Jul1521   Brenin: I tried 22 Bxf7 Kxf7 (else Qxg6+ or Bxe8) 23 Qe6+ Kg7 (Kf8 24 Bh6 mate) and now 24 Nh4 looks lethal, e.g. 24 ... Qd6 25 Bh6+ Kxh6 26 Nf5+ Kg5 27 h4+ leads to mate. However, 25 ... Kh7 refutes this line. The game continuation 23 Bh6 is less obvious, but better: Black has plenty of responses, but none save him. 

Jul1521   mel gibson: I wasn't sure about this one.
Stockfish 13 agrees with the text:
22. Bxf7
(22. Bxf7 (♗b3xf7 ♘c6e5 ♕e4xe5+ ♕c7xe5 ♘f3xe5
♗e7xg5 ♗f7xe8 ♖c8xe8 ♘e5f3 ♖e8xe1+ ♖a1xe1 ♗g5f6 b2b4 ♗b7d5 ♖e1d1
♗d5e6 h2h4 ♗f6b2 a3a4 ♔g7f6 ♘f3d4 ♗e6g4 ♖d1d3 a7a6 a4a5 b6xa5
b4xa5 ♔f6g7 ♘d4f3 ♗g4e6 ♖d3d6 ♗e6c4 ♖d6c6 ♗c4d3 ♘f3g5 ♗b2f6
♘g5e6+ ♔g7f7 ♘e6c7 ♗f6d8 ♖c6d6 ♔f7e7 ♖d6xd3 ♗d8xc7 ♖d3a3 ♔e7f6
♖a3a2 ♔f6e6 ♔g1f1 ♗c7d8 f2f3 ♗d8xa5 ♖a2xa5 ♔e6e7 ♔f1f2 ♔e7f8
♔f2e3) +6.73/47 459)
score for White +6.73 depth 47. 

Jul1521   drollere: this one took an embarrassing amount of brain damage, but i found 22. Bxf7 Kxf7 (otherwise Qxg6+), 23. Bh6, and i can't find a defense against the simultaneous mate threats Qc4+, Qd5+ and Qe6+, e.g. 23. .. Ne5, 24. Nxe5+ and 25. Qxg6+ 

Jul1521   Walter Glattke: 22.Qc4 Bxg5 23.Nxg5 Rxe1+ 24.Rxe1 Rf8 25. Qc3+
B) 22.Bxf7 Kxf7!? 23.Nh4 Bxg5 24.Qxg6+ Kf8 25.Qxg5 + / 23.Bh6 Nd8 24.Ng5+ Kf6 25.Nh6+ Kf7 26.Ng5+ = 25.Qd4+ Kf5 no refutation of 22.Kxf7 found, ask stockfish for that 

Jul1521
  al wazir: If 23...Bf6, then 24. Qd5+.
Not 24. Ng5+ because of 24...Bxg5 25. Qd5+ Kf6 26. Qxg5+ (26. Bxg5+ Kg7) Kf7. 

Jul1521
  al wazir: But what happens after 23...Nd8 ?
If 24. Ng5+, then 24...Kf6 25. Qd4+ (25. Nh7+ Kf7) Kf5 26. f3/h3 (threatening g4#)
Qf4. 

Jul1521   drollere: <22.Bxf7 Kxf7!? 23.Bh6 Nd8>(?)<24.Ng5+>(?)<Kf6 25.Nh6+ Kf7 ... no refutation of 22.Kxf7 found> 24. Ne5+ is better at that position, if i read your note correctly, since capturing the N loses the Q (the exchange with attack) and the K has only 3 retreat squares, f6, e6 or g8. 25. Qxg6+ is then either mate (f6), or forces the K to d5 (from e6) where the R's and QB will either mate or take the Q, and g8 is mate on h8. if 24. .. Ke6 25. Qxg6+ Bf6, 26. Ng4+ and there are many ways to win from there. 

Jul1521   Walter Glattke: Yes, 24.Ne5 Ne6 25.Ng4++# not Ng5 here 

Jul1521   saturn2: I had 22 Bxf7 Kxf7 and the clearance of g5 for the knight however by 23. Bf4 (did not see Bh6) Bd6 (Qd8 24.Rd1) 25.Qd5+ and white wins back the piece has a pawn more and attack. 

Jul1521   Walter Glattke: Corr. I mean 24.Ne5+ Kf6 25. Qd4+ and then 25.Nd6 26. Ng4++# so 22.Bxf7 Kxf7 23.Bh6! wins. 

Jul1521   GlennOliver: Even given that the initial move 22. Bxf7 could readily be intuited, the continuing sequence to forced mate was very far from apparent. If Martinez actually saw that full sequence, prior to making move 22, then chapeau. 

Jul1521   Autoreparaturwerkbau: I went for the right line up to 24th move, but i never even considered 25.♘h7+! ... what a stunner! So, only half a point for me today. 

Jul1521
  chrisowen: Loquacious i vents lavish Bxf7 hobble achtung munch ivory aisle its i vains lavish loquacious i vents lavish yard its bray draw its velocity veins loins lavish quick knights jekyll germinates puffy again i devotion gapoff its joy arrive bide its mow i v rants vents clear infit totups awooga jobus i vents cinch jiffy afford icicle flick feline ae its dont arm chew its fiddly Bxf7 cervix; 

Jul1521
  chrisowen: Calm it re down bf7 us x 

Jul1521
  chrisowen: Firing bunny pops no? 

