|
< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> |
| Mar-26-14 | | cro777: In a very interesting endgame, Karjakin missed to win in the position after 45.g4 
click for larger viewHere, instead of 45...Bf2, he should have played 45...a3 folloewd by 46...Be3 (after 46.Kc2) and then ...e4 and ...Ke5. For instance, 45...a3 46. Kc2 Be3 47. Rh8 e4 48. fxe4 Ke5
click for larger view |
|
| Mar-26-14 | | csmath: Except that 47. Rh7? would not be the move. Try analyzing something else. You have a good line though, this could indeed be winning. |
|
Mar-26-14
 | | TheAlchemist: Hmm, now that I look at it more closely you are right, I guess I didn't pay attention enough and it seems Topalov was really walking a fine line. <cro777> What about 47.Kb3 in your line? And say 47...e4 48.Re8? |
|
| Mar-26-14 | | Ulhumbrus: Suppose that after 46 Rh8 Black plays - as the commentators on the official site may have indicated - 46...e4. This clears the way for Black's king to march to g2 and support the advance of the h pawn. |
|
| Mar-26-14 | | csmath: This is almost a study. Bringing bishop to f4 and then playing pawn to e4. I do not think anybody would find that over the board. |
|
| Mar-26-14 | | csmath: I think
45. ... a3
46. Kc2 Be3
with bishop going to f4 and pawn to e4!
is indeed a winning line. I cannot find any line that draws for white. This seems to be a missed win but a study like win. |
|
| Mar-26-14 | | csmath: Change of analysis, with respect to what cro777 found. 15. g5!?
[first new move, both players are prepared. Black is in waiting mode to see what white has in the position.] 18. ...Qa8
[binding but not very dangerous, Karjakin has a draw in mind.] 19. Qh1?!
[Yet again Topalov shows that he is nervous. He is generally active player but on this tournament he seems to be impatient.] 22. Rxb7
and the game seems to be headed for a draw. But ...
34. b5!?
[Topalov decides to try something.]
35. a5!?
[Consistent and insistent! Not too many players would try this.] 38. Ra2?!
[38. c5 dxc5 39. Bxc5 and white cannot hope for anything but draw.] 
click for larger view39. ...Bxe3
[...a4 40. Rxb6 Rxb6 41. Rxb6 Rxb6 42. Bxb6 a3 43. Ba5 a2 44. Bc3 and in this crazy position draw is the outcome since h-passer cannot be controlled other than by king. For example 44. ...Kf5 (the most dangerous) 45. Kf2 Ke6 46. d4! exd4 (46. ...a5? 47. d5! and white wins!) and now white has f-passer and the game is clearly a draw.] 41. ...Bc5
click for larger view45. g4?
[White gambled too much. 45. gxh4 gxh4 46. Rd8 a3 47. Kc2 a2 48. Kb2 Kg7 49. Rd7 and white is barely holding.] 45. ...Bf2?
[...Bg1 46 Rf8 Kg7 47. Rd8 a3 48. Kc2 h3 49. Rd7 and again white is barely holding. But 45. ...a3! 46. Kc2 Be3 
click for larger viewwith Bf4 and e4! seems to be winning in all lines.]
