Sep0615
  Penguincw: Damn it, it's Sunday!
Anyway, the first few moves were pretty obvious, but I got the move order wrong (not the first time this has happened). I wanted to remove the defender of e6, and then sac on it, but apparently you sac on e6, <and then> remove the c5 knight. Close enough! 

Sep0615   nalinw: I got the first few moves too  and in the right order. However I couldn't see what to do after
22. ... Qxc5 except a perpetual check 

Sep0615
  An Englishman: Good Evening: 22...Qxc5 baffled me as well. How does White win after that? 

Sep0615
  al wazir: Same here.
However, I made the first two moves in the wrong order. 21. Rxc5 Rxc5 22. Nxe6 fxe8 23. Qxe6+ just transposes into the game line, but after 21...Qxc5 22. Nxe6 Qe7 23. Ng7+ Kd8 24. Nf5 Qd7, white has one ♙ for the exchange with no attack. 

Sep0615
  triangulation: It took me half an hour but I think I got this one. After 22 ..Qxc5 23 Qxe6+ Kf8 24 Qxf6+ Ke8 25 Qe6+ Kf8 26. Rd1 white threatens 27 Rd5/7 against which there is no good defense If the rook moves off the back rank to defend with 26 ..Rc7/6 then simply 27. Rd8+ 10 For 26 ..Qc7 27 Rd7 Qxd7 28 Qxd7 and white should win fairly easily after winning one or two more pawns. For 26 ..Qc4 27 Rd5 Qc6 28 Rf5+ Kg7 29 Qf7+ 10
For 26 ..Qc6 27. Rd6 and we transpose to one of the above scenarios. After 22 ..Rxf5 I won't write down my analysis as the game continuation makes it fairly clear on how to win. Let me know if this analysis is correct. I don't have any engine to check. 

Sep0615   morfishine: 21.Rxc5 Rxc5 22.Nxe6 fxe6 23.Qxe6+ Kf8 24.Qd6+ Kg7 25.f4 ***** 

Sep0615   leow: 33... b5 seems to drop a pawn unneccesary? 

Sep0615   The Kings Domain: Good game, and nice combination by white. Shouldn't black have played 25)... Qe5 to save his rook? Perhaps he bungled due to time pressure. 

Sep0615   morfishine: <The Kings Domain: Good game, and nice combination by white. Shouldn't black have played 25)... Qe5 to save his rook?> If 25...Qe5, 26.Rd7+ Kh8
27.Qf7 Qxe4 28.Qxf6+ or 26...Kh6 27.Qxg8
***** 

Sep0615   gofer: I was torn between f4, Nxe6 and Rd1. I felt that Rd1 still
left black with a lot to answer due to the dire state that
Nc5 is in, but I also "saw"...
<21 Nxe6 ...>
21 ... Nxe6
22 Rxc8+
<21 ... fxe6>
<22 Rxc5 ...>
... preparing for Qxe6+, and then being able to play
23 Rd1 ...
... but like others I had thought that
22 ... Qxc5 was stronger than
22 ... Rxc5
So the only question is whether the black queen has to defend Pf6 or not.
If it does then the advantege was "clear to see" as h4 does win back the
rook sacrifice AND tear black's defences to pieces.
22 ... Qxc5
23 Qxe6+ Kf8
24 Qxf6+ Ke8
25 Qe6+ Kf8
26 Rd1
click for larger view~~~
<The Kings Domain>
25 ... Qe5
26 Rd7+ Kh8
27 Qf7 Rxg3+
28 Kf1
click for larger view 

Sep0615   The Kings Domain: morfishine and gofer: Thanks. 

Sep0615   TheFocus: This guy has an acs to grind. 

Sep0615   wooden nickel: It's not all that hard guessing the first few moves, the insanity is in seeing that a temporary rook sac wins back the rook after picking up a couple of pawns.
This move is hard to calculate so many moves ahead,
White to move and win
click for larger view25.h4!
...25.f4? (would fail because the queen needs f4 to give check) Qg4 26.Rd7+ Kh6 27.Qxf6+ Qg6 

Sep0615   wooden nickel: <TheFocus: This guy has an acs to grind.>
I'll say! 

