< Earlier Kibitzing · PAGE 25 OF 25 ·
Later Kibitzing> 
Mar2716
  Tiggler: Excuse me, after 102...Ka4, I should have said. 

Mar2716
  Tiggler: <Honza Cervenka> <I would say that in Caruana's seat I would have won.> I don't think you could have got there before move 116. 

Mar2716   poulh: I'm wrong. 109 Bc3 Rb4+ 110 BXR stalemate 

Mar2716   suenteus po 147: Let's do The Time Warp again! Karjakin vs Caruana, 2013 

Mar2716   poulh: Wait a minute. 109 Bc3 Rb4+ 110 Kc5 Rb5+ 111 Kc6 and wins. 

Mar2716
  Honza Cervenka: click for larger viewThe position after 104...Rh5 (diagram), which was the only reasonable move preventing immediate mate, is butter and bread of R+B vs R endings. The winning technique is not so difficult to remember, and it is even possible to find it over the board: 105.Rb2! (threatening with mate from the other side) 105...Rh3 (the only alernative 105...Ka3 is followed by "subtle" 106.Rb1 with a threat of mate on a1, and after the only defense 106...Rh2 white plays 107.Rb6 Ka2 108.Ra6+ Kb1 109.Ra1+ Kc2 110.Ra2+ and the Rook on h2 falls, or 107...Rb3+ 108.Bc3 and black has nothing better than 108...Rxc3+) 106.Bf2! (covering h4 and forcing thus the Rook to leave hfile as 106...Ka3 loses quickly after 107.Re2 Ka4 108.Re8 Rh5 109.Be1 Ka3 110.Re2 with threat 111.Bb4+ and 112.Ra2#) 106...Rf3 107.Bc5 Rf4+ 108.Bd4 Rf3 (the difference between the Rook on the hfile and ffile is that Bishop now covers critical points f6 and f2 limiting thus defensive possibilities of black) 109.Rb4+ Ka3 (109...Ka5? 110.Rb7 and the game is over) 110.Rb7 Ka2 111.Rb2+ Ka3 112.Re2 Rf4 (or 112...Ka4 113.Be3 Ka3 114.Bc5+ Ka4 115.Ra2+ Ra3 116.Rxa3#) 113.Re6 Ka2 114.Re1 and the only way to prevent mate is 114...Rxd4+. 

Mar2716
  Eyal: <Tiggler> But you should take into account that before the actual mate, Black has to give up the rook in many lines, and once there's a piece capture the 50 moves rule isn't a problem anymore. Going with the tablebase's "main line" (i.e., the most precise moves from both sides) we get 105. Rb2 (instead of Caruana's Bf2) Rh3 106. Bf2 Rf3 107. Bc5 Rf4+ 108. Bd4 Rf3 109. Rb4+ Ka3 110. Rb6 Ka2 111. Rb2+ Ka3 112. Re2 Rf4 113. Re1: click for larger viewand here it says mate in 9, but in order to avoid immediate mate Black has to give up the rook. So it seems that with the most precise play Caruana could have won even taking into account the 50 moves rule  but the margin of error (or imprecision) was very small indeed. 

Mar2716
  chancho: From Chessbase:
<The moment when Peter Svidler played 102...Ka4 and realized he lost the draw ...> http://en.chessbase.com/Portals/All... <... only to have it returned by Caruana on move 111> http://en.chessbase.com/Portals/All... 

Mar2716
  Eyal: <... only to have it returned by Caruana on move 111> But as already noted it wasn't really on move 111, because by that stage
the 50 moves rule would have kicked in anyway before Caruana could win. 

Mar2716
  Eyal: From the press conference:
Svidler: "Now [after 102...Ka4] I suddenly realized that I'm getting the Philidor... and I started looking at the score sheet, I stopped looking at the board" Caruana: "I've solved this Philidor position so many times, and I just forget every time what I'm supposed to do" https://www.youtube.com/watch?v=O75... 