and game is drawn. Truly dangerous attempt by Topalov. =====
Yet again Topalov shows that he is as serious as he can be and that no position is drawn for him until proven so. However he could have lost in this game after his 45th move. |
|
| Mar-26-14 | | csmath: Very, very interesting and wild endgame. Karjakin could have won this but Topalov was rewarded for bravery when Karjakin missed a study-like win. |
|
| Mar-26-14 | | cro777: <TheAlchemist: What about 47.Kb3 in your line?> After 45...a3 46.Kc2 Be3 47.Kb3
click for larger viewGM Erwin L'Ami suggested the following winning line:
47...Bf4 48.Rh8 e4 49.dxe4 Ke5 50.Kxa3 Kd4 51.Kb4 a5+ 52.Kb5 a4 |
|
| Mar-26-14 | | devere: <csmath: This is almost a study. Bringing bishop to f4 and then playing pawn to e4. I do not think anybody would find that over the board.> I thought of it almost immediately, and I am not a grandmaster. The Black king must penetrate the White position to win, and playing e4 is the only way to do it. It was a missed opportunity for Karjakin. |
|
Mar-26-14
| | TheAlchemist: <csmath> Again, sorry about the mess. I was thinking of maybe 45...a3 46.Kc2 Be3 47.Kb3 Bf4 48.Ra8 (with the dream of maybe bringing the rook to a1 in time by capturing both pawns first, though it doesn't seem possible) e4 49.fxe4 (I tried with this to maybe have the f-file to cut off the king and the potential d3-d4, but there's never time) Ke5 50.Kc3 and here I was a little stuck. If 50...a2 to deflect the king then 51.Re8+ first (same move after 50...h3), and if 50...Bg3 51.Rf8 Be1+ 52.Kb3 Kd4 53.Rf3. Then I tried 48...h3, and only after 49.Rh8 h2 50.Kxa3 and only now 50...e4. Here I think white cannot stop the black king. <cro777:
47...Bf4 48.Rh8 e4 49.dxe4 Ke5 50.Kxa3 Kd4 51.Kb4 a5+ 52.Kb5 a4> Thanks! And if 50.Kc3 the cute 50...Bd2+ works. |
|
Mar-26-14
 | | tamar: The difficulty for Black, as Karjakin demonstrated in the press conference, is that if he gets any of the timing of the ...e4 line wrong, he actually can lose some of the positions. I did some analysis with Komodo, and it appears that even if Black orchestrates a timely ...e4, White has drawing chances. A key line appears to be 45...a3 46 Kc2 Be3 47 Kb3 Bf4 48 Rd8 Ke6 49 Rh8 e4
50 Re8+ Kd7 51 Rxe4 Bg3 52 Re2 h3 53 Rc2 Bf4 54 Re2 h2 55 Rh1 Kc6 56 d4 d5 57 c5 Kb5 58 Rd1 a2 59 Rh1 Bg3 60 Rc1 Bf2 61 c6 a1Q 62 Rxa1 Bxd4 63 c7 Bxa1 64 c8Q h1Q 
click for larger viewWin or draw? I'm not sure that even if Black can escape the checks, he can win from here. |
|
Mar-26-14
| | tamar: Note that Komodo plays 48 Rd8 to lure the King to e6 so that when e4 occurs, Black has check on e8, and captures on e4 with the rook. This appears to be a finesse directed at L'Ami's winning attempt as it keeps the King out.
<cro777>...<GM Erwin L'Ami suggested the following winning line:47...Bf4 48.Rh8 e4 49.dxe4 Ke5 50.Kxa3 Kd4 51.Kb4 a5+ 52.Kb5 a4> |
|
| Mar-26-14 | | csmath: <devere: I thought of it almost immediately, and I am not a grandmaster. The Black king must penetrate the White position to win, and playing e4 is the only way to do it. It was a missed opportunity for Karjakin.> You can think of that move but to calculate it timely and have a courage to play it is a different thing. I am not surprised that Karjakin just went for a safe draw. This is surely the ending I will remember for a long time. |
|
| Mar-26-14 | | Rhialto: <Win or draw? I'm not sure that even if Black can escape the checks, he can win from here.> I would guess that Black wins pretty easily in the diagrammed position, the idea is to walk the king over to g3, when Black escapes the checks: 65.Qb8+ Kc5 66.Qc7+ Kd4 67.Qa7+ Ke5 68.Qe7+ Kf4 69.Qf7+ (69.Qc7+ Kxf3) 69...Kg3 and now what? Note that if you try and arrange Qd6+ or Qe5+ in response to ...Kg3 (so that Black can't take on f3) then Kg2 escapes checks again, because the king there can escape to the h-file against side checks on d2 or e2 (but not c2 or b2). I am sure there are lots of variations but it looks to me like the Black king can always squirm his way over to g3. At that point the f3 and g4 pawns are doomed, and surely White cannot collect all three pawns, any one of which is enough to win once the Queen escapes h1. |
|
| Mar-26-14 | | csmath: And we are yet to see a complete analysis in all lines. I went through some morass but I am not sure it is all. It does seem that everything wins for black in 45...a3 46.Kc2 Be3 line provided that he plays e4 at some point. This is worth including in an endings book. |
|
| Mar-26-14 | | SirRuthless: I am not convinced that there arent improvements in the lines provided here. I let stockfish chew on the position after the e4 break to depth 40 and it seems to be finding endgames where both sides queens and black cannot break through due to whites pawns being the wrong color and an active white king.This thing may have just been drawn though karjakin would have been the only one with chances with correct play. |
|
Mar-26-14
| | tamar: <SirRuthless> What is Stockfish' sequence leading up to ...e4? This is a very complex ending, but lately both Komodo and Stockfish have shown increased skill at long endgames, which basically separates them from other engines, even Houdini. Komodo is still grinding away at the position after both sides Queen, and has been able to avoid perpetual check at the cost of two pawns. 
