Candidates Round #13
This was an intense round with three high-octane draws and one inconsequential win. Tailender Nakamura beat tailender Topalov to pull himself back to 50 per cent. But that mattered a lot less than the three draws played out by Caruana, Karjakin and Anand, A win by any member of this troika would have altered the last round dynamics.
Caruana-Svidler played a sizzler and Caruana finally flubbed a well-known technical win, albeit one that is notoriously difficult. if he had converted that position, the American would have pulled ahead.
As it happened, that 100-move-plus draw might have set things up for Karjakin, who would also have the solo lead if he beat Aronian. Only, Karjakin was under massive pressure and barely survived through another 100-mover, where he defended with his customary coolth (if one may borrow a word coined by the late Sir Terry Pratchett).
Those two games went on for a very long time. All four players looked utterly physically exhausted by the end. The draws meant that the Karjakin-Caruana clash in Rd 14 would be critical.
Meanwhile, Anand-Giri was drawn in "only" 50-odd moves but it was also an exciting game. Giri got adventurous and sacrificed a piece against Anand. It was hard to make sense of what was going on, But that also burnt out into a draw. If Anand had put this one away, he would have tied Caruana and Karjakin and had everything to play for in the last round. But with the benefit of hindsight, Anand never looked like winning.
Let's look at the games in ascending order of complexity and importance.
Topalov was not in time trouble (he very rarely is). Nor are these variations hard to calculate. It is an indicator of his very poor form (and probably, lack of motivation by this stage) that he played 33.Rg3 Re2 34.Ra1 Nxe5 35.Ra8+ Kf7 36.Bh7 g5 37.Bg8+ Kf6 38.Rf8+ Kg7 39.Re8 Ng6 40.Bxe6 Nf4 (0–1). If 41. Bd7 Rxe8 42. Bxe8 Ne2+ kills it
Anand-Giri looked nice and equal in the diagram.
On normal moves like 24. -- Qc5 25. Bd3, it should just burn out. Instead, Giri tried the amazing 24.-- Bxf2+!?
After the more or less forced sequence 24. Bxf2+ 25.Kxf2 Qb6+ 26.Kf1 Nh5 27.g4 fxg3 28.Bd3 Rxe1+ 29.Rxe1 Nf4 30.Nd4 g6 31.Be4 Qf6 it's hard to tell who is better, if indeed anybody is. Anand returned the exchange on e1 and then sacrificed his entire queenside to pile onto f7. If Giri doesn't take the perpetual towards the end, Ng4-h6 is a potent threat from white's side. Also black must beware of white marching Kh6. But of course, Giri took the perpetual.
This game pretty much finished Anand's challenge in the event. Of course, when Anand's game ended, Aronian was pressing against Karjakin and Caruana-Svidler was unclear. But both Caruana and Karjakin would have to lose to give Anand a reasonable chance in Rd 14.
Aronian had built a nice edge out of an English Opening
Crazy tactics start at this point. Both players assessed the normal 25, -- c6 26. Rxb5 cxb5 27. Nd6 s good for white. The engines agree that this is a little superior for white.
The actual play went 25.-- Nxa3?! 26.c6! Qe7 27.Qa2 bxc6 28.Rxc6 Bf7 29.Rc5! This hits Nd5 and cuts off the Queen's defence of the Kt-a3. It triggers the next tactical flurry. Black went with 29.-- Nxe3!? 30.Qxa3 Rxb2 31.Qxb2 Nxg2 32.Kxg2 a3 33.Qb7 Qd8 34.Qxc7 Qxc7 35.Rxc7 Bd5 36.Rc5 a2 37.Bc3 Bg8 38.Ba1 Rb8 39.Ra5 Rb1 40.Bc3 Rd1 and somehow black is still hanging on. Aronian must have thought he was winning. He tried all sorts of reshuffles and complicated piece placements over the next umpteen moves. But Karjakin just kept finding good defensive resources.
Caruana - Svidler was more up and down.
