Testposition King Safety

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
Eelco de Groot
Posts: 4265
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Testposition King Safety

Post by Eelco de Groot » Sun Oct 09, 2016 9:26 pm

Given by someone on Fishcooking. All pieces still on board, except two pawns, still in the opening, Sicilian perhaps?

https://groups.google.com/forum/#!topic ... J1X9lhkku0

[D]rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ - 0 13

Difficult for normal engines, but if you have aggressive settings, maybe not so hard, but then you have to find the correct follow up moves!
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

Werewolf
Posts: 1306
Joined: Thu Sep 18, 2008 8:24 pm

Re: Testposition King Safety

Post by Werewolf » Sun Oct 09, 2016 9:46 pm

I was wondering if this was a French.

Anyway, is the winning line something like this:
1.Bh7+ Kh7
2. Qh5+ Kg8
3. o-o-o Ne4
4.Ne4 de
5. Rd-g1 g6
6. Qe2 with h4 to follow.

?

User avatar
Eelco de Groot
Posts: 4265
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Testposition King Safety

Post by Eelco de Groot » Sun Oct 09, 2016 10:06 pm

Hi Carl, it could be a French. I really don't know. It would be nice to know, if it was something playable!

With the line Rainbow Serpent is at the moment not so happy, but that could change, now it is going up a little for White again. Could be this version is not so good, but at long depths I would expect that not to matter much.

[pgn]
[Event "?"]
[Site "?"]
[Date "2016.10.10"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ -"]

1. Bxh7+ Kxh7 2. Qh5+ Kg8 3. O-O-O Ne4 4. Nxe4 dxe4 5. Rdg1
*
[/pgn]
[D]rnb2rk1/1pq1bpp1/p3p3/4P1PQ/3NpP2/4B3/PPP4P/2K3RR b - -

28/64 3:39 -1.19 5...g6 6.Qe2 Bd7 7.h4 Nc6 8.h5 Nxd4
9.Bxd4 Bb5 10.Qg4 Qc4 11.hxg6 fxg6
12.Qh3 Kf7 13.Qh7+ Ke8 14.Qxg6+ Kd7
15.Rd1 Qxa2 16.Rh7 Rae8 17.Bc5+ Kc8
18.Rxe7 Rxe7 (1.058.279.954) 4818

29/64 4:36 -1.11-- 5...g6 6.Qe2 (1.327.285.374) 4795

29/64 4:44 -1.03-- 5...g6 6.Qe2 (1.362.045.207) 4789

29/64 4:45 -0.91-- 5...g6 6.Qe2 (1.368.941.742) 4788

29/64 4:48 -0.74-- 5...g6 6.Qe2 (1.381.611.655) 4787

29/64 4:52 -0.51-- 5...g6 6.Qe2 (1.401.262.666) 4785

29/64 5:36 -0.31 5...g6 6.Qe2 Bd7 7.h4 Nc6 8.h5 Nxd4
9.Bxd4 Bb5 10.Qh2 Kg7 11.f5 gxf5
12.g6 fxg6 13.Rxg6+ Kf7 14.Qg2 Ke8
15.Rxe6 Qc4 16.Qg6+ Kd7 17.Rxe7+ Kxe7
18.Qd6+ Ke8 (1.602.473.797) 4761

30/64 7:27 -0.40++ 5...g6 (2.110.910.761) 4722

30/64 7:38 -0.31-- 5...g6 6.Qe2 (2.165.697.121) 4718

30/64 7:55 -0.39++ 5...g6 (2.242.063.685) 4716

30/64 8:06 -0.39 5...g6 6.Qe2 Nc6 7