click for larger viewThis is tantalizingly close to tablebase certainty, as only Queens have to come off for the six piece evaluation to be available. Komodo continues 74...Qb1+ 75 Ka4 Qa1+ 76 Kb3 Qb2+ 77 Ka4 Be5 78 Qa1+
Now White can play 78 Qa3+ Qxa3+ 79 Kxa3, but cannot because that is a tablebase win for Black in 26 moves 
click for larger view So White plays 78 Kb4 and the analysis continues... |
|
| Mar-26-14 | | cro777: <tamar: Note that Komodo plays 48 Rd8 to lure the King to e6 so that when e4 occurs, Black has check on e8, and captures on e4 with the rook.> 45...a3 46 Kc2 Be3 47 Kb3 Bf4 48 Rd8 Ke6 49 Rh8 e4 50 Re8+ Kd7 51 Rxe4 Bg3 52 Re2 h3 53 Rc2
click for larger view53...h2 54. Rc1 d5 55. c5 d4 56. Kxa3 Bf2 57. Rh1 Bg1
click for larger view |
|
Mar-26-14
| | tamar: <cro777> Those moves look very familiar to me. Are they Stockfish's line also? Komodo traced the same path, but chose 54...Bf4 55 Re2 h2 56 Re1 Kc6 57 d4 instead of 54...d5 The only difference I can see is that Komodo preferred not to have the rook on the c file, and then valued the extra piece more than the final position you gave. But if it works, 54...d5 would be a cleaner way to win. |
|
| Mar-26-14 | | cro777: After 45... a3! <Forcing the white king to c2 to stop the a-pawn.> 46. Kc2 Be3! <It turns out f4 is the ideal square for the bishop.> 47. Kb3 Bf4 48. Rd8 Ke6
click for larger viewGM Jan Gustafsson gives the following line
49. Re8+ Kd7 50. Rh8 Kc6−+
click for larger view |
|
| Mar-26-14 | | cro777: <Tamar> I have tried different approaches. According to my analysis, 54...d5 (in this and similar lines) gives Black winning chances. |
|
Mar-26-14
| | tamar: <cro777> There are a few tricks even after 54...d5, but I am hoping this line works. It reminds me of an amalgam of Fischer Spassky game 13 elements, with the rook trapped by the bishop and pawn, and the King trying to control all the possible passed pawns, but it looks like it might work. In Gustaffson's line, 49 Re8+ looks premature, as it is only useful when Black has pushed ...e4. It gives the King a push where it wants to go anyways. |
|
| Mar-26-14 | | cro777: <Tamar> I don't think 49.Re8+ is best try for White, but Gustafsson's analysis is interesting as a sideline. |
|
Mar-27-14
| | tamar: Komodo confirms that <cro777>'s line is a win. 45...a3 46 Kc2 Be3 47 Kb3 Bf4 48 Rd8 Ke6 49 Rh8 e4 50 Re8+ Kd7 51 Rxe4 Bg3 52 Re2 h3 53 Rc2 h2 54. Rc1 d5 55. c5 d4 56. Kxa3 Bf2 57. Rh1 Bg1 Offers improvement with 56...Kc6 57 Kb3 Bf4 58 Rh1 Kxc5 59 Kc2 Kc6 Instead of putting the bishop on g1, this improvement stops all counterplay based on f4. As someone else noted, the square f4 is the ideal square for the bishop, and it returns to it again and again in the winning lines. |
|
 |
|
< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> |
Kibitzing
