< Earlier Kibitzing · PAGE 3 OF 3 ·
Later Kibitzing> 
Oct2905   snowie1: Us hillbillies understand en passant, since it specifies one certain move. But we shouldn't use words that only suggest a condition...like LOL 

Oct2905   phire: I think 36...Re3! was the better move...i see that some of you have found that too 

Oct2905   percyblakeney: Shredder sees 36. ... Re3, Bb4 and Rf4 as almost equally good, but prefers Re3. 

Oct2905
  patzer2: If 38. Re2!?, what is Black's best reply?
click for larger view(After 38. Re2!?, what's Black's best 38...?) 

Oct2905   Saruman: <patzer2> Maybe 38.Nxg4 or Rxe4, which both seem winning. 

Oct2905
  patzer2: Hmmnn! Fritz 8 is spoiling the fun here. After 36...Bc4! Fritz 8 finds 37. Rcc1!? Rfxc3 38. Nxc3 Rxc3 39.
Rxc3 Bxc3 40. Rd6 Bxa5 41. Rxa6 (Fritz has it as 1.44 @ 18 depth & 1289kN/s). However, despite Fritz 8's optimistic assessment, the resulting position doesn't look like such an easy forced win for Black: click for larger viewAnalysis by Fritz 8 (after 37. Rcc1 Rfxc3 38. Nxc3 Rxc3 39.
Rxc3 Bxc3 40. Rd6 Bxa5 41. Rxa6 in the diagram position): 1. (1.44): 41...Bb4 42.Rb6 Bc5 43.Rb7+ Be7 44.Kg3 Nd3 45.Rb6 g5 46.Kf3 Ne5+ 47.Kg3 Bc5 48.Rb7+ 2. (1.38): 41...Bd8 42.b4 Nxg4 43.b5 Ne3+ 44.Kf3 Nc4 45.Ke4 g5 46.Rc6 Ne5 47.Ra6 Nc4 

Oct2905   FORTRAN 77: <patzer2> 38.... rxc3 

Oct2905   jorgegatica: The practical sense of zugzwang considers the definitions mentioned previously and there are many games as examples like famous SämischNimzowitch and another of AlekhineNimzowitch but...
the purist origin defined it in a mutual and simultaneously way. That means that a real zugzwang position is achieved if and only if the moving side is forced to select between the complete set of legal moves and all of them are losing. And in the hypothetical situation that he could pass (without moving) then the other side would have the same problem. 

Oct2905
  patzer2: <Saruman> Well after 38. Re2!? both your selections appear to lead to difficult endgames with some winning chances, but neither one is the strongest reply: 1. <38 Re2!? Nxg4> 39. Rxe3
Nxe3+ 40. Kf3 Nf5! 41. Ne2 Rxc1 42. Nxc1 Bxa5 (1.62 @ 16 depth per Fritz 8) 2. <38. Re2 Rxe4> 39. Rxe4 Rxc3 40.
Rxc3 Bxc3 41. b4 ( 1.25 @ 16 depth per Fritz 8)
Like yourself, I also initially chose the 38. Re2!? Nxg4! line. However, apparently much stronger is the Fritz 8 solution <38. Re2 Rexc3!> 39. Nxc3 Nd3 40. Rf1 Rxc3 41. Ra2 Rxb3 (3.53 @ 13 depth) with two active pieces for the Rook. 

Oct2905
  patzer2: <Fortran77> Indeed after 38. Re2!?, the strongest reply is 38...Rxc3! If you found it without a computer I'm impressed. 

Oct2905   Saruman: <patzer2> Honestly, I hadnt calculated much further than 38.Rxe4 39.Rxe4 Rxc3 40.Rxc3 Bxc3 41.Ra4 Ke6 (Nc6?? Rc4 ) in order to prepare Nc6, which I thought should win. 

Oct2905
  patzer2: I'll have to concede that 36...Re3! is much stronger than 36...Bb4!? Following 36...Re3! 37. Rd4 Nxg4!, the threat of 38...f5 is much stronger than in the 36...Bb4!? continuation.
Fritz 8 assesses the position after 36...Re3! 37. Rd4 Nxg4 as 3.25 @ 15 depth, when Black wins without the difficult endgame possibilities in the 36...Bb4!? 37. Rcc1!? lines. 

