< Earlier Kibitzing · PAGE 18 OF 18 ·
Later Kibitzing> 
Jan2813   Olavi: Well HÃ¼bner found those improvements without a comp, otherwise he wouldn't have missed 24.Bb7 in <CharlesSullivan>'s third diagram. And I think he has stated that too. But it's always easier to correct an existing analysis. 

Jan3013
  AylerKupp: <CharlesSullivan> It is true that Huebner, or for that matter Kasparov or Byrne, didn't have computers. Critter 1.6 has the following lines after 14.Rfd1 Nd3 and either 15.Nd4 or 14.Nf4: After 15.Nd4: [0.71], d=31: 15. ... Ne4 16.Nxe4 dxe4 17.Bb2 Rc8 18.a4 Rc7 (instead of Kasparov's 18...Qg5) 19.Bf1 Qc8 20.Ra2 Ne5 21.Bxa6 Qxa6 22.Ba1 Qc8 23.Qe2 a6 24.Rad2 Rd8 25.Kg2 Nd3 26.Bb2 Qb7 27.Kg1 Qd5 28.Rxd3 exd3 29.Qxd3 b5 30.axb5 axb5 31.Qd2 Bf8 32.Qa5 Rc2 33.Rb1 Ra8 34.Qxa8 Qxa8 35.Nxc2
click for larger viewA long, possibly meaningless line, but I don't see how White can survive after 35...Qa2. Apparently the horizon effect in action since the [0.71] eval following 15...Ne4 is therefore much too low (yet another example why you <MUST> check all engine analysis). Starting from this position Critter evaluates the resulting position at [9.33], d=28 after 35...Qa2 36.Rf1 Qxb3 37.Nd4 Qxb2 38.Nc6 b4 39.Nxb4 Qxb4 and the rest is not particularly interesting, R+P vs. Q+B.
click for larger viewAfter 15.Nf4: [0.60], d=30: 15...Ne4 16.Nxe4 dxe4 17.Rab1 Rc8 18.Bb4 (same as Huebner's suggested line) 18...Nxf4 19.gxf4 Qxd2 20.Rxd2 Bd3 21.Rbd1 a5 22.Bd6 Re6 (here Critter deviates from Huebner's suggested 22...Bc3) 23.Bh3 f5 24.Be5 (a strange move, why does White need to give up a pawn, even if it's doubled? Is 24.Ba3 that much worse?) 24...Bxe5 25.fxe5 Rxe5 26.Bf1 Bxf1 27.Kxf1 Kf7 28.Rd7+ Re7 29.R1d6 Rc1+ 30.Kg2 Rc2 31.Rxe7+ Kxe7 32.Rxb6 Rxa2 33.Rb7+ Ke6 34.Rxh7 f4 35.Rh3 g5 36.exf4 gxf4 37.Kf1 Kf5 38.Rc3 f3 39.Ke1 Kf4 40.Rc7 Re2+ 41.Kf1 e3 42.Rc4+ Ke5
click for larger viewAnd this is indeed a draw. At d=26 Critter evaluates the position at [0.00] after 43.fxe3 Rxe3 44.h4 Rxb3 45.Ra4 Rb5 46.h5 Kf6 47.h6 Kg6 48.h7 Kg7 49.Rf4 Rb1+ 50.Kf2 Rb2+ 51.Kxf3 Rb3+ 52.Ke2 Kxh7 53.Rh4+ Kg6 54.Rg4+ Kh5. So if Black is to have any winning chances, they must probably be found somewhere after 26.Bf1 when Black is, at least temporarily, a pawn up.
click for larger viewMaybe 37...f3 instead of 37...Kf5 to gain a tempo (White must play 38.Ke1 to prevent mate) and cut off the White rook from the qside, at least temporarily.
click for larger view 

