patzer2: Here's a look at today's Saturday puzzle (25. ?) and game with the chessgames.com opening explorer (OE), Stockfish 8, Deep Fritz 15 and Houdini 5.01:<1. Nf3 Nf6 2. d4 g6 3. Bg5 Bg7 4. c3> This quiet opening move has decent winning results (36.5% wins vs 25.8% losses out of 422 games in the OE) and is the second most popular.
Most often played is 4. Nbd2 = (0.00 @ 36 depth, Houdini 5.01 x 64) as in White's win in B Sambuev vs Fishbein, 2017 or Black's win in E Torre vs G Jones, 2017.
<4...O-O 5. Nbd2 d5 6. e3 c6> The computers evaluate 6...c6 as level, but this seldom played move has not been good for Black. Of 21 games with 6...c6 in the OE, White won 10 and Black won one.
The popular move and my preference is 6...Nbd7 as in the draw in Kramnik vs Radjabov, 2017.
<7. Be2> This is good, but also playable is 7. Bd3 as in White's win in Balashov vs Gufeld, 1975 or Black's win in A Saric vs T Grzincic, 2001.
<7...Bg4> This is OK, but slightly better IMO is 7...Bf5 as in Black's wins in O Yaksin vs D Baramidze, 2007 and Kamsky vs M Vachier-Lagrave, 2013.
<8. h3 Bxf3 9. Nxf3> We appear to leave opening theory, as this is the only game with this move in the OE.
Equally strong is 9. Bxf3 as in White's win in Khenkin vs Zaur Bayramov, 2017.
<9. Nbd7 10. O-O Re8 11. Bf4 Ne4 12. Bh2 e5 13. g4 Qa5 14. Qc2 Rad8 15. b4 Qc7 16. c4 Qd6 17. cxd5 cxd5 => Up to now, Black has done well holding it level against a much stronger opponent.
<18. Bb5 f6?!> This is not too bad, as Black may be able to survive with only a slight disadvantage after 19. dxe5 Qxb4 20. Rab1 Qf8 21. e6! 
(+0.56 @ 32 depth, Stockfish 8.)
The main point of being critical of 18...f6>?! is there's a much better move available. Instead of 18...f6?, Black should play 18... Rc8 19. Qb2 h5 20. Nxe5 Nxe5 21. f3! Nc4! 22. Bxc4 Ng3!23. Qf2 Rxc4 24. Bxg3 Qxb4 = to 
(-0.33 @ 33 depth, Stockfish 8) with a good game.
<19. Rfd1!> This is a good move, which becomes an excellent move when Black takes the bait and plays 19...Rc8? to take the "open file."
An equally good or better option for White might be the above mentioned computer suggestion 19. dxe5 Qxb4 20. Rab1 Qf8 21. e6! (+0.56 @ 32 depth, Stockfish 8.)
<19... Rc8?> Allowing 20. dxe5 Qxb4 21. Qd3 
to 
(+1.92 @ 30 depth, Stockfish 8), this seems to be the losing move.
Instead, Black should play 19... Qxb4 20. Rab1 Qf8 21. dxe5 fxe5 22. Kg2 Ndc5 23. Bxe8 Qxe8 24. Qb2 b6 25. Bxe5 Na4 26. Qd4 Bxe5 27. Nxe5 Nac3 28. Rbc1 Nxd1 29. Rxd1 Qb5 30. g5 Qc5 31. h4 (+045 @ 32 depth, Stockfish 8) with only a small edge for White.
<20. dxe5 Qxb4 21. Qd3 Ndc5 22. Qxd5+ Re6 23. Bd7 Nc3 24. Bxe6+ Kh8 25. Qd8+!!> This solves the Saturday May 20, 2017 chessgames.com puzzle.
<25...Rxd8 26. Rxd8+ Bf8 27. Rxf8+ Kg7 28. exf6+!> With best play, 28. exf6+! yields mate-in-six.
My winning try was 28. Rf7+ Kh6 29. Bf4+ g5 30. Bf5 Ne2+ 31. Kh2 Qxf4+ 32. exf4 gxf4 33. Rxh7#.
Another strong winning option is 28. Rg8+ Kh6 29. Bf4+ g5 30. Nxg5 Ne2+ (30... fxg5 31. Bxg5#) 31. Kh2 Qxf4+ 32. exf4 Nxe6 33. Nf7#.
<28... Kh6> If 28... Kxf8, it's mate-in-three after 29. Bd6+ Ke8 30. f7+ Kd8 31. f8=Q#
<29. Bf4+> This wins, but even faster is mate-in-five with 29. g5+ Kh5 30. Rh8 Ne2+ 31. Kg2 Nf4+ 32. Bxf4 Nxe6 33. Rxh7#.
<29... Qxf4 30. exf4 Nxe6 31. Rc8 1-0> Black resigns.
It's mate-in-nine after 31...Ne2+ 32. Kh2 N2xf4 33. Re1 b6 34. Re5 a6 35. g5+ Nxg5 36. Nxg5 Kh5 37. Nf7+ Kh4 38. Rc4 g5 39. Rxg5 b5 40. Rxf4# (Deep Fritz 15 @ 23 depth).
Kibitzing
sac: 25.Qd8+
(-3.07 @ 32 depth, Stockfish 8
