[Event "Corus A"]
[Site "Wijk aan Zee NED"]
[Date "2008.01.26"]
[Round "12"]
[White "Adams, Mi"]
[Black "Polgar, Ju"]
[Result "0-1"]
[ECO "C42"]
[WhiteElo "2726"]
[BlackElo "2707"]
[Annotator "Regan,Kenneth"]
[PlyCount "97"]
[EventDate "2008.01.12"]
1. e4 e5 2. Nf3 Nf6 3. Nxe5 d6 4. Nf3 Nxe4 5. d4 d5 6. Bd3 Nc6 7. O-O Be7 8.
Nc3 Bf5 9. Re1 Nxc3 10. bxc3 Bxd3 11. Qxd3 O-O 12. Rb1 Na5 13. Qf5 Re8 14. Bf4
g6 15. Qh3 Nc4 16. Ne5 Bd6 17. Nxc4 Bxf4 18. Rxe8+ Qxe8 19. Qf3 dxc4 20. Qxf4
Qe2 21. h3 b6 22. Qf3 Qxf3 23. gxf3 a5 24. Re1 Kf8 25. Re4 Re8 26. Kf1 f5 27.
Rxe8+ Kxe8 {One of the most complicated K+P endings between top-level players
in memory. An analyst would wish to have Nimzovich and Kmoch, poets of pawn
structures, at the table, but GM Mihail Marin in his Wijk round-12 report for
ChessBase.com (http://www.chessbase.com/newsdetail.asp?newsid=4416) gave a
deep analysis in one day's time without them.} 28. f4 $5 {In an endgame, every
Pawn advance comes with a cost. This move seems to make f2-f3 incumbent later,
to deny Black's King access from d5 to e4, but then the e3 square becomes a
hole. The reader may still try to deepen GM Marin's analysis of 28.Ke2 in
place of this rightly-queried move. We believe with Marin that it holds, but
do not know---and do not know whether the surprise drawing idea we give at
Move 33 and later is needed here. But we do know that your White defense will
appreciate having a pawn at f2!} (28. Ke2 f4 ({On} 28... Ke7 29. Ke3 Kf6 {,}
30. Kf4 {tries to work around GM Marin's idea that Black can force f4 by White,
and we leave this as unclear.}) 29. Kd2 Ke7 {
Here is just one illustrative line.} 30. Ke2 b5 31. Kd2 h5 32. Kd1 Kf6 33. Ke2
Kg5 34. Kf1 b4 35. Kg2 $1 {usually, taking on b4 is fatal} bxc3 36. Kf1 Kf5 37.
Ke1 Ke6 38. Kd1 Kd5 39. Kc1 Kxd4 40. Kd1 Ke5 41. Ke2 Kf5 42. Ke1 g5 43. Kf1 $1
{Here, the only move!} (43. Ke2 g4 44. hxg4+ hxg4 45. Kf1 g3 $19) 43... c6 44.
Kg2 $1 {And now, can Black do anything?} (44. Ke2 g4 $19 45. hxg4+ hxg4 46.
fxg4+ Kxg4 {loses in a position that is mutual Zugzwang}) (44. Ke1 g4 $19))
28... Kd7 29. Ke2 Kd6 (29... Kc6 {
shows a Black idea that almost never succeeds:} 30. Kd2 Kb5 31. Kc1 Ka4 32. Kb2
b5 33. a3 h6 34. h4 c6 35. f3 h5 36. Ka2 b4 37. axb4 axb4 38. Kb2 b3 39. cxb3+
cxb3 40. c4 Kb4 41. d5 $11) 30. Kd2 Kc6 31. f3 {Not to be criticized---Black
can force it anyway at no cost, as demonstrated by Marin.} Kd6 32. Ke3 Kd5 33.
Ke2 {Marin's key position, in which we introduce a key idea. Usually creating
an outside passed pawn for the opponent is suicide in a King and Pawn ending.
In this ending, however, doing so gives White's King counterplay along the a-
and h-files, and the position is blocked enough to slow the other King's usual
mayhem on the other wing. This "Suicide Draw" idea refutes Marin's analysis
and winning assessment of the rest of the endgame. It also exemplifies Marin's
own stated caution about relying on chess engines, as Deep Fritz 10 evaluates
Black as over a Rook ahead in some of the drawn positions that result!} Kd6 (
33... b5 {Marin.} 34. Kd2 c6 35. a3 c5 36. dxc5 Kxc5 37. Kc1 b4 38. Kb2 $1 {
KWR.} ({Marin gives only} 38. axb4+ $2 axb4 39. cxb4+ Kxb4 40. Kb2 c3+ 41. Ka2
h6 42. h4 h5 43. Kb1 Kc4 44. Ka2 Kd4 45. Kb3 Ke3 46. Kxc3 {when} Kxf4 $1 {
also wins the resulting Queen endings, without needing his tricky 46...Kxf3 47.