Jan3013
  AylerKupp: (continued) And indeed Rybka 4.1 agrees (hooray for me!). It evaluates the resulting position after 37...f3 39.Ke1 at [13.72], d=28 after 38...Re2+ (forcing White back and gaining another tempo due to the mate threat) 39.Kf1 Rb2 40.Ke1 Rxb3 41.Rh5 a4 42.Ra5 a3 43.h4 Kf7 44.Ra8 Kg6 45.Ra5 Kh6 46.Ra7 (White is in zugswang) 46...Kh5 47.Ra6 Kxh4 48.Ra4 Kh5 49.Ra5+ Kg4 50.Ra8 Rb1+ 51.Kd2 Rb2+ 52.Kc3 Rxf2 53.Rg8+ Kf5 and the 2 passed pawns get through.
click for larger viewSliding forward, at first glance Rybka seems to find a better way than Critter starting from the position after 26.Bf1. It evaluates the resulting position at [0.67], d=28 after 26...Bxf1 27.Kxf1 Rec5 (instead of Critter's 27...Kf7) 28.Rd6 R5c6 29.Kg2 R8c7 30.Kg3 Kg7 31.Rxc6 Rxc6 32.Rd7+ Kh6 33.Rb7 a4 34.bxa4 Rc4 35.Ra7 Rc5 36.Ra6 Ra5 37.Rxb6 Rxa4 38.Rb2 Kg5 39.Rc2 Ra3 40.Kg2 Kg4 41.h3+ Kh4
click for larger viewAn advantage for Black to be sure (1 pawn island instead of 3) but I don't think that it is enough to win. Rybka doesn't seem to have a clue how to proceed, apparently basing its evaluation on positional considerations only rather than a constructive plan of action. Starting from this position it evaluates the resulting position at [0.76], d=37 after 42.Rc7 h5 43.Rc6 Kg5 44.Rc2 Ra5 45.Rb2 h4 46.Kg1 Kf6 47.Kg2 Ra3 48.Kf1 Ra6 49.Kg2 Ra5 50.Rc2 Ra3 51.Rc6+ Kg5 52.Rc2 Ra6 53.Kg1 Kh5 54.Kg2 Ra8 55.Rd2 Ra3 56.Rc2 Ra5 57.Rb2 and Black has made no progress. Yet another example of needing to check engine analysis.
click for larger viewOf course, none of these lines are forced (although I couldn't find any obvious improvements for either side in the initial 15.Nf4 line other than 37...f3) so they're hardly conclusive. But I think that they do show that Black retains the better chances throughout despite what Huebner and Kasparov (without computers, of course) thought. 

Jan3013
  AylerKupp: (continued) To change the subject somewhat, as <kingscrusher> indicated slightly over a year ago, he updated his analysis of the game and you can find it here: http://www.youtube.com/watch?v=6S7_.... I think it is his best video ever, with lots of analysis of Fischer's late game possibilities. Very highly recommended. You owe it yourself to set aside an hour of your time, get comfortable (48 minutes long), and watch it. It makes Fischer's play look even more remarkable than it appeared at the time. And Byrne should also be commended for seeing that his position at the end was hopeless, even if everybody else at the scene didn't think so. As <Olavi> said, it's always easier to correct an existing analysis, with or without computers. And it is usually more fun, even if likely meaningless, to do the subsequent analysis without the clock ticking. And compare the cramping effect of Fischer's Nd3 with the similar cramping effect of Kasparov's Nd3 in Karpov vs Kasparov, 1985. You can also see a very informative 30min video on the game with commentary by Kasparov at http://www.youtube.com/watch?v=6z4Z.... Great minds and great chessplayers obviously think alike. 

Jan3113   CharlesSullivan: <AylerKupp> Good to have your analysis! (A) Regarding 15.Nf4, I am still fairly certain that White can hold the game. You give 15...Ne4 16.Nxe4 dxe4 17.Rab1 Rc8 18.Bb4 Nxf4 19.gxf4 Qxd2 20.Rxd2 Bd3 21.Rbd1 a5 22.Bd6 (Huebner) 22...Re6 23.Bh3 f5 24.Be5 Bxe5 25.fxe5 Rxe5 26.Bf1 Bxf1 27.Kxf1 Kf7 [Houdini 1.5a gives 27...Rec5 (0.23) as best after 37 iterations] 28.Rd7+ Re7 29.R1d6 Rc1+ 30.Kg2 Rc2 31.Rxe7+ Kxe7 32.Rxb6 Rxa2 [Houdini 1.5a prefers 32...f4 33.exf4 e3 (0.12) after 8 hours, 40 iterations] 33.Rb7+ Ke6 34.Rxh7 f4 [or 34...g5 35.Kg3=]:
click for larger view
Now 35.exf4 [instead of 35.Rh3?] seems perfectly adequate for equality: 35...e3 36.Rh3 Rxf2+ [or 36...exf2 37.Rg3 Kf6 38.Kf1=; or 36...e2 37.Re3+ =] 37.Kg1 Rxf4 38.Rxe3+ Kf6 39.Rc3=. (B) Regarding 15.Nd4, I agree that White ends up with a difficult game, but I cannot find a win for Black. You give 15...Ne4 16.Nxe4 dxe4 17.Bb2 Rc8 18.a4 (Huebner) 18...Rc7 [My previous post analyzed Kasparov's 18...Qg5. I've also looked at 18...h5 19.Bh3! and 18...Nxb2, both of which also lead to drawish positions]
click for larger view
A preliminary, overnight analysis gives  instead of 19.Bf1  19.a5 [Houdini 1.5a (0.60) after 36 iterations, 16 hours] 19...b5 20.b4 Qc8 [I've started looking at 20...Rc4] 21.Ra3 Bb7 22.Rc3 a6 23.Bf1 Rc4 24.Rxc4 bxc4 25.Bc3 Qd7 26.Rb1 looks solid, albeit passive, for White. So far the computers give 26...Qd5 [transposing is 26...Bc6 27.Bg2 Qd5] 27.Bg2 Bc6 28.Nxc6 Qxc6 29.Bxg7 Kxg7 30.Qc3+ f6 31.Rd1 (0.70). 

