White to play and win

[acase]

The goal here is to show a winning score from the root position, and for extra credit to figure out what study and who the author is that I borrowed from.

I suspect that there will be an engine that will achieve the goal and that someone will figure out the study, as usual I'll post a full PGN later with all the details.


[pgn][Event "Modified Study by whom?"]
[Site "Columbus, Ohio"]
[Date "2023.04.05"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "1-0"]
[BlackElo "2000"]
[Time "20:51:23"]
[WhiteElo "2000"]
[TimeControl "0"]
[SetUp "1"]
[FEN "8/p6p/1p2P1Qp/r2RKBpP/pPk1N3/p2b2p1/P2n2R1/1q1r4 w - - 0 1"]
[Termination "normal"]
[PlyCount "0"]
[WhiteType "program"]
[BlackType "human"]

1-0[/pgn]

[peter]

The goal here is to show a winning score from the root position, and for extra credit to figure out what study and who the author is that I borrowed from.

I suspect that there will be an engine that will achieve the goal and that someone will figure out the study, as usual I'll post a full PGN later with all the details.


[pgn][Event "Modified Study by whom?"]
[Site "Columbus, Ohio"]
[Date "2023.04.05"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "1-0"]
[BlackElo "2000"]
[Time "20:51:23"]
[WhiteElo "2000"]
[TimeControl "0"]
[SetUp "1"]
[FEN "8/p6p/1p2P1Qp/r2RKBpP/pPk1N3/p2b2p1/P2n2R1/1q1r4 w - - 0 1"]
[Termination "normal"]
[PlyCount "0"]
[WhiteType "program"]
[BlackType "human"]

1-0[/pgn]

Took a Cfish because of it's fine persistent TT- feature, Sopel brought up way back for his FT2. Found the compile here

https://www.open-chess.org/viewtopic.php?p=26581#p26581

Classical evaluation, Syzygy ProbeLimit=0, after some Forward- Backward:

8/p6p/1p2P1Qp/r2RKBpP/pPk1N3/p2b2p1/P2n2R1/1q1r4 w - - 0 1

Analysis by Cfish 160223 64 POPCNT:

1.Nxd2+ Rxd2 2.Bxd3+ Qxd3 3.Qxd3+ Rxd3 4.Rxa5 bxa5 5.b5 Re3+ 6.Kd6 Rd3+ 7.Kc6 Rc3 8.e7 Kb4+ 9.Kd5 Re3 10.Kd6 Rd3+ 11.Ke6 Re3+ 12.Kd7 Rd3+ 13.Ke8 Re3 14.Rc2 g4 15.Rg2 Kc4 16.Rg1 Kd3 17.Kd7 Ke2 18.Rxg3 Rxg3 19.e8Q+ Re3 20.Qa8 Kf2 21.Qxa7 g3 22.b6 g2 23.b7 g1Q 24.b8Q Qg4+ 25.Kc6 Qxh5 26.Qf4+ Qf3+ 27.Qxf3+ Kxf3 28.Qxh7 Re6+ 29.Kd7 Rf6 30.Qg7 Rf4 31.Qxh6 Rc4 32.Qe6 Rb4 33.Qg6 Kf2 34.Kc6 Rb2 35.Qe4 Kg3 36.Qe1+ Kh2 37.Qe5+ Kh3 38.Qxa5 Rf2 39.Qh5+ Kg2 40.Qg5+ Kf3 41.Kb5
+- (5.35 ++) Depth: 41/62 00:02:01 9934MN

Edit: in meantime I saw HekaStockfish (Majkel Novak) get a winning eval with 5.b5 in output- line, which seems to be crucial, started with empty hash soon too, also with classical (non- NNUE) eval and Find Mate option checked, regards

[Ajedrecista]

Hello Andrew:

I think that I found the original problem thanks to Peter's solution:

[pgn][Event "sp.p MSV-2014"]
[Site ""]
[Date "2014"]
[Round ""]
[White "Becker, Richard"]
[Black ""]
[Result "1-0"]
[SetUp "1"]
[FEN "8/p6p/4P2p/p3K1pP/pPk5/p2r2p1/P5R1/8 w - - 0 1"]
[PlyCount "69"]
[EventDate "2014"]