Kd2 Ke4!!}) 38... bxa3+ (38... bxc3+ 39. Kxc3 Kb5 40. Kb2 Kc6 41. Kc3 Kc5 42.
h4 h6 43. Kd2 Kb5 44. Kc3 Ka4 45. Kxc4 Kxa3 46. Kd5 a4 47. c4 Kb3 48. c5 a3 49.
c6 a2 50. c7 a1=Q 51. c8=Q $11) 39. Kxa3 {Fritz 10 instantly regards White's
38.Kb2 as having blundered the equivalent of 2 pawns, and here gives Black a
usually-winning edge over 3 pawns, thinking 10 moves (20 ply) ahead and beyond.
But let Fritz think longer, and its evaluation will flatline to 0. 00. In
chess terms, why? Black cannot force the a-pawn further than a4, so Black's
only plan is the usual one of running to the other wing. But White's pawns at
c3, f3, and f4 block any passage other than the Strait of Magellan at g7, by
which time White has gobbled up a5 and c4 and queened a c-pawn long before.
Thus Black's King can only tack around a5, and once the engine enumerates
enough moves to get a repetition in each branch, it sees the positional
reality of White's idea.} Kb5 40. Kb2 h6 41. h4 $11 {Starting Fritz here gives
"0.00" fairly quickly, usually without initial 3-pawns-up opinions. But back
at Move 38 those opinions may have hidden the presence of 38.Kb2 as a viable
option. Below we find a case where Deep Fritz 10 sometimes gives Black a
score over 5 pawns ahead, to depths well past 30, and yet the position is
drawn!}) 34. Kd2 Ke7 35. Ke3 Ke6 36. Ke2 {After some pre-time-control tacking,
the key point in the actual game. Because the "Suicide Draw" idea makes most
of White's options viable after moves other than 36...g5!?, we now regard
Judit Polgar's choice as an excellent practical one, especially pre-control
and with White's 28.f4 having already shown an aversion to letting Black play .
..f4!} g5 $5 {A fine try, because the slower 36...Kf6 or etc. give White
multiple ways to hold. All White really needs is to realize that certain
positions with the Kings at d5 and d2 are Zugzwangs, together with the "Kb2!"
idea after Black has played ...c5. Plus avoid fxg5 opening the K-side when
Black can recapture in a way that prevents blocking on f4/h4. The active moves
a3 and/or h4 turn out to be useful early on, rather than weaknesses or losses
of tempo.} (36... Kf6 {appears best among other tries. Although early moves
a3 and.or h4 seem steadiest, we give a holding main line where White shuttles
the King to illustrate the ideas in the position. The branch at 46.h4?, with a
return mistake 50...c5?, also leads to the most extreme successful case of the
"Suicide Draw" defense, whereby Deep Fritz 10 has given evals as extreme as -5.
22 beyond depth 30 in a drawn position!} 37. Kf2 (37. Ke3 g5 $1 {
transposes into the game}) (37. Kd2 g5 38. Ke3 Kg6 $1 {loses similarly}) (37.
h4 $5 {is the only move besides 37.Kf2 to hold, and leads to a correspondence
of squares around mutual Zugzwang positions:} b5 (37... h6 38. Kf2 $1 b5 39.
Ke3 $1 (39. a3 {may also hold}) 39... g5 40. hxg5+ hxg5 41. fxg5+ Kxg5 42. d5
$1 b4 43. Kd4 Kf4 44. cxb4 axb4 45. Kxc4 Kxf3 46. Kxb4 f4 47. a4 Ke4 48. Kc3
Kxd5 49. a5 c5 50. a6 Kc6 51. Kd3 $11) 38. Ke1 Kg7 39. Kf2 c6 (39... b4 40. d5
$1 $11) (39... Kh6 40. d5 $1 $11) 40. a3 {At this point, the only move...} (40.