Mar2716
  Tiggler: <Eyal> Yes, well done. I agree completely. 

Mar2716
  Tiggler: If indeed Philidor solved this in the eighteenth century, then no wonder many people in France still say he was the greatest ever. 

Mar2716
  hoodrobin: Tomorrow will be the day! 

Mar2816   activechess55: <Honza Cervenka> Thanks for insightful analysis. Last capture or pawn move occured on 66th move. That means White has to execute capture or mate by 116th move. If he manages to capture rook before 116th move, the 50 move counting would begin afresh. That would mean Caruana really blew it on 111th move ! 

Mar2816
  Richard Taylor: <Marmot PFL: <This kind of position is still full of dynamic possibilities. I think it is actually easier to play against or with the Kings Gambit for example (for a GM) than this kind of position. I think the position suits Caruana especially in this round. But Svidler will make it a good fight.>
Just about what I was thinking. Caruana wants a small but lasting edge he can build on, not an all out attack.> He never played King's Gambit to my knowledge, and playing it now would be almost insane. Yes, it isn't wise to play riskily or to change one's openings in mid match unless the chips are really down. But the Ruy Lopez was played by Tal and he had many great wins with it...also Kasparov. So it is a myth that it is dull. It just depends how the game goes. At high chess level we mostly see games tending toward draws. This doesn't mean they are not interesting. But also we have seen some good wins. Caruana's good win over Anand from the English which is also a well known and pretty well analysed opening but has many possibilities. 

Mar2816
  Richard Taylor: I just looked at the game. Svidler came under pressure. I've forgotten if and when B and R vs R is a win or draw. I think it is in some positions. 

Mar2816   Ulhumbrus: Perhaps Caruana needs to practise playing Philidor's position not with a computer but with a training partner from both colours and all possible angles until he has the procedure nailed to his forehead 

Mar2816   RolandoFurioso: <Eyal> <Tiggler>
Starting in the position after 104...Rh5 and following the shortest path to mate (M18), it is indeed so that the rook will get captured in move 113 and thus well within the bound of the 50moves rule. Since, however, the mate in 18 lies outside the 50move bound, the critical question is whether Black could possibly follow a formally suboptimal line of play here that allows White to mate in 13, 14, 15 or 16 moves while delaying the capture of its rook so that it lies outside the 50moves bound. If so, then the position after 104...Rh5 would be a draw anyway. Though I suspect that this is really the case here!
The point is that one should consult a tablebase giving the distance to conversion (capture of the rook OR mate) rather then just the distance to mate in this very case since the distance to mate lies outside the 50move limit. 

Mar2816
  Eyal: <RolandoFurioso> Ok, that's a good point. So I looked at it again with what you said in mind, and it seems that if Black tries to delay the piece capture as much as possible he would still lose with precise play by White  ¯<exactly> on the 50th move... The opportunity to "improve" in this way arises after 112.Re2 in the line that I gave: click for larger viewand now 112...Ka4(!) [instead of ...Rf4] 113. Be3! Ka3 114. Bc1+/c5+ Ka4 115. Ra2+ Ra3 116. Rxa3#. 

Mar2816   RolandoFurioso: <Eyal>
Thanks for this excellent analysis! Indeed, in the line that you give, after112...Ka4(!) [instead of ...Rf4] 113. Be3! Ka3 114. Bc1+/c5+ Ka4 115. Ra2+ Ra3 that is, immediately before the rook gets captured, we have the position click for larger viewwith the FEN 50moves counter set to 98, which means that we are just 2 plies away from the 50moves bound. This delay of the rook capture comes at the "expense" of being mated in 12 rather than 18 moves. I verified this by a Stockfish analysis: Stockfish finds M18 (without tablebases) from the original position after 104...Rh5 (50moves counter amounts to 76) click for larger viewand as well with the 50movesrule counter being incremented to 77 but no longer when further incremented to 78, thus confirming your analysis. Very nice and instructive! This proves that Caruana had indeed the chance to win in this position; but there was absolutely no room for error, assuming that Svidler would have followed this line of "optimal" play where "optimal" is to be interpreted relatively to the 50moves rule rather than to a naive distancetomate tablebase application. See as well http://chessprogramming.wikispaces.... 

