< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> 
Sep2715   diagonalley: hmmmmmm... not only 'insane' but machine v machine... this ain't chess, it's an academic exercise... HAH! it even ended in a draw! ... i'm going back to bed 

Sep2715
  jahhaj: 23. Rh8, amazing. What does the computer say? Oh wait ... 

Sep2715   tea4twonty: 24♖d7 threatens #2 .... monster ♘ & ♖ 

Sep2715   The Kings Domain: Interesting game by machines. I guessed the puzzle the first time around but changed my mind numerous times afterwards. I still don't get why black sacrificed its rook. 

Sep2715   Al2009: It seems that 24. Bd8! is winning faster than 24. Rd7 a) 24...Rxd8 25. Rxd8 Qxd8 26. Nxd8 Ba8 27. Nb6
b)24...Ng6 25. Qg4 

Sep2715
  morfishine: This game needs to be enginetested 

Sep2715
  agb2002: White has a rook for a bishop and a pawn.
Black threatens 23... Nxf4.
The first idea that comes to mind 23.Qxf3 loses to 23... Qe5 with the double threat 24... Qa1# and 24... Qxe6. Another idea is 23.Rh8+, trying to deflect either the king or the knight: A) 23... Kxh8 24.Qf7 with the double threat 25.Rh1 and 25.Bf626.Bxg7+ looks winning. B) 23... Nxh8 24.Rd7 with the threat 25.Qxf8+ Rxf8 26.Rxg7# but I don't know what to do after 24... Nf6. This is all I can do today. 

Sep2715
  Once: I stared at 23. Rh8+ as an ant might stare at a skyscraper. Coo! Ain't that a big one! I wonder what it does? Computer vs computer. On a Sunday. I admit it. I was intimidated. 

Sep2715   wooden nickel: 24.Bd8 looks like a logical follow up but doesn't quite make it due to 24... Ng6 (23.Rh8+ Nxh8 24.Bd8)
click for larger viewIf 24... Qb5, then 25.Qxf8+ Kh7 26.Rh1+ Kg6 27.Nf4#
click for larger view 

Sep2715   RandomVisitor: Black threatens Nxf4 and the white queen has no good square to move to, for various reasons as shown below. Black threatens 1. a knight fork on f2, 2. to capture the white bishop on g5 if Ne6 moves, 3. Qe5 followed by Qa1+, 4.Qf5, threatening to capture either Bg5 or Ne6. The deflection Rh8+ Nxh8 is really an exchange sacrifice, since the followup Rd7 Ng6?! threatens Qf7+. Nf6 Bxf6 returns material. Since white was up the exchange before this mess, the position returns to a sortof balance. click for larger viewKomodo9.264bit:
<+0.00/23 23.Rh8+ Nxh8 24.Rd7 Nf6 25.Bxf6 Ng6> 26.Qh2 Qf5 27.Nxf8 Qf4+ 28.Qxf4 Nxf4 29.Ng6 Nxg6 30.Rxg7+ Kf8 31.Nc5 f2 32.Ne6+ Ke8 33.Nc7+ Kf8 34.Ne6+ 2.61/23 23.Qg4 Nf2 24.Qc4 Qb5 25.Rxf3 Bxf3 26.Nxf8+ Qxc4 27.bxc4 Rxf8 28.Rd6 Kh7 29.Rxa6 Ne4 30.Ra5 Re8 31.Rb5 Nxg5 32.Rxg5 Re1+ 33.Kb2 Ne5 34.Nb6 Kh6 35.Rg3 Re3 36.Rh3+ Kg6 37.Rg3+ Kf7 38.Rg1 Be4 39.c5 Kf6 40.Rd1 Rc3 41.Nd5+ Bxd5 42.Rxd5 3.00/23 23.Qxf3 Qe5 24.Rd4 Qxe6 25.Be3 Nc3 26.Qh5 Ne2+ 27.Qxe2 Qxh3 28.Bd2 a5 29.Nb6 Ra6 30.Qb5 Qh1+ 31.Kb2 Qc6 32.Qc4+ Qxc4 33.Nxc4 Be7 34.c3 Bf6 35.Rd7 Bc6 36.Rd6 Ne5 37.Rxf6 Nxc4+ 38.bxc4 gxf6 39.cxb4 axb4 40.Bxb4 f5 41.Bc3 Kf7 3.62/23 23.Qe3 Qe5 24.Kb1 Qxe6 25.Rxf3 Bc6 26.Rxf8+ Rxf8 27.Nc5 Qf5 28.Nxe4 Bxe4 29.Qd2 Rc8 30.Rc1 Ne5 31.Be3 Nf3 32.Qd6 Bxc2+ 33.Ka1 Qe5+ 34.Qxe5 Nxe5 35.Bd4 Nd3 36.Rg1 Rc7 37.Bb6 Rf7 38.Be3 Nf4 39.Rc1 Nd5 40.Bd4 3.68/23 23.Qh2 Qf5 24.Be3 Qxe6 25.Rxf3 Nc3 26.Qh3 Qxh3 27.Rxh3 Nxd1 28.Kxd1 Bd6 29.Bc5 Nf4 30.Rg3 Bc7 31.Bxb4 Nd5 32.Be1 Bxg3 33.Bxg3 Ne3+ 34.Kc1 Rc8 35.c3 Nf5 36.Bf4 Kf7 37.Kb2 Bf3 38.c4 Re8 39.Nc5 Re2+ 40.Ka3 