Kf1 {loses} Kh6 41. Kg2 b4 42. Kf2 bxc3 43. a3 Kg7 $19) 40... Kf6 41. Ke2 {
41.Ke3 and 41.Ke1 are fine too, with 42.Ke2! next.} Ke6 42. Ke1 $3 (42. Kd2 Kd5
$1 {is Zugzwang}) (42. Kd1 Kd6 $1 43. Kc1 b4 44. Kb2 Kd5 $19 {
and the loss of tempo is murder.}) 42... Kd6 43. Kd1 $1 Kd5 44. Kd2 c5 45. dxc5
Kxc5 46. Kc1 b4 47. Kb2 $1 {The "Suicide Draw" strikes again.}) 37... h6 {
Other tries seem no more challenging.} (37... g5 38. Kg3 $11 {
since 38...h6 39.fxg5+ hxg5?? 40.f4! actually loses.}) 38. Ke3 (38. Kg3 {loses}
Ke6 39. Kf2 Kd5 40. Ke2 (40. h4 $5 b5 $1 $19) (40. Ke3 Kc6 {transposes}) 40...
Kc6 {The lone case we see where the ...Kb5-a4 idea works} 41. Kd1 Kb5 42. Kc1
Ka4 43. Kb2 g5 44. fxg5 (44. a3 b5 $1 {wins} 45. d5 gxf4 46. h4 h5 {
The Zugzwang is finally achieved!}) 44... hxg5 45. Kc1 (45. a3 {loses} b5 46.
d5 f4 $19) (45. Ka1 Ka3 $19) 45... Ka3 {Brave!} 46. f4 gxf4 47. h4 f3 48. Kd2
Kxa2 49. h5 a4 50. h6 a3 51. h7 f2 $1 52. Ke2 Kb2 53. h8=Q a2 $19 {White has qu
eened first, but has no checks! Nor can White make a tempo for activity,
because ...f1Q+! will queen with check for Black. Then White's Pawns fall.}) (
38. a3 $5 {Can Black do better than} Ke6 39. h4 Kd5 40. Ke3 b5 41. Kd2 {forced}
c5 42. dxc5 Kxc5 43. Kc1 b4 44. Kb2 $1 {with the thematic defense again?}) (38.
h4 $5 {transposes into 37.h4 h6 38.Kf2, which holds as well}) 38... g5 39. Kf2
gxf4 40. Kf1 {White's King need only avoid being caught one file to the right
of Black's King, or two files to the left, with Black to move.} Ke6 41. Ke2 Kd5
(41... b5 42. a3 Kd5 43. Kd2 $1 {defends by mutual Zugzwang}) 42. Kd2 {
Safe with a3 and ...b5 not played} Kc6 43. Kc1 Kb5 44. Kb1 Ka4 45. Kb2 c6 (
45... b5 46. h4 h5 47. a3 c6 48. Ka2 b4 49. axb4 axb4 50. Kb2 $11) 46. a3 $1 ({
5-Pawn Swing, this way-->} 46. h4 {This deprives White of a reserve tempo, and
loses even though it takes Black "forever" to wend the King back to the center.
However, hasty play by Black with 50...c5? can still throw away the win.} Kb5
47. Kc1 Ka6 48. Kb2 b5 49. Kc1 Kb7 50. Kb2 {~~*But*, DF10's PV continued} c5 $2
{Start your favorite engine afresh in this position, and see if it gives evals
like -4.70---5.22 from Deep Fritz 10 at depths beyond 30!} ({Correct is} 50...
Kc7 $1 51. Kc1 Kd7 52. Kd2 Ke6 53. Ke2 Kf6 54. Kf2 Kg6 55. Kg2 b4 $19 {
an idea given also by Marin}) 51. dxc5 Kc6 {
A subtle correspondence still lurks here.} 52. a3 $3 (52. Kc1 Kd5 $3 53. c6
Kxc6 54. a3 $1 Kc5 {and this is Zugzwang of White}) 52... Kd5 53. Kb1 $3 Kxc5
54. Kc1 {White is ready to meet 54...b4 by 55.Kb2!, and at the same time will
force Black to play ...a4 to win the h-pawn.} Kd5 55. Kd2 Kc6 56. Kd1 $1 Kd6
57. Kd2 Kd5 58. Kd1 $1 Ke6 59. Ke2 Kf6 60. Kf2 Kg6 61. Kg2 Kh5 62. Kh3 a4 63.
Kh2 {Undoubtedly what DF10 has missed is that Black cannot win this position.