Black has generated serious pressure in the centre. He seems better because the Qd5 has more influence than Qa4. But the last move 34.. - Rae8 misses a cute little tactic. White played 35.dxe5 Nxe5 36.Nxe5 Rxe5 37.Bxf4! and now the apparently excellent 37. -- Rxe4? 38. Rxd5 Rxe1+ Kh2 39. Nxf4 is refuted by an X-ray 37. -- Rxe4? 38. Qxe8+! and the white queen's placement suddenly makes sense. So Svidler had to play 37.-- Qb5 38.Qxb5 Rxb5 39.Bd6 Kh8 40.g4 Nf6 41.Nxf6 Rxe1+ 42.Rxe1 Bxf6 43.Re8+ Kg7 44. Bb4 Re5 45. Rc8 and white is better because he can target the queenside pawns. But this is by no means easy to win.
Caruana soldiered on to reach this position
White is two pawns up but even this is probably not winning. The pure rook endgame with bishop and rook is technically drawn (though people have been known to lose it). Black keeps his rook on the h-file (as distant as possible) and rook goes to h1 and starts checking from the first rank, or the h-file if white tries to walk in with his king. It's impossible to make progress against good defence.
The pure bishop endgame would also be drawn because the a8 queening square cannot be controlled, Black just sacrifices his bishop for the c-pawn at some stage and sits on a8 with his king. White will deliver stalemate.
So white has the job of keeping both pieces on and trying to finagle something. Black played for another known technical draw with 63.Kd3 Rh5 64.c5 Kb7 65.Rg6 Bxa5 66.Bxa5 Rxc5 and the R+B Vs Rook ending is drawn. Note that the 50-move rule has kicked in with the last capture on 66.
People, meaning a genius called Andre Danican Philidor started looking at this material set up in the 18th century. It's drawn with correct play from most situations and Philidor found several of the most critical set ups. Computer analysis gives us an exact countdown of any given set up. All five piece positions have long since been solved.
There is also the 50-move rule. This favours the defender because quite often, forced wins go over 50. For a while, this material used to be played out with a special extension to 75 moves because of that factor. But it's long since reverted to 50-moves.
In practice, White will always push the black king to the edge of the board and guard checks against his own king by interposing bishop. But black has a choice of squares to step to with his king. Judicious retreat will hold the draw. He must also keep his rook active of course and there are also stalemate defences in some situations.
Despite centuries of study and complete solutions, this is an incredibly hard position to play, Offhand, I can recall several superGMs, such as Grischuk, Ulf Andersen and Jan Timman who have played this out and made serious errors. We can add Svidler and Caruana to that list.
Svidler has just made a terrible error with 102-.. Ka5-a4?. Instead 102.--Ka6 103. Rb1 Rc8! holds the draw because the white king can't get "mating opposition" with Kc6. For that matter, 102. -- Rg8 or 102. Re8 holds the draw.
After 103.Kc4! Rh4+ 104.Bd4 Rh5 white could win with the Philidor method which is outlined in in any standard textbook such as Basic Chess Endings.
He can play 105. Rb2 Rh3 (to stop Ra2#) 106. Bg3! (prevents Rh4+ and causes zugzwang) Rf3. Pulling the rook away from the edge of the board is vital. Now white plays 107. Bc5! Rf4+ 108. Bd4 Rf3 109. Rb4+! Ka5 110 Rb7 and mates due to f6 control. (This is one reason why Rf3 needs to be forced).
Or 109. Rb4+ Ka3 110. Rb7 Ka2 111. Rb2+! Ka3 112. Re2 (Bc5+ is the deadly threat) Ka4 113. Be3! (Another reason to force Rf3) Ka3 114. Bd6+! Ka4 115. Ra2+ Ra3 116, Rxa3# In both cases, mates is within the 50-move limit.
Instead Caruana frittered the big chance away with 105.Bf2 Rg5 106.Rh7 Rg4+ 107.Bd4 Rg5 108.Rh8 Rb5 109.Ra8+ Ra5 110.Rb8 Rh5 111.Bf6 Ka5 112.Bc3+ Ka6 113.Bd4 Rh6 114.Be3 Re6 115.Rb3 Rc6+ 116.Kd5 Draw.
So it all boiled down to the last round.