1. b5 $1 (1. bxa5 $2 {unblocks the a5 square. See variation 11...Kb4.1...} Re3+ 2. Kd6 Rd3+ 3. Kc6 Re3 4. Kd7 Rd3+ 5. Ke8 Re3 6. e7 a6 7. Rg1 Kb4 8. Rd1 Kc3 9. Kd7 Kc2 {=}) 1... Re3+ (1... Kc5 2. e7 Re3+ 3. Kf6 Kd6 (3... Rf3+ 4. Ke6 Re3+ 5. Kd7 Rd3+ 6. Ke8 Rf3 7. Rc2+ Kb4 8. Re2 Rf2 9. Re4+ Kc3 10. Kd7 {+-}) 4. Rd2+ Kc7 5. Rd3 $1 Rxd3 (5... Re2 6. Rxg3 Rf2+ 7. Ke6 Re2+ 8. Kf7 Rf2+ 9. Ke8 Rxa2 10. Rc3+ Kb7 11. Re3 Rf2 12. Kd7 a2 13. Re1 Rd2+ 14. Ke6 a1=Q 15. Rxa1 Re2+ 16. Kd7 Rd2+ 17. Ke8 {+-}) 6. e8=Q Rd6+ 7. Ke5 g2 (7... Rd2 8. Qf7+ Kb6 9. Qf3 Kxb5 10. Qxg3 {+-}) 8. Qf7+ Rd7 9. Qf2 Rd2 10. Qxa7+ Kd8 11. Qg1 {+-}) 2. Kd6 $1 (2. Kf6 $2 Rf3+ 3. Kg7 Re3 4. Kf7 (4. Kxh6 Rxe6+ 5. Kxh7 Rf6 6. h6 Rf2 7. Rxg3 Rxa2 {=}) 4... Rf3+ 5. Ke8 Rf2 6. Rxg3 Rxa2 {=}) 2... Rd3+ 3. Kc6 {(Kc7)} (3. Ke7 $2 Rf3 4. Kd7 Rf2 5. e7 Rxg2 6. e8=Q Rd2+ {=}) 3... Re3 4. Kd7 Rd3+ 5. Ke8 Re3 ( 5... Rf3 6. e7 Rf2 7. Rxg3 Rxa2 (7... Re2 8. Rxa3 {+-}) 8. Rg4+ Kb3 9. Re4 Rc2 (9... Rf2 10. Kd8 Rd2+ 11. Kc7 a2 12. e8=Q Rc2+ 13. Kb8 a1=Q 14. b6 $1 {+-}) 10. b6 $1 axb6 11. Kd8 Rd2+ 12. Kc7 a2 13. e8=Q a1=Q 14. Qf7+ Kc2 15. Rc4+ Qc3 (15... Kb2 16. Qf6+ Kb3 17. Qxa1 Kxc4 18. Qxa4+ {+-}) (15... Kb3 16. Rc1+ Kb2 17. Qf6+ Kxc1 18. Qxa1+ Kc2 19. Qxa4+ {+-}) 16. Qxh7+ Rd3 17. Rxc3+ Kxc3 18. Qxh6 {+-}) 6. e7 Rf3 7. Kd7 $1 {The first maneuver to lose a tempo.} (7. Re2 $2 {(Rc2+?)} Kd3 8. Rg2 Kc4 {Loss of time.}) 7... Rd3+ 8. Ke6 Re3+ 9. Kf7 Rf3+ 10. Ke8 {Repeating the position, but with BTM. 10...} Re3 (10... Kb4 11. Re2 Rf2 12. Re4+ Kc3 13. Kd7 g2 14. Re3+ Kb2 15. Re1 g1=Q 16. Rxg1 Rd2+ 17. Kc6 Re2 18. Rg2 $1 Rxg2 19. e8=Q Kxa2 20. Kb7 Kb1 21. Qf8 Rb2 22. Qf1+ Kc2 (22... Ka2 23. Kxa7 Rb3 24. b6 {+-}) 23. Qc4+ Kd2 24. Kxa7 {-+}) 11. Rg1 g4 (11... Rd3 12. Re1 g2 13. Kf7 g1=Q 14. Rxg1 Rf3+ 15. Ke6 Re3+ 16. Kd7 Rd3+ 17. Kc6 {+-}) (11... Re2 12. Rxg3 Kb4 (12... Rxa2 13. Rg4+ {etc.}) 13. Rd3 Rxa2 (13... Kc4 14. Rxa3 {+-}) 14. Rd4+ Kb3 15. Re4 {etc.