Jan3113
  AylerKupp: <Charles Sullivan> This has always been one of my favorite games. And I consider 18...Nxg2 one of the best moves ever played, possibly the best ever given the concept behind it. Fischer must have been able to see this move and its follow up before playing 12...e5, since both 14.Rfd1 and 15.Qc2 are, on the surface, very reasonable moves for White. Poor Byrne apparently didn't have any idea of what was coming! It's difficult (impossible?) in games to reach a definite conclusion using engine analysis with such long lines. So many possibly promising lines get pruned during the search and, of course, different engines have different evaluation functions and search tree pruning heuristics that they come up with different lines and evaluations. In particular, the moves given by the engines towards the end of the lines can be highly suspect, like my 37...f3 instead of Critter's 37...Kf5. So I laugh when people vehemently argue about the relative merits of two 20move long unforced engine lines. I had run a Houdini 1.5a analysis a short time ago to see what it thought was White's best 14th move. It evaluated 14.Rad1 to be the best (like everyone else did), at [0.25], d=31 but it considered Byrne's 14.Rfd1 to be second best, evaluating it at [0.54], d=31 after 14.Rfd1 Nd3 15.Nf4 Ne4 16.Nxe4 dxe4 17.Rab1 Rc8 18.Bb4 Nxf4 19.gxf4 Qxd2 20.Rxd2 Bd3 21.Rbd1 a5 22.Bd6 Re6 23.Bh3 f5 24.Be5 Bxe5 25.fxe5 Rxe5 26.Bf1 Bxf1 27.Kxf1 Rec5
click for larger viewAt the time I thought that this looked pretty grim for White to me. A pawn down and 3 pawn islands vs. 2 in a KRR+KRR ending. But I didn't look further. In your (A) line I foolishly assumed that 35.exf4 would lead to a loss after 35...e3 but again I didn't bother to look further so I assumed that Critter was probably right about 35.Rh3. Obviously things are not as simple as they looked to me on the surface! I had Rybka 4.1 analyze the position after 35.exf4 35...e3 and, indeed, it couldn't find a win after 36...e3, coming up at low ply with [0.11], d=22: 36.Rh3 Rxf2+ 37.Kg1 Kf5 38.Rxe3 Rxf4, pretty much a draw.
click for larger view35.exf4 e3 36.Rh3 e2 doesn't lead to anything either. And Rybka doesn't find anything better than 34...f4 at d=26, evaluating the resulting position at no better than [0.10] for Black. So, yes, this line after 15.Nf4 seems to hold for White. With regards to your (B) line I decided to give Critter another chance, starting with the position after 15.Nd4 Ne4 16.Nxe4 dxe4 17.Bb2 Rc8 18.a4. It still selected 19.Bf1 as White's best move [0.71], d=25 over Houdini 1.5's 19.a5 at [0.88], d=25. So Critter and Houdini pretty much agree, a difficult game for White but no forced win for Black. So I tried Stockfish 2.3.1. Stockfish at d=34 prefers 19.a5 over 19.Bf1, with an eval of [1.13] vs. [1.37]. This is Stockfish's line: 19.a5 b5 20.b4 Rc4 21.Bc3 Qc7 22.Ra3 Nxf2 (better late then never !) 23.Qxf2 Rxc3 24.Rxc3 Qxc3 25.Qd2 Qc8 26.Qe1 Qg4 27.Nc6 Qe6 28.Nd4 Qd5 29.Ne2 Qc4 30.Rd7 Bf8 31.Rxa7 Bc8 32.Qf2 Re7 33.Ra8 Rc7 34.Qf4 Bxb4. But this is another example of the horizon effect; starting the analysis from this position Stockfish evaluates the resulting position at [0.00], d=30 after either 35.Kf2 or 35.h4, both petering out to a draw by repetition. So it looks like you may be right again, no forced win for Black after 15.Nd4 either, at least not this way. Did you look at the <kingscrusher> video? After 15.Nf4 Ne4 16.Nxd4 exd4 17.Rab1 Rc8 he doesn't look at 18.Bb4 but instead came up with the following after 18.Nxd3 Bc3 19.Qe2 Bxd3 20.Qg4 h5 21.Qh3 Bxb1 22.Rxd8 Rcxd8 and White is helpless against 23...Rd8 and 24...Bd3. Yet another example of the dynamic potential of Black's pieces. 

