doubts about Black's chances in the 19...Qd8 variation, right? As well you should, I think - the only question there, as far as I can tell, is if Black can achieve an inferior position that isn't just lost.

Now let's start dismantling Black's other tries. For your entertainment, let's take your line with 19...R8c5. I'll grant the whole thing: 20.g5 Nh5 21.Nf4 Rxc3 22.Rxc3 Rxc3 23.Nxh5 gxh5 24.Qxh5 Kf8 25.Qh6+ Ke8 26.Qxh7 Rxc2+ 27.Kxc2 Be6 28.a3 Qxg5 29.Qh5 Qg2+ "with compensation".

I don't think so: 30.Kb1, and now Black has to worry about the simple but effective attacking idea of Qh8+ with Rc1, threatening Qc8# and general mayhem. Here's my main line:

30.Kb1 Bc4 31.Rc1 Qe2 32.Ka1 Qd3 33.Qh8+ Kd7 34.Qc3 with an ending that's probably lost for Black, and is at least highly undesirable.

Once we've agreed that this line's more or less dead and buried, I'll start chipping away at the official main line with 19...Be6. Keep me posted.

Victor Reppert said...

Dennis: I agree on that line, but you haven't dealt with the alternative line starting with a deviation on move 22 for Black which I labeled unclear. That one has been holding up so far when I tried analyzing it, though it is kind of strange.

Dennis Monokroussos said...

Okay, time to dispose of that line: 19...R8c5 20.g5 Nh5 21.Nf4 Rxc3 22.Rxc3 and now the sideline with 22...Rxg5. I agree that the position at the end of your line continuing with 23.Nd5 isn't completely clear, but I think the following simple approach is quite powerful:

23.Nxh5 Rxh5 (23...gxh5 24.f4+-) 24.Rxh5 and now:

(a) 24...gxh5 25.a3 f6 26.f4 Qb5 27.b3 and White is winning: his king is safe, but Black has to worry about White's rook penetrating both along the c-file and the g-file.

(b) 24...Qxh5 25.Qxh5 gxh5 26.Kd2 h4 27.Ke3 Bc6 28.Kf4 h5 29.Kg5 h3 30.Rd3 and White will round up an h-pawn and slowly but surely use the extra exchange to make progress. Maybe Black can save a draw, but it's going to be a miserable undertaking. A possible continuation: 30...Bb5 (30...Bd7 31.Rd2 f6+ 32.Kxh5 Kf7 33.Kh4 +/-) 31.Rd2 Bf1 32.e5!! Bg2 33.exd6 exd6 34.Kh4 Kf8 35.Kg3 Ke7 36.Rd4 and White wins.

Victor Reppert said...

Ok this isn't looking good either. Time to focus on 19...Be6.