Sep2715
  patzer2: <RandomVisitor: 16.Nxe6! might win for white.> It's certainly an improvement. Deep Fritz 14 x 64 gives 16. Nxe6! fxe6 17. Bf5! exf5 (not 17...Rxh1?? 18. Bg6+ Ke7 19. Qd6#) 18. Rxh8 Ng4 19. Nb6 Nde5 20. Rxf8+ Kxf8 21. Qxb4+ Ke8 22. Rxd8+ Rxd8 23. Qc5 Bc6 24. Bf2 (+2.61 @ 25 depth). 

Sep2715
  patzer2: I didn't spend much time on today's Sunday puzzle, and immediately went for the 23. Qxf3 "show we what you got hot shot" approach. Fritz and <Rv>'s Komodo analysis demonstrate that after 23. Qxf3, Black wins immediately with 23...Qe5! (diagram below). click for larger viewBlack's dual threats 24...Qa1# and 24...Qxe6 are decisive. If 23. Qxf3 Qe5! 24. Nd4 (diagram below),
click for larger view24...Nxg5 25. Qxe5 Nxe5 (5.92 @ 23 depth, Deep Fritz 14) wins easily. So White has nothing better than the forced draw with 23. Rh8+! that I missed, but which my computer found in the blink of an eye. To make matters worse, even after knowing 23. Rh8+! forced the draw, I missed that Black is lost by deviating early from the computer
's precise followup.
Specifically, Black loses after 23...Kxh8? 24. Qf7 or 25...gxf6? 26. Qg4+ . 