}) (11... Kb4 12. Kd7 Rd3+ 13. Ke6 Re3+ 14. Kd6 Rd3+ (14... Re2 15. Rxg3 {+-}) 15. Ke5 Re3+ 16. Kd4 Rxe7 17. Rb1# {mate! The point of 1.b5! instead of 1.bxa5?}) 12. Rg2 {The beginning of the second maneuver.} (12. Rc1+ $2 Kd3 {=}) 12... Rf3 (12... Kb4 13. Rc2 Re1 (13... g2 14. Rxg2 g3 15. Rc2 {+-}) 14. Rd2 Kc4 15. Kd7 Kc3 16. Rg2 Rd1+ 17. Kc6 Re1 18. Rxg3+ Kb2 19. Rg2+ Ka1 20. Rg1 $1 Rxg1 21. e8=Q Rb1 (21... g3 22. Kb7 Rb1 23. Qe5+ {transposes}) 22. Kb7 g3 23. Qe5+ Kxa2 24. Kxa7 Rb3 (24... g2 25. Qd5+ Ka1 26. Qxg2 Rxb5 27. Ka6 {+-}) 25. b6 Kb1 26. Qe4+ Kb2 27. Qd4+ Kb1 28. Qxa4 Kb2 29. Qxb3+ Kxb3 30. b7 {+-}) 13. Kd7 $1 (13. Re2 $2 Kd3 14. Re1 Kd2 15. Re4 Kd1 {=}) 13... Rd3+ 14. Ke6 Re3+ 15. Kf7 Rf3+ 16. Ke8 Re3 (16... Rc3 17. Rd2 Rd3 18. Re2 {+-}) 17. Rg1 {Again, repeating the position with BTM. 17...} g2 (17... Re2 18. Rxg3 Rxa2 19. Rxg4+ {+-}) (17... Kb4 18. Kd7 Rd3+ 19. Ke6 Re3+ 20. Kd6 {+-}) 18. Rxg2 g3 19. Rg1 Rd3 20. Kf7 $1 {The third maneuver, but this time the wK's rundlauf goes clockwise.} (20. Re1 $2 g2 21. Kf7 Rf3+ {=}) (20. Rc1+ $2 Kb4 21. Kf7 (21. Rc2 Re3 22. Rd2 Kc3 23. Rg2 Kc4 {Loss of time.}) 21... Rf3+ 22. Ke6 Re3+ 23. Kd6 Re2 24. Rd1 Kc3 25. Rg1 Kb2 26. Rxg3 Kxa2 {=}) 20... Rf3+ 21. Ke6 Re3+ 22. Kd7 Rd3+ 23. Ke8 {The third position is repeated with BTM. 23. ..} Re3 24. Rc1+ $1 (24. Rg2 $2 Rf3 $1 25. Kd7 (25. Re2 Kd3 26. Re1 Kd2 27. Re4 Kd1 {=}) (25. Rc2+ Kd3 26. Kd7 Kxc2 27. e8=Q Kd2 $1 {=}) 25... Rd3+ 26. Ke6 Re3+ 27. Kf7 Rf3+ 28. Ke8 Re3 {Loss of time.}) 24... Kd3 (24... Kb4 25. Rc2 ( 25. Kd7 $2 Rd3+ 26. Ke6 Re3+ 27. Kd6 Re2 28. Rd1 Kc3 29. Rg1 Kb2 {=}) 25... Re1 26. Rd2 Kc4 27. Kd7 Kc3 28. Rg2 {+-}) 25. Kd7 Kd2 26. Rf1 $1 (26. Rc8 $2 Rd3+ 27. Ke6 Re3+ 28. Kf7 Rf3+ 29. Kg7 g2 30. e8=Q g1=Q+ {=}) 26... Rd3+ 27. Ke6 Re3+ (27... Ke2 28. Rf8 g2 29. Rg8 Re3+ 30. Kd7 {+-}) 28. Kf7 Kc2 (28... Ke2 29. Rf4 g2 30. Rg4 {+-}) 29. e8=Q Rxe8 30. Kxe8 Kb2 31. Rf7 Kxa2 32. Rxa7 Kb2 33. b6 a2 34. b7 a1=Q 35. b8=Q+ {wins} 1-0[/pgn]