Mar3116   RookFile: Interesting stuff. If Caruana had won then he would have enjoyed draw odds against Karjakin. 

Apr0116
  cro777: Caruana: "I saw the Philidor position so many times and I just forget every time what I'm supposed to do!" The most important winning method in the rook and bishop versus rook endgame was discovered by Philidor long time ago: Philidor's study, 1749 (White to move wins)
click for larger viewThis is the basic winning position (king on the sixth and bishop on the fifth rank), referred to as the Philidor position. As Caruana pointed out, he saw this position many times. It is theoretically winning for White: with best mutual play White mates in 21. 1.Rf8+! <Black threatened to establish a secondrank defense> Re8 2.Rf7! Re2 <This is the best square for the rook> 3.Rg7 <Zugzwang> Re1 4.Rb7 Rc1 5.Bb3
click for larger viewThis is the ideal formation: the rook defends the bishop and the bishop takes away the check on d1. 5…Rc3 6.Be6 Rd3+ 7.Bd5 Rc3 8.Rd7+ Kc8 9.Rf7 Kb8 10.Rb7+ Kc8 11.Rb4 Rd3 12.Ra4 and Black is forced to sacrifice the rook. Of course, the mirror images of the basic Philidor position on the other three sides of the board also win. Caruana – Svidler: position after 104...Rh5 (The last capture was on move 66. White needs to capture the enemy rook before move 116)
click for larger viewThis is the Philidor position. With best mutual play White mates in 18 moves. 105. Rb2! Rh3 106. Bf2
click for larger view The ideal formation has been reached: the rook defends the bishop and the bishop controls the h4 square. 106…Rf3 107. Bc5 Rf4+ 108. Bd4 Rf3 109. Rb4+ Ka3 110. Rb6 Ka2 111. Rb2+ Ka3 112. Re2 Rf4
113. Re1 and Black has to sacrifice the rook
113... Rxd4+ 114. Kxd4 Kb3 115. Rc1 Kb2 116. Rc3 Ka2 117. Kc4 Kb2 118. Kb4 Ka2 119. Rb3 Ka1 120. Kc3 Ka2 121. Kc2 Ka1 122. Ra3# Caruana failed in his task: instead of 105.Rb2 he opted for 105.Rf2? This is incorrect: the bishop should be defended by the rook (105.Rb2! Rh3 106.Bf2). It is worth noting that when Caruana got a winning position, he had constantly about a minute on the clock, plus the 30second increment. 

Apr0116   epistle: Well at least we now have an endgame nonstudy named after Caruana. 

Oct2620
  g15713: E. 1.0
Black to move after 102. Rb7
click for larger viewThis a theoretical draw of Rook + Bishop versus Rook ending Syzygy 'https://syzygytables.info/?fen=7r/... Syzygy says any of these moves to draw at 102 for Black Ka6, Rc8, Re8, Rg8, Rh4 or Rh6
But 102...Ka4? should have loss as follows:
103. Kc4 Rh5+ 104. Bd4 Rh5 105. Rb2 Rh3 106. Bf2 Rf3 107. Bc5 Rf4+
108. Bd4 Rf3 109. Rb4+ Ka3 110. Rb6 Ka2 111. Rb2+ Ka3 112. Re2 Ka4
113. Be3 Ka3 114. Bc1+ Ka4 115. Ra2+ Ra3 116. Rxa3# checkmate 

Oct2620
  g15713: 103. Kc4 Rh5+ should be <Rh4+> 



< Earlier Kibitzing · PAGE 25 OF 25 ·
Later Kibitzing> 