Sep2715   gofer: I have looked at this for a while and really wanted to play Bd8, but can't due to
Nxf4, so then I looked at lots of other choices and then thought what about the
queen infiltrating into f7 if we play Rh8+ Kxh8 Qf7, with the threat of Rh1+! Finally, the light dawned, combine both threats...
<23 Rh8+ ...>
But is that enough?!?!?
~~~
Nope, I look at the possible combinations after either Kxh8 and Nxh8 and they
are very complicated. Way too much for my simple brain. Even though I got the
first move I can take no credit for this, nothing is clear at all... 

Sep2715   sfm: The final perpetual is well worth knowing. It has occurred a handful of times in my own games. 

Sep2715
  patzer2: Correction: After 23. Qxf3 Qe5! 24. Nd4 (last diagram in my post above), Black's best is 24...Qxg5+ (not 24...Nxg5? 25. Qxb7 ) 25. Qe3 Bb7 (3.49 @ 23 depth, Deep Fritz 14). 

Sep2715   awfulhangover: Ok, I had to give up, and went to see the solution.
I still understand nothing. The finish is almost inhuman. 

Sep2715
  Penguincw: Answer to my own question: the last there was a Computer vs. Computer POTD was Aug/15/2014: FIBChess vs Falcon, 2004. However, I think computer vs. computer POTDs should only be on Mondays. It's not fair since they can calculate deeper than humans. :p 

Sep2715   houtenton: <wooden nickel> I agree on 24.Bd8! Good move, I also looked at it. No refutation came up?
<Penguincw> A computer never let it come to a mondaymove (food for thought, if this is good english). 

Sep2715
  whiteshark: I would have gone down fighting 

Sep2715
  whiteshark: <patzer2: <RandomVisitor: 16.Nxe6! might win for white.> It's certainly an improvement. Deep Fritz 14 x 64 gives 16. Nxe6! fxe6 17. Bf5! exf5> Hiarcs continues with the cruel <17...Rh6> click for larger view eg 18.Rhe1 Be7 19.Bxe6
tbc... 

Sep2715
  patzer2: <whiteshark> Thanks for the interesting look at 16. Nxe6! fxe6 17. Bf5! Rh6! Deep Fritz 14 with a movebymove look gives 18. Rhe1 Be7 19. Bxe6 Nf8 20. Bc4 Qxd2+ 21.Bxd2 Rh5 22. gxf6 gxf6 23. Bxb4 Re5 24. Bxe7 Kxe7 25. Nc5 Bc6 26. Nxa6 Bb7 27. Nc5 Bc6 28. Kd2 Rb8 29. b3 Nd7 30. Nd3 Rxe1 31. Rxe1+ Kd6 32. Re6+ Kc7 33. a4
Rh8 34. Ke3 Rh2 35. Nb4 Ne5 36. Nxc6 Nxc6 37. Bd3 Ne5 38. a5 Kd7 39. Rb6 Kc7
40. Rb4 Nxd3 41. cxd3 Ra2 42. Rc4+ Kd6 43. b4 f2 44. Rf4 Rb2 45. Rxf2 Rxb4 46.
Rxf6+ Kd5 47. a6 (+3.01 @ 31 depth). 

Sep2815   Al2009: <houtenton + wooden nickel> By the way...
After 24. Bd8! Ng6
White can simply play 25. Bxa5 Nxf4 26. Nxf4 and then, after gaining pawn b4 (which is doomed) White's position seems surely much better (full control of d file, 3 abc connected pawns). 

Sep2815   houtenton: <Al2009> I agree! It feels good, 3 men found a possible winning line. Are we better than a computer? There is still a lot of work to do, there is f.i. the black pawn on f3 and black has the bisshoppair too. 

Sep2815
  HeMateMe: the Falcon and the Shreddman. 

Sep2915
  patzer2: <Al2009> After 24. Bd8 Ng6 25. Bxa5 Nxf4 26. Nxf4 (diagram below) click for larger viewBlack has 26...g5 (5.13 @ 21 depth, Deep Fritz 14) when play might continue 27. Nh5 Kf7 28. Bc7 Rc8 29. Bh2 g4 30. Ng3 Bh6+ 31. Kb2 Bg7+ 32. Kb1 Nxg3 33. Bxg3 Be4 34. Rc1 Bh6 35. Nb6 Rc6 (8.45 @ 22 depth, Deep Fritz 14). So it would seem 24.Bd8 doesn't improve White's chances. Best, if you can see through the complications, is the forced draw with 24. Rd7 =. 


< Earlier Kibitzing · PAGE 2 OF 2 ·
Later Kibitzing> 