(A human needs to fill in: Black can run the King to d5 and play ...b4 cxb4+
Kd4, but after ...Kc3 or ... Ke3 White's b-pawn runs too fast.) Can readers
compose "Suicide Draws" that give more-extreme evals from engines?}) 46... b5
47. h4 $1 {Now this is safe, with BK having walled himself in, and ...h5
produces a total K-side block.} h5 48. Ka2 b4 49. axb4 axb4 50. Kb2 b3 51.
cxb3+ cxb3 52. c4 Kb4 53. d5 $11) 37. Ke3 $2 {The losing move} (37. fxg5 $1 {
Vladimir Barsky (http://www.chesspro.ru/_events/2008/weik24.html) gave only
analysis of this capture, but gave all that needs saying, including the 47.
Kb2! move which I (KWR) had also found.} f4 38. Kf2 Kf5 {Now Marin wrote that
unless White tries 39.h4, "Black would simply switch back to the
aforementioned plan"---but the "Suicide Draw" applies equally well in the
position that results.} 39. a3 (39. h4 Kg6 40. Kg2 Kh5 41. Kh3 b5 $1 42. a3 c6
{White must cough up the h- and g-pawns, and here the outside passer really
does win, because the Queenside is fluid rather than blocked. Marin continues:}
43. Kh2 Kxh4 44. Kg2 Kxg5 45. Kh3 h5 $1 46. Kg2 Kf5 47. Kh3 Ke6 48. Kh4 Kd5 49.
Kg5 (49. Kxh5 b4 $19) 49... b4 $1 50. cxb4 axb4 51. axb4 Kxd4 52. Kxf4 {
, and now} h4 $1 {eats f3 and then the Q-side.}) 39... Kxg5 40. Kg1 Kf6 (40...
h5 41. Kf2) 41. Kf2 Ke6 42. Ke2 b5 43. Ke1 Kd5 44. Kd2 c5 45. dxc5 Kxc5 46. Kc1
b4 47. Kb2 $1 $11 {The Black pawn at f4 plays the same blocking role as the
White pawn at f4 in the first instance of this defense after Marin's 33...b5.})
(37. Kf2 {also loses, because Black has the reserve tempo ...h6, and so can
win by a second instructive mutual-Zugzwang position:} gxf4 38. a3 b5 39. Ke2
Kd5 40. Kd2 h6 $1 {This arises also in lines with 36...Kf6 or 36...h6. Black
to move only draws, but here White loses!} 41. Kc1 b4 42. Kb2 bxa3+ 43. Kxa3
Ke6 {Starting Black's King from d5 not c5, and having the c7-pawn to slow
White's counterplay, gives Black just enough time to win the Queen endings
that result---full details are in the PGN file at http://www.cse.buffalo.edu/
~regan/chess/analyses/AdamsPolgar108full.pgn}) 37... Kf6 $1 {Now Black is
winning, but Polgar had to see the next move too, else 39.Kd4 would cause
concern.} 38. d5 (38. h4 $5 {This was the first "Suicide" idea I posted at
http://susanpolgar. blogspot.com/2008/01/strongest-ever.html on the day of the
game, but Black can accept the gambit as well as win by my intended (and still
easier) 38...h6(!) or 38...g4(!).} gxh4 39. Kf2 b5 40. Kg2 b4 {
Black cannot gain time by defending h4 first.} 41. Kh3 Ke6 42. Kxh4 bxc3 43.
Kg5 c6 $1 {The key time-gaining move, also stopping 44.d5+!} 44. a4 h5 45. Kxh5
Kd5 46. Kg5 Kxd4 47. Kxf5 Ke3 48. Kg4 Kd2 49. f5 Kxc2 50. f6 Kb2 51. f7 c2 52.
f8=Q c1=Q $19 {Black's pawn on c4 is faster than White's pawn on f3.}) (38. Kf2
{loses as with 37.Kf2.}) 38... h5 $1 39. a3 (39. Kd4 g4 40. hxg4 (40. Ke3 b5)
40... fxg4 41. fxg4 hxg4 42. Ke4 b5 43. f5 Kg5 44. f6 Kxf6 45. Kf4 a4 46. Kxg4
Ke5 $19) (39. a4 g4 $19) 39... g4 40. fxg4 fxg4 41. hxg4 hxg4 42. Ke4 b5 43. f5
(43. Kd4 Kf5 44. Ke3 a4 $19) 43... Kg5 44. f6 Kxf6 45. Kf4 a4 46. Kxg4 Ke5 47.
Kf3 Kxd5 48. Ke3 Ke5 49. Kf3 0-1