Links:

https://arves.org/arves/index.php/en/aw ... vilag-2014 (Where I found the PGN).
https://arves.org/arves/images/PDF-Awar ... g-2014.pdf (Special Prize, page 5 of the PDF).
https://www.arves.org/arves/images/PDF/ ... y_2016.pdf (Page 62 of the issue, Problem No. 20584).

Regards from Spain.

Ajedrecista.

[peter]

Hello Andrew:

I think that I found the original problem thanks to Peter's solution:

[pgn][Event "sp.p MSV-2014"]
[Site ""]
[Date "2014"]
[Round ""]
[White "Becker, Richard"]
[Black ""]
[Result "1-0"]
[SetUp "1"]
[FEN "8/p6p/4P2p/p3K1pP/pPk5/p2r2p1/P5R1/8 w - - 0 1"]
[PlyCount "69"]
[EventDate "2014"]

1. b5 $1 (1. bxa5 $2 {unblocks the a5 square. See variation 11...Kb4.1...} Re3+ 2. Kd6 Rd3+ 3. Kc6 Re3 4. Kd7 Rd3+ 5. Ke8 Re3 6. e7 a6 7. Rg1 Kb4 8. Rd1 Kc3 9. Kd7 Kc2 {=}) 1... Re3+ (1... Kc5 2. e7 Re3+ 3. Kf6 Kd6 (3... Rf3+ 4. Ke6 Re3+ 5. Kd7 Rd3+ 6. Ke8 Rf3 7. Rc2+ Kb4 8. Re2 Rf2 9. Re4+ Kc3 10. Kd7 {+-}) 4. Rd2+ Kc7 5. Rd3 $1 Rxd3 (5... Re2 6. Rxg3 Rf2+ 7. Ke6 Re2+ 8. Kf7 Rf2+ 9. Ke8 Rxa2 10. Rc3+ Kb7 11. Re3 Rf2 12. Kd7 a2 13. Re1 Rd2+ 14. Ke6 a1=Q 15. Rxa1 Re2+ 16. Kd7 Rd2+ 17. Ke8 {+-}) 6. e8=Q Rd6+ 7. Ke5 g2 (7... Rd2 8. Qf7+ Kb6 9. Qf3 Kxb5 10. Qxg3 {+-}) 8. Qf7+ Rd7 9. Qf2 Rd2 10. Qxa7+ Kd8 11. Qg1 {+-}) 2. Kd6 $1 (2. Kf6 $2 Rf3+ 3. Kg7 Re3 4. Kf7 (4. Kxh6 Rxe6+ 5. Kxh7 Rf6 6. h6 Rf2 7. Rxg3 Rxa2 {=}) 4... Rf3+ 5. Ke8 Rf2 6. Rxg3 Rxa2 {=}) 2... Rd3+ 3. Kc6 {(Kc7)} (3. Ke7 $2 Rf3 4. Kd7 Rf2 5. e7 Rxg2 6. e8=Q Rd2+ {=}) 3... Re3 4. Kd7 Rd3+ 5. Ke8 Re3 ( 5... Rf3 6. e7 Rf2 7. Rxg3 Rxa2 (7... Re2 8. Rxa3 {+-}) 8. Rg4+ Kb3 9. Re4 Rc2 (9... Rf2 10. Kd8 Rd2+ 11. Kc7 a2 12. e8=Q Rc2+ 13. Kb8 a1=Q 14. b6 $1 {+-}) 10. b6 $1 axb6 11. Kd8 Rd2+ 12. Kc7 a2 13. e8=Q a1=Q 14. Qf7+ Kc2 15. Rc4+ Qc3 (15... Kb2 16. Qf6+ Kb3 17. Qxa1 Kxc4 18. Qxa4+ {+-}) (15... Kb3 16. Rc1+ Kb2 17. Qf6+ Kxc1 18. Qxa1+ Kc2 19. Qxa4+ {+-}) 16. Qxh7+ Rd3 17. Rxc3+ Kxc3 18. Qxh6 {+-}) 6. e7 Rf3 7. Kd7 $1 {The first maneuver to lose a tempo.} (7. Re2 $2 {(Rc2+?)} Kd3 8. Rg2 Kc4 {Loss of time.}) 7... Rd3+ 8. Ke6 Re3+ 9. Kf7 Rf3+ 10. Ke8 {Repeating the position, but with BTM. 10...