Oct2905   Saruman: ...And also completly overlooking Fritz's 41.b4! 

Oct2905
  chessgames.com: This is actually an error. We knew that 36...Re3! is winning so we wanted to start the problem on move 37, but it didn't get fixed in time. And so either answer is acceptable. 

Oct2905
  patzer2: <Saruman> I too found myself in this position trying to calculate a difficult ending when an easier middle game tactic was available. At least for me, the lesson is that if I find a difficult win to look for another stronger and easier winning line. Also, I shouldn't hesitate to win two active pieces for a Rook when it's to my advantage.
<ChessGames.com> Thanks for the instructive puzzle. 

Oct2905
  kevin86: This was a very complicated position for so few pieces.I think even the computers have a difference of opinion on it,lol. 

Oct2905   Dudley: I was encouraged that Fritz also chose 36...Re3 as a solution followed by 37...Bb4 as that was my final answer also, although 36...Bb4 was my first thought. It seemed too obvious 

Oct2905
  Richard Taylor: Bb4 was the move I went for  I missed Re3. Very clever. But OTB Bb4 would be good enough... 

Nov0105   Averageguy: <chessgames.com><This is actually an error. We knew that 36...Re3! is winning so we wanted to start the problem on move 37, but it didn't get fixed in time. And so either answer is acceptable.> Who Hoo! So I did get it! 

Nov1407   arnaud1959: <chessgames.com> I don't see how this game can be a "zugzwang" exemple. Zugzwang doesn't only mean that one has no good moves. It's even more: it means that one loses (or only draws) because he has to move. Like in many pawn endings. 

Mar0212   Orhtej: This game is not a zugszwang..white loses even if he want to pass to move,so it is cLear that Black's threat is unstopable.. 

Mar0212
  AylerKupp: I agree with <Orhtej>. I believe that a zugswang position is one where not only will the side with the move suffer a (significant?) deterioration of their position if they are required to make a move but also that they DON'T suffer a (significant?) deterioration of their position if they are NOT required to make a move. The most straightforward example I can think of is the following K vs. K+P position:
click for larger view
White to move will lose since the only legal move White has is 1.Kg2 allowing 1...Ke2 and the black pawn queens. However, if White were allowed to pass on his move then the only move Black has that does not lose the pawn is 1...Kf3 which results in a stalemate. However, in this game White is probably lost regardless of whether he is forced to move or not. After 37...Re3 Houdini 1.5a at relatively low ply (d=24) considers the best play for both sides to be 38.Re2 Nxg4 (Houdini, unlike Fritz 8, considers 38.Rexc3 to be only White's 5th best move, evaluating the resulting position at [1.13]) 39.Rxe3 Nxe3+ 40.Kf3 Nf5 41.Ne2 Nd4+ 42.Ke3 Rxc1 43.Nxc1 Nb5 44.Nd3 Bxa5 45.Nec5 Bb6 46.Ke4 Nc3+ 47.Kf5 Ne2 (maybe 47...a5 is better? Eliminating the qside pawns would seem to reduce Black's advantage of B vs. N) 48.b4 a5 49.bxa5 Bxa5 50.Nf4 (and exchanging knights doesn't seem to be to White's advantage) 50...Nxf4 51.Kxf4 Bb6 52.Na4 Bf2 53.Kf5 Bd4 and evaluates the resulting position at [1.64].
click for larger viewBut this isn't a simple win for Black in spite of the 2pawn and B vs. N advantage. White has built a fortress on the light squares and it isn't easy for Black to make progress. Starting the analysis from this position Houdini evaluates it at only [1.00], d=28 after 54.Ke4 Bg1 55.Kf5 Bf2 56.Nc3 Bg3 57.Ne4 Bc7 58.Nc5 Bd8 59.Nb7 Be7 60.Na5 Bf8 61.Nc6 Be7 62.Nd4 Bd6 63.Nc6 Bc7 64.Nb4 Bd8 65.Nd5 Be7 66.Nb6 Bd6 67.Nd7 Ba3 68.Nb6 Be7 69.Nc4 Bb4 70.Nb6 Ba3 71.Nd5 Bd6 72.Nc3 Bb4 73.Ne2 Be7 74.Nd4 Bd6. Then again, Houdini is not a particularly good endgameplaying engine.
click for larger viewIn the game's ending position if White were allowed to pass then his position would be even worse ([4.36], d=25) after 38...Nd3 39.Nd5 Rxc2+ 40.Rxc2 Rxe4 since Black is a piece ahead and will soon lose additional pawns.
click for larger viewSo in the game's ending position White would definitely NOT benefit by passing on his turn to move so it wouldn't be zugswang. At least not by my definition/opinion. 