Feb0113   CharlesSullivan: <AylerKupp> Yes, I looked at the kingscrusher video a couple of months ago. Actually, his analysis got me thinking that there was more going on in the position at moves 14 & 15 than I had previously thought. 

Feb2113
  morfishine: <AylerKupp> Thanks for posting Kingscrushers video on this game! 

Apr0713   Tigranny: Wow! Amazing game, probably among my top favorites! The true Game of the Century! 

Apr0713   RookFile: In retrospect you realize that a3 bishop of white's was pretty worthless. 

Apr1313   IndigoViolet: From Game 48's Introduction in <MSMG>: <K.F. Kirby [Kenneth Kirby ], editor of the South African Chess Quarterly, summed up the astonishment and admiration of the chess world when he wrote: <The Byrne game was quite fabulous, and I cannot call to mind anything to parallel it. After White's eleventh move I should adjudicate his position as slightly superior, and at worse completely safe. To turn this into a mating position in eleven more moves is more witchcraft than chess! Quite honestly, I do not see the man who can stop Bobby at this time...>> 

Dec3113   PJs Studio: This game made a big impression on me as a kid. I couldn't believe Fischer took the bishop instead of the rook on move18... when studying the game. The depth of the calculation was so deep Fischer could see the net cast. Byrne resigned and the IMs commenting on the game thought he was crazy for resigning. then...Fischer showed them the continuation. I bet they felt stoopid! Tactics became my lifeblood after the discovery of this game. Funny but true: I found that studying modern GMs annotations of their own games in New in Chess (NiC) back in the early 90's that almost all decisions they make (after the opening) are based on large amounts of tactical considerations. 

Mar2014   chathuranga: Byrne is a piece up,cloud't he play 22.Nf3 to defend?can somebody answer? 

Mar2014   Petrosianic: No, he can't defend. That's why he resigned. 

Mar2014
  Jim Bartle: 22...Qh3+ and there's no defense. 

Mar2114   chathuranga: On Qh3+,23.Qg2,then I think there is no sudden mate.Am I right? 

Mar2114
  perfidious: <chathuranga: On Qh3+,23.Qg2,then I think there is no sudden mate.Am I right?> As Fischer pointed out in his final note, after 22....Qh3+ 23.Qg2 comes 23....Re1+ 23.Rxe1 Bxd4 and wins. 

Mar2214
  shivasuri4: <chathuranga>, after 22.Nf3 Qf5 23.Qf4 Qxf4 24.gxf4 Bxc3, White will have a significant material disadvantage, with many hanging pieces. You are right in that there's no immediate mate. 