} Re3 (10... Kb4 11. Re2 Rf2 12. Re4+ Kc3 13. Kd7 g2 14. Re3+ Kb2 15. Re1 g1=Q 16. Rxg1 Rd2+ 17. Kc6 Re2 18. Rg2 $1 Rxg2 19. e8=Q Kxa2 20. Kb7 Kb1 21. Qf8 Rb2 22. Qf1+ Kc2 (22... Ka2 23. Kxa7 Rb3 24. b6 {+-}) 23. Qc4+ Kd2 24. Kxa7 {-+}) 11. Rg1 g4 (11... Rd3 12. Re1 g2 13. Kf7 g1=Q 14. Rxg1 Rf3+ 15. Ke6 Re3+ 16. Kd7 Rd3+ 17. Kc6 {+-}) (11... Re2 12. Rxg3 Kb4 (12... Rxa2 13. Rg4+ {etc.}) 13. Rd3 Rxa2 (13... Kc4 14. Rxa3 {+-}) 14. Rd4+ Kb3 15. Re4 {etc.}) (11... Kb4 12. Kd7 Rd3+ 13. Ke6 Re3+ 14. Kd6 Rd3+ (14... Re2 15. Rxg3 {+-}) 15. Ke5 Re3+ 16. Kd4 Rxe7 17. Rb1# {mate! The point of 1.b5! instead of 1.bxa5?}) 12. Rg2 {The beginning of the second maneuver.} (12. Rc1+ $2 Kd3 {=}) 12... Rf3 (12... Kb4 13. Rc2 Re1 (13... g2 14. Rxg2 g3 15. Rc2 {+-}) 14. Rd2 Kc4 15. Kd7 Kc3 16. Rg2 Rd1+ 17. Kc6 Re1 18. Rxg3+ Kb2 19. Rg2+ Ka1 20. Rg1 $1 Rxg1 21. e8=Q Rb1 (21... g3 22. Kb7 Rb1 23. Qe5+ {transposes}) 22. Kb7 g3 23. Qe5+ Kxa2 24. Kxa7 Rb3 (24... g2 25. Qd5+ Ka1 26. Qxg2 Rxb5 27. Ka6 {+-}) 25. b6 Kb1 26. Qe4+ Kb2 27. Qd4+ Kb1 28. Qxa4 Kb2 29. Qxb3+ Kxb3 30. b7 {+-}) 13. Kd7 $1 (13. Re2 $2 Kd3 14. Re1 Kd2 15. Re4 Kd1 {=}) 13... Rd3+ 14. Ke6 Re3+ 15. Kf7 Rf3+ 16. Ke8 Re3 (16... Rc3 17. Rd2 Rd3 18. Re2 {+-}) 17. Rg1 {Again, repeating the position with BTM. 17...} g2 (17... Re2 18. Rxg3 Rxa2 19. Rxg4+ {+-}) (17... Kb4 18. Kd7 Rd3+ 19. Ke6 Re3+ 20. Kd6 {+-}) 18. Rxg2 g3 19. Rg1 Rd3 20. Kf7 $1 {The third maneuver, but this time the wK's rundlauf goes clockwise.} (20. Re1 $2 g2 21. Kf7 Rf3+ {=}) (20. Rc1+ $2 Kb4 21. Kf7 (21. Rc2 Re3 22. Rd2 Kc3 23. Rg2 Kc4 {Loss of time.}) 21... Rf3+ 22. Ke6 Re3+ 23. Kd6 Re2 24. Rd1 Kc3 25. Rg1 Kb2 26. Rxg3 Kxa2 {=}) 20... Rf3+ 21. Ke6 Re3+ 22. Kd7 Rd3+ 23. Ke8 {The third position is repeated with BTM. 23. ..} Re3 24. Rc1+ $1 (24. Rg2 $2 Rf3 $1 25. Kd7 (25. Re2 Kd3 26. Re1 Kd2 27. Re4 Kd1 {=}) (25. Rc2+ Kd3 26. Kd7 Kxc2 27. e8=Q Kd2 $1 {=}) 25... Rd3+ 26. Ke6 Re3+ 27. Kf7 Rf3+ 28. Ke8 Re3 {Loss of time.}) 24... Kd3 (24... Kb4 25. Rc2 ( 25. Kd7 $2 Rd3+ 26. Ke6 Re3+ 27. Kd6 Re2 28. Rd1 Kc3 29. Rg1 Kb2 {=}) 25... Re1 26. Rd2 Kc4 27. Kd7 Kc3 28. Rg2 {+-}) 25. Kd7 Kd2 26. Rf1 $1 (26. Rc8 $2 Rd3+ 27. Ke6 Re3+ 28. Kf7 Rf3+ 29. Kg7 g2 30. e8=Q g1=Q+ {=}) 26... Rd3+ 27. Ke6 Re3+ (27... Ke2 28. Rf8 g2 29. Rg8 Re3+ 30. Kd7 {+-}) 28. Kf7 Kc2 (28... Ke2 29. Rf4 g2 30. Rg4 {+-}) 29. e8=Q Rxe8 30. Kxe8 Kb2 31. Rf7 Kxa2 32. Rxa7 Kb2 33. b6 a2 34. b7 a1=Q 35. b8=Q+ {wins} 1-0[/pgn]