Mar0212
  AylerKupp: Well, in my previous post I was wrong (at least according to Houdini 1.5a) in thinking that after 38.Re2 Nxg4 39.Rxe3 Nxe3+ 40.Kf3 Nf5 41.Ne2 Nd4+ 42.Ke3 Rxc1 43.Nxc1 Nb5 44.Nd3 Bxa5 45.Nec5 Bb6 46.Ke4 Nc3+ 47.Kf5 perhaps 47...a5 is better than Houdini's recommended 47...Ne2 since the qside pawns are not eliminated. See position below.
click for larger viewBut Houdini disagrees, evaluating the resulting position at d=27 at only [0.49] after 48.Na4 Nxa4 49.bxa4 Be3 50.Nb2 Bd2 51.Nc4 Bb4 52.Ne3 Bd6 53.Nc4 Bc7 54.Ne3 Be5 55.Nd5 Bd4 56.Nf4 Bc5 57.Ne2 Be3 58.Nc3 Bd4 59.Ne4 Be5 60.Nc5 Bd4 61.Nb7 Bb6 62.Nd6+ Kg8 63.Nc4 Bc7 64.Kg6 Bd8 65.Kf5
click for larger viewIt seems that Black's (bad) bishop is partly tied down to the defense of his Pa5 and Black's can't make any progress in spite of his 2pawn advantage and pawns on both sides of the board. Maybe White's resignation was premature? So I decided to try Rybka 4.1 after the game ending 37...Re3. Rybka was more optimistic about Black's chances than Houdini, evaluating the position at [2.82] at d=23 after 38.Re2 Nxg4 39.Rxe3 Nxe3+ 40.Kf3 Nf5 41.Kf4 (here Rybka deviates from Houdini's 41.Ne2 but it's effectively only a transposition of moves since they both arrive at the same position after Houdini's 45.Bb6 and Rybka's 46.Nec5, with Rybka making an extra move for both White and Black) 41...Nd4 42.Ke3 Nb5 43.Ne2 Rxc1 44.Nxc1 Bxa5 45.Nd3 Bb6+ 46.Nec5 a5 (and here Rybka agrees with me that it's better to try to keep the apawn. That's one reason why I think that Rybka is a better endgame engine than Houdini. ƒº ) 47.Ke4 Nc3+ 48.Kf3 (and here is where I think that Rybka makes a mistake for White; 48.Kf5 establishes a kside bind on the light squares and denies Black any freedom of movement there) 48...Bc7 49.Nb2 f5 (now Black's king has an entry to g5 and White's Ph5 is vulnerable) 50.Nc4 Kf6 51.Na6 Bd8 52.Nc5 Nd5 53.Nb7. So in this instance Rybka's endgame play may not be as good as Houdini's.
click for larger viewBut is Rybka's [2.82] eval of the position realistic? Rybka certainly thinks so. Starting from the last diagram Rybka's eval is [3.58] at d=24 after 53...Bc7 54.Nc5 Nb4 55.Ne3 Bh2 56.Kf2 Nc6 57.Nd5+ Ke5 58.Nc3 Kd4 59.N5a4 Bd6 60.Nb5+ Ke5 61.Ke3 Bb4 62.Nb2 Kf6 63.Nc7 Bc5+ 64.Kd3 Ne5+ 65.Kc3 f4 66.Nd3 (this has to be a mistake, every piece exchange helps Black) 66...Nxd3 67.Kxd3 Kf5 68.Ke2 Ke4
click for larger viewAnd this position is clearly won for Black although 68...Kg4 getting 3 connected passed pawns might have been even more direct. Once Black's king is allowed to become active by 48.Kf3 instead of 48.Kf5) it's all over for White. 