Sep0615   patzer2: Here's a look with Deep Fritz 14:
<20...Rc8?> This is the decisive mistake. Instead, Black can hold with 20... b6! when play might continue 21. f4 Qh6 22. Nb3 Nxb3 23. axb3 h4 24. Kh1 Rd8 = (0.00 @ 26 depth). <21. Nxe6!!> I correctly guessed this move for my Sunday puzzle solution. <21...fxe6 22. Rxc5 Rxc5 23. Qxe6+ Kf8> If 23... Kd8, then 24. Rd1+ Kc7 25. Rd7+ Kb8 26. Qb6 when play might continue 26...Rb5 27. Rxb7+ Kc8 28. Qc7#. <24. Rd1 Kg7 25. h4> I got this far in my solution calculation, but did not anticipate <25...Qg6> I was hoping for 25... Qe5? 26. Rd7+ Kh8 27. Qf7 when play might continue 27...Rxg3+ 28. Kf1! (Not 28. fxg3? Qxg3+ 29. Kf1! Qf3+ 30. Kg1 Qg3+ 31. Kf1 Qf3+ 32. Kg1 Qg3+ = with a draw by repetition). <26. Qe7+ Kh6 27. Qxc5 Qxe4 28. Rd6!> This is the only clear winning move here. Did we have to see this to solve the Sunday puzzle? I wonder if Meier calculated this far ahead before playing 21. Nxe6!! <28...Qf3 29. Qe5 Kg6 30. Qe6 Rg7 31. c4> to (+2.21 @ 25 depth). This move (31. c4) is apparently winning, but much stronger is the Fritz suggestion 31. Rd5! (+4.05 @ 25 depth) threatening 32. Rg5+! . <31...Qf5 32. Qe8+> Fritz indicates this is the strongest continuation, and Meier shows outstanding technique in bringing home the full point after <32...Kh6 33. Kg2 b5 34. Rd5 Qg6 35. Qe3+ Kh7 36. cxb5 axb5 37. Rxb5 Rd7 38. Qf3 Kh6 39. a4 10> 

Sep0615
  Jimfromprovidence: Good comments.
Learned why both 22...Qxc5 and 25...Qe5 do not work.
FWIW, 31 Rd5!? looks like a good alternative to the text. click for larger view 

Sep0615   dfcx: material is even. black's knight and rook are pinned in the C file. white can play 21.Nxe6.
Black definitely won't play 21...Nxe6 22.Rxc8+, that left the other choice 21...fxe6 22.Rxc5 22...Rxc5 23.Qxe6+ Kf8 (Kd8?? 24.Rd1+ Kc7 25.Rc7+ Kb8 26.Qb8 wins).
24.Rd1 Kg7 25.Rd7+ Kh8 and it's getting complicated. I don't see how white can win or even draw from here. 

Sep0615   dfcx: Ouch, missed 25.h4!, I was close. 

Sep0615   BOSTER: < triangulation: After 22...Qxc5 23.Qxe6+>.
Black'd play 23..Qe7 and offer any rook, and he'd play ending without two pawns. 

Sep0615
  agb2002: The material is identical.
The rook on c4 xrays the rook on c8 which is defenseless. The white queen and knight converge on e6. These details lead to consider 21.Nxe6: A) 21... Nxe6 22.Rxc8+ wins the exchange and a pawn with a won position. B) 21... fxe6 22.Rxc5
B.1) 22... Qcx5 23.Qxe6+ Kf8 (else 24.Qxg8+ or 25.Qxc8+ winning two pawns at least) 24.Qxf6+ Ke8 25.Qe6+ Kf8 26.Rd1 (threatens 27.Rd5 and 27.Rd7) B.1.a) 26... Qc6 27.Rd6 Qe8 28.Qf5+ delivers mate or wins the queen for the rook +  [Q+3P vs 2R]. B.1.b) 26... Qxc3 27.Qf5+ Ke8 (27... Ke7 28.Rd7+ Ke8 29.Qf7#; 27... Kg7 28.Rd7+ Kh6(8) 29.Q(R)h7#) 28.Rd7 Rg7 29.Rxg7 Qxg7 30.Qxc8+ Ke(f)7 31.Qxb7+ and 32.Qxg7 +  [3P]. B.2) 22... Rxc5 23.Qxe6+
B.2.a) 23... Kd8 24.Rd1+ Kc7 25.Rd7+ Kb8 (25... Kc8 26.Rg7+ Kb8 27.Qxg8+ Rc8 28.Rxg5 Rxg8 29.Rxg8+ wins) 26.Qb6 and mate soon. B.2.b) 23... Kf8 24.Qd6+ (24.Rd1 Kg7 25.Rd7+ Kh7 26.Qe7 Rg7 looks losing for White) 24... Kg7 25.f4 Qg6 (25... Rc6 26.Qd7+ wins) 26.Qxc5 or 26.f5 recovers the rook and wins one pawn at least. C) 21... Qe5 22.Ng7+ wins a pawn at least (22... Rxg7 23.Qxc8+ Ke7 24.Qxc5+ wins). 

Sep0615
  whiteshark: <25.h4!> is indeed the insane part as it disconnects ♖c5♕g5 link, when 25...Qe5 will face 26.Rd7+ followed by 27.Qf7. 

Sep0615
  Once: Sunday delivers its usual dose of sobering humility. I got the first couple of moves but nowhere near the solution. Excellent puzzle, way above my paygrade. It is Monday tomorrow, isn't it? 

Sep0615
  An Englishman: Good Afternoon: Thanks for solving the 22...Qxc5 puzzle. 26.Rd1 is such a hard move to find because it's so very quiet and unassuming. 

Sep0615   Mendrys: I did get the first few moves up to 24. Rd1 but don't give myself full credit as I didn't calculate what may happen after 22. Qxc5 nor did I see up to 24. h4!. That being said this is the type of combination that I can see myself playing during a game. With a Queen staring the opposing King in the face and the relative incoordination of the black pieces a draw should be had at a minimum. A difficult puzzle but certainly not one where I would say "Wow, there's no way I would have even considered the text move." 