Links:

https://arves.org/arves/index.php/en/aw ... vilag-2014 (Where I found the PGN).
https://arves.org/arves/images/PDF-Awar ... g-2014.pdf (Special Prize, page 5 of the PDF).
https://www.arves.org/arves/images/PDF/ ... y_2016.pdf (Page 62 of the issue, Problem No. 20584).

The position has been quoted for quite some times in forum:

https://talkchess.com/forum3/search.php ... %2F8+w+-+-

BTW, I should have tried SF dev with Andrew's changed one position at first, with empty hash and default setting winning eval and 5. or 6. b5 arises within short time anyhow too at most ones of a few more trials with 30 threads. Especially good is Koivisto 9.2 and Berserk 11.1 at it, regards

[acase]

Well done Peter and Jesus, I was aware that Berserk 11.1 solved it quickly (almost immediately), but didn't try Koivisto 9.2, and Jesus is of course correct about the study that was "hidden" in there. Here is the PGN I have:

[pgn][Event "Modified Study by whom?"]
[Site "Columbus, Ohio"]
[Date "2023.04.05"]
[Round "?"]
[White "Workstation"]
[Black "Workstation"]
[Result "*"]
[BlackElo "2400"]
[Time "20:51:23"]
[WhiteElo "2400"]
[TimeControl "0"]
[SetUp "1"]
[FEN "8/p6p/1p2P1Qp/r2RKBpP/pPk1N3/p2b2p1/P2n2R1/1q1r4 w - - 0 1"]
[Termination "unterminated"]
[PlyCount "44"]
[WhiteType "human"]
[BlackType "human"]

1. Nxd2+ Rxd2 (1. .. Kxb4? {Leads to a quicker loss} 2. Nxb1 Rxd5+ 3. Kxd5
Bxf5+ 4. Rd2 hxg6 5. Rxd1 gxh5 6. Rd4+ Ka5 7. e7 Bg6 8. Nc3 {etc.}) 2.
Bxd3+ Qxd3 3. Qxd3+ Rxd3 4. Rxa5 bxa5 {After black's 4th move, This is the
root position of a study by Richard Becker from the WCCI 2013-2015, Special
Prize MSV -2014} (4. .. Re3+ 5. Kd6 bxa5 6. b5!! {Leads to the main line})
5. b5!! (5. bxa5? {=}) 5. .. Re3+ 6. Kd6 Rd3+ 7. Kc6 Re3 (7. .. Rc3 8. e7
Kb4+ 9. Kd5 Re3 10. Kd6 Rd3+ 11. Ke6 Re3+ 12. Kd7 Rd3+ 13. Ke8 {Pretty much
the same idea as the mainline.}) 8. Kd7 Rd3+ 9. Ke8 Rc3 10. e7 Rf3 11. Kd7
Rd3+ 12. Ke6 Re3+ 13. Kf7 Rf3+ 14. Ke8 Re3 15. Rg1 g4 16. Rg2 Rf3 17. Kd7
Rd3+ 18. Ke6 Re3+ 19. Kf7 Rf3+ 20. Ke8 Re3 21. Rg1 g2 22. Rxg2 g3 *
[/pgn]