8

Shown below is a game I have been trying to solve, Black has been better for the whole game but resigns at the end. On checking the analysis stockfish says black is still better

Stockfish Analysis on Lichess

[Event "Hourly Blitz Arena"]
[Site "https://lichess.org/RS8OKA3q"]
[Date "2021.06.09"]
[White "okotieno"]
[Black "Kryptonopolis"]
[Result "1-0"]
[UTCDate "2021.06.09"]
[UTCTime "15:28:08"]
[WhiteElo "2018"]
[BlackElo "2363"]
[WhiteRatingDiff "+11"]
[BlackRatingDiff "-10"]
[Variant "Standard"]
[TimeControl "300+0"]
[ECO "B21"]
[Opening "Sicilian Defense: McDonnell Attack"]
[Termination "Normal"]
[Annotator "lichess.org"]
[FEN ""]

    

  1. e4 c5 2. f4 { B21 Sicilian Defense: McDonnell Attack } d5 3. e5?! { (-0.07 → -0.59) Inaccuracy. exd5 was best. } (3. exd5 Nf6 4. Nc3 Nxd5 5. Nxd5 Qxd5 6. Qf3 Qe6+ 7. Qe3 Nc6) 3... Nc6 4. Bb5 Bd7 5. Bxc6 Bxc6 6. d3?! { (-0.62 → -1.17) Inaccuracy. d4 was best. } (6. d4 Nh6 7. Nf3 e6 8. c3 Nf5 9. O-O cxd4 10. cxd4 Bb5) 6... d4 7. Nf3 Nh6 8. Nbd2 e6 9. Ne4 Be7 10. h3?! { (-0.63 → -1.44) Inaccuracy. Qe2 was best. } (10. Qe2) 10... O-O? { (-1.44 → -0.04) Mistake. f5 was best. } (10... f5) 11. g4 Qd5 12. Kf2 c4?! { (-0.76 → 0.01) Inaccuracy. f6 was best. } (12... f6 13. exf6) 13. g5?? { (0.01 → -2.34) Blunder. Re1 was best. } (13. Re1 Rac8) 13... Nf5 14. Rg1 Kh8 15. Qf1 Rac8 16. Bd2?! { (-2.69 → -3.52) Inaccuracy. Qe2 was best. } (16. Qe2 Rfd8 17. h4 cxd3 18. cxd3 Qb5 19. Rd1 Bxe4 20. Qxe4 Rc2+ 21. Rd2 Rdc8 22. h5 g6) 16... cxd3?! { (-3.52 → -2.75) Inaccuracy. b5 was best. } (16... b5 17. Qe2 c3 18. bxc3 dxc3 19. Be3 b4 20. h4 Rfd8 21. h5 a5 22. h6 g6 23. Nf6) 17. cxd3 Bb5 18. Rc1?! { (-2.30 → -3.39) Inaccuracy. a4 was best. } (18. a4 Ba6 19. b4 b6 20. Qe2 Rc2 21. Rgb1 Rfc8 22. Ne1 R2c7 23. a5 Bb7 24. b5 g6) 18... Qxa2 19. Rxc8?! { (-3.01 → -3.78) Inaccuracy. Ng3 was best. } (19. Ng3) 19... Rxc8 20. Bc1? { (-3.78 → -7.51) Mistake. Ng3 was best. } (20. Ng3 Nxg3) 20... Rc2+ 21. Nfd2 Qa6 22. Nf6 Bxd3 23. Qg2 Qc6 24. Qg4 Rxc1 25. Qh5 Bxf6 26. Qxf7 { Black resigns. } 1-0
    2.

How should black continue to prevent the checkmate on the backrank?

Amit Dash
  • 389
  • 2
  • 14
Owen Kelvin
  • 807
  • 1
  • 6
  • 17

1 Answers1

16

White threatens 27. Qf8#. Black has several ways to stop the immediate mate, so let's look at them:

  1. 26...Qc8? 27. Rxc1 Qg8 (27...Qxc1? 28. Qf8#) 28. Qxg8+ Kxg8 29. gxf6 gives back a decisive amount of material.
  2. 26...Qc5? 27. Qe8+ forces mate.
  3. 26...Be7 27. Rxc1 Qd7 leaves Black with a still-comfortable material advantage.
  4. 26...Rf1+ is a variant of the above idea. After 27. Nxf1 or Rxf1 Black plays 27...Be7 and maintains a still-comfortable material advantage.
  5. 26...h6? 27. Qf8+ Kh7 28. g6#
  6. 26...h5? 27. Qf8+ Kh7 28. g6+ Kh6 29. Qh8#
  7. 26...Nh6 27. gxh6 Bh4+ is the computer's main line. 28. Rg3 (only legal move) Bxg3+ 29. Kxg3 Rg1+ 30. Kh4 Qc8 leaves Black up more than a decisive amount of material.

The last line is admittedly somewhat difficult to spot, but 26...Be7 is still winning.

Allure
  • 26,534
  • 1
  • 68
  • 146
  • Bh4+ is what I was missing after white picks the knight, thanks – Owen Kelvin Jun 10 '21 at 06:22
  • Thoughts on 26. … h5 (instead of h6)? It doesnt look like it falls to the g6 mate im (5), but I could see 27. Qxh5+ being annoying/starting to give up those material advantages – D. Ben Knoble Jun 10 '21 at 17:45
  • 1
    @D.BenKnoble 26. …h5 would be a blunder because of the same variation provided by Allure 26...h6? 27. Qf8+ Kh7 28. g6+ Kh6 29.Qh8# – Owen Kelvin Jun 10 '21 at 19:31
  • 26...Nh6 27. Qf8+ Ng8 (the declined sacrifice becomes the only blocker that self-pins to stop mate) is mind-bendingly precise but good to study and get comfortable with.
    • – BaseZen Jun 10 '21 at 19:59
  • on the first line, with Qg8 (forced) I don't see what's wrong with, e.g. Qb8? Why it is forced? I imagine Qb8 leads to gxf3 gxf3. Qxf3+ Ng7, which is the same result in material losses for black. – njzk2 Jun 10 '21 at 21:17
  • 1
    @njzk2 I meant Black has to move the queen and cannot play Qxc1. I'll edit. – Allure Jun 10 '21 at 21:26
  • @njzk2, 26...Nh6 27. Qf8+ Ng8 28. gxf3 continues with 28... Rxg1 and White can no longer checkmate. – Stadem Jun 10 '21 at 21:31
  • @Stadem ah, yes, didn't see Rxg1, thanks – njzk2 Jun 10 '21 at 21:37