Mar2214   RandomVisitor: After the suggested improvement 14.Rad1:
click for larger viewDeep Rybka 4.1 x64:
<[0.21] d=27 14...Qc8> 15.Bd6 Nd3 16.Nxd5 Nxd5 17.Bxd5 Rd8 18.Bxa8 Qxa8 19.Ba3 Bf6 20.e4 Qxe4 21.Qe3 Qc6 22.Be7 Re8 23.Rxd3 Bxe7 24.Nd4 Qd7 25.Re1 Bxd3 26.Qxd3 Rd8 27.Qc4 Bc5 28.Nf3 Qf5 29.Kg2 a5 

Mar2214   RandomVisitor: Fischer: [after 14.Rad1] "Finally I found 14...Qc8!  the only move to keep the pressure. Now on 15. Nxd5 Nxd5 16. Bxd5 Rd8 17. f4 Rxd5! 18. Qxd5 Bb7! 19. Qd8+ (if 19. Qd2 Qh3! 20. Nd4 Ng4 21. Rfe1 [or 21. Nc2 h5 with a strong attack] Nxe3! should win) Qxd8 20. Rxd8+ Rxd8 21. fxe5 Bxe5 with a better ending. And on 15. Rc1 Qd7! 16. Rcd1 Red8 Black has finagled a precious tempo, since his Queen is on d7 instead of d8. After 14...Qc8! relatively best is 15. Bb2 (if 15. Qc1 Ne4 16. Nxd5 Bxe2 17. Bxe4 Kh8! wins the Exchange. One possible line is 18. Qxc8 Raxc8 19. Ne7 Rc7 20. Rc1 Rd7 21. Rfe1 Bf3!) although Black keeps the initiative with ...Qf5." 

Mar2214   RandomVisitor: After <14.Rad1 Qc8 15.Bd6> Nd3 16.Nxd5 Nxd5 17.Bxd5 Rd8 18.Bxa8 Qxa8 19.Ba3 Bf6 20.e4 Qxe4 21.Qe3 Qc6 22.Be7 Re8 23.Rxd3 Bxe7 24.Nd4 Qd7 25.Re1 Bxd3 26.Qxd3 Rd8 27.Qc4 Bc5 28.Nf3: click for larger view Rybka 4.1 x64:
<[0.17] d=27 28...Qf5> 29.Kg2 Kg7 30.Re2 Rd3 31.Ne5 Rd4 32.Qc2 Qe6 33.Nf3 Qc6 34.Qc3 Kg8 35.Rc2 Rd1 36.Rc1
[0.16] d=27 28...a5 29.Re2 Qf5 30.Kg2 Kg7 31.Qc2 Qd5 32.h4 Qb7 33.Qc3+ Bd4 34.Qc2 Rc8 35.Qe4 Qxe4 36.Rxe4 Bc5 37.Re2 Rd8 38.Ne5 f5 39.Nc4 Kf7 40.Rc2 Bd4 41.Kf3 Can anyone find an improvement for black? 

Mar2514   RandomVisitor: After 14.Rfd1 Nd3:
click for larger viewDeep Rybka4.1 x64
<[0.75] d=27 15.Nd4> Ne4 16.Nxe4 dxe4 17.Bb2 Rc8 18.a4 Qd7 19.Bf1 Qd5 20.a5 Nxb2 21.Qxb2 Bxf1 22.Rxf1 bxa5 23.Qd2 Rc5 24.Ra4 Bxd4 25.Qxd4 Qxb3 26.Qxc5 Qxa4 27.Qxa7 Kg7 28.Qc7 Qb5 29.Ra1 Re5 30.Qc2 

Mar2614   RandomVisitor: After 11...Re8: <Rybka4.1> [0.10] d=28 12.Rc1 Rc8 13.Rc2 h6 14.Re1 g5 15.h3 e6 16.Qd2 Qd7 17.g4 Bf8 18.Bxf8 Rxf8 19.Rcc1 Nb4 20.Red1 Nd3 21.Rc2 Rc7 22.Ng3 Nb4 23.Rcc1 Rfc8 24.a3 Nd3 [0.11] d=28 12.Re1 Rc8 13.b4 Bc4 14.b5 Na5 15.Nf4 e6 16.Nd3 Bxd3 17.Qxd3 Nc4 18.Bb4 Rc7 19.Rec1 Qd7 20.a4 Rec8 21.Rc2 Bf8 22.Na2 Bd6 23.Rac1 Qe7 Note that after 12.Rc1, 12...e5 cannot be played because of 13.Nxd5 Nxd5 14.Rxc6. 

Apr0814   MrJafari: Before seeing this game,I thought that Fischer was a great player,but after,now I think he was a chess magician too! This game impressed me... 

Apr1914   Ed Frank: I want to play one game, just one game that makes me happy and aghast as this one. 


< Earlier Kibitzing · PAGE 18 OF 18 ·
Later Kibitzing> 