Mar0212
  AylerKupp: But who's right in their evaluation, Houdini or Rybka? The only way I know to try to determine that is to go for a tie breaking opinion. So here's what Stockfish 2.2.2 calculated starting from the gameending position: 38.Re2 Nxg4 39.Rxe3 Nxe3+ 40.Kf3 Nf5 41.Ne2 Rxc1 42.Nxc1 Ke6 (here is Stockfish's rather obvious contribution to improving Black's play, activate the king) 43.Nd3 Bxa5 44.b4 Nd4+ 45.Kf2 (45.Ke3 would have been more forcing and saved a tempo besides) 45...Bb6 46.Nec5+ Kf5 (now that Black's king is active it's all over for White since the loss of the Ph5 can't be avoided) 47.Nxa6 Kg5 48.Ke3 f5 49.Nac5 Nc2+ 50.Ke2 Nxb4 51.Ne6+ Kf6 52.Nxb4 (possibly a mistake, every piece exchange helps Black. Maybe 52.Nf4 is better.) 51...Kxe6 53.Nd3 (and why not the desperado 53.Nxg7?) 53...Kd5 54.Kf3 Bc7 55.Nf4+ Ke5 (and 55...Bxf4 56.Kxf4 Ke6 is a forced win for Black per the Nalimov tablebases, but Stockfish at this point still doesn't have tablebase search capabilities) 56.Ke3 Bd6 57.Kf3 Kf6
click for larger view
Still a simple win for Black since White's hpawn goes after ...Kg5. So I decided to give Houdini the benefit of Stockfish's 42...Ke6 and see if it could find a win from that position. Houdini agrees somewhat, raising its eval from [1.13] after 38.Re2 Nxg4 39.Rxe3 to [1.60], d=26] after 38.Re2 Nxg4 39.Rxe3 Nxe3+ 40.Kf3 Nf5 41.Ne2 Rxc1 42.Nxc1 Ke6 43.Nd3 Bxa5 44.b4 Nd4+ 45.Kf2 Bb6 46.Nec5+ Kd5 47.Nxa6 Kc4 (erring, IMO, in going to the qside instead of the kside) 48.Ke3 Ne6+ 49.Kd2 Ba7 50.Ke2 f5 51.Kd2 Kb5 52.Nac5 Nxc5 53.bxc5 Bxc5 (now that the qside situation is cleared up, Houdini should hustle Black's king over to the kside) 54.Nf4 Kc4 55.Ke2 Bd6 56.Ke3 Bc7 (and here 56...Bxf4 57.Kxf4 is a win for Black per the 6man Nalimov tablebases even after 57...Kd4 58.Kxf5 but Houdini only supports 5man Gaviota tablebases so it can't figure this out) 57.Kf3 Bb8 58.Ke3 Be5 59.Ng6 Kd5 60.Ne7+
click for larger viewThis should be a win for Black since he's 2 pawns up and has B vs. N advantage but Houdini has gotten itself into a position where it is still not for Black to make progress. So what are my conclusions?
(a) Black should win after 37...Re3 but it's not that easy. Ivanchuk won the Gold medal and Nikolic the Silver medal in this 2004 Individual European Chess Championship, and Nikolic holding Ivanchuk to a draw would not have changed this result. (b) If you're using Houdini to help you analyze your endgame, make sure that you review its suggested lines and evaluations for accuracy. 

Jan1317   Saniyat24: After Black's 30th move the four Knights are facing each other. I haven't seen such a position before in any chess match...! 



< Earlier Kibitzing · PAGE 3 OF 3 ·
Later Kibitzing> 