Jul1521   Walter Glattke: Not easy, 26.Ng4++ Kf7 27.Rxe6 

Jul1521   agb2002: Black threatens Bxg5.
White has five pieces ready to attack the black king, starting with 22.Bxf7: A) 22... Kxf7 23.Bh6 (threatens 24.Qe6#)
A.1) 23... Bf6 24.Qd5+ Re7 25.Qxe6#.
A.2) 23... Bf8 24.Ng5+ Kg8 (24... Kf6 25.Qf3+ Qf4 25.Qxf4#) 25.Qxg6+ Kh8 (25... Bg7 26.Qh7+ Kf8 27.Qh8#) 26.Rxe8 Rxe8 27.Qxe8 wins. A.3) 23... Kg8 24.Qxg6+ Kh8 25.Qg7#.
A.4) 23... Bh4 24.Qc4+ Kf6 25.Qxh4+ Kf7 (25... Kf5 26.Qg5#; 25... g5 26.Qxg5+ Kf7 27.Qg7#) 26.Qc4+ Kf6 27.Qc3+ Ne5 (27... Kf7 28.Qg7#; 27... Kf5 28.Nh4+ Kg4 29.Qh3#) 28.Nxe5 A.4.a) 28... Qxc3 29.bxc3 (29.Nd7+ leaves the knight trapped) 29... Rxe5 30.Bg7+ Kxg7 31.Rxe5 +  [R+P vs b]. A.4.b) 28... Rxe5 29.Bg7+ followed by 30.Qxe5 in A.4.a. A.5) 23... Rh8 24.Qe6+ Ke8 25.Qxg6+
A.5.a) 25... Kd7 26.Rad1+ Nd4 (26... Bd6 27.Qe6+ Kd8 28.Rxd6+ wins) 27.Rxd4+ Bd6 28.Qe6+ Kc6 29.Rc4+ wins. A.5.b) 25... Kd8 26.Rad1+ as above.
A.6) 23... Qd6 24.Qc4+ Kf6 25.Ng5 (25.Rad1 Nd4, unclear) 25... Nd8 (25... Ne5 26.Qf4#; 25... Rf8 26.Ne4+ wins) 26.Qh4 Qb8 (prevents Ne4+ while keeping f4 under control; 26... Qc7 27.Ne6+ wins) 27.Nh7+ Kf7 (27... Kf5 28.Qh3#) 28.Rxe7+ Rxe7 29.Qf6+ Ke8 (29... Kg8 30.Qxe7 Qc7 31.Nf6+ Kh8 32.Bg7#) 30.Qf8+ Kd7 31.Nf6+ looks winning (31... Ke8 32.Re1+). A.7) 23... Qd7 24.Qc4+ Kf6 25.Ng5 looks winning (25... Nd8 26.Nh7+ Kf5 27.Qf4#; 25...Ne5 26.Nh7+ Kf5 27.Qf4+ Ke6 27.Qxe5+ Kf7 28.Qg7#). B) 22... Bxg5 23.Qxg6+ (or 23.Bxe8) 23... Kh8 (23... Kf8 24.Qg8#) 24.Nxg5 wins. C) 22... Ne5 23.Qxe5+
C.1) 23... Qxe5 24.Nxe5 Bxg5 25.Bxe8 wins decisive material. C.2) 23... Kxf7 24.Qe6+ Kg7 (24... Kf8 25.Bh6#) 25.Bxe7 wins decisive material. 

Jul1521   King.Arthur.Brazil: I found the sacrifice and the next move: 22. ♗xf7 ♔xf7 23. ♗h6. I saw that 23...♗f6 24. ♕d5+ ♖e6 25. ♕xe6# was impossible, but I didn't try Black answer 23...♗d6?. I though about the answer 23...♘d8 24.♘e5+ which forced Black to lose the ♕: ♕xe5 25.♕xe5 since 24...♔g8 25.♕xg6+ and checkmate next. Or 24... ♔e6 25. ♕g6+ ♔d5 26. ♕d3+ ♔e6 27. ♘d7+ ♔f7 28. ♕h7#. However, the keymove 25.♘H7+ I probably would not find. Certainly, my namesake was very upset about today game. 

Jul1521   King.Arthur.Brazil: Oops, I forget to say that after 25...♕xh7 26. ♕f3+ ♗f4 27. ♕xf4#. So, Black ♕ is slaved, cannot capture... <polyglOtwen> I guess this bunny has tied legs... 

Jul1521   johnnydeep: "Got" the first two moves, or perhaps "guessed" is the better word. White's position seemed good after that with several subsequent mating attacks. Not bad for me "for a Thursday" ;) 

Jul1521
  Jimfromprovidence: I missed the follow up after 23...Qd6 24 Qc4+ Kf6, namely 25 Ng5. click for larger viewThat is because I did not see the threat 26 Qf7#. Subsequently I saw that 25...Rf8 is no good in this position because of 26 Nh7+ Kf5 27 Qe4#. 

Jul1521
  NMABQjmgChess: Am really glad that the engine is back on line. Need it to trace through all the variations. 

Feb2323
  Jabot: This game is very similar to Petrosian vs Balashov, 1974 . Have to wonder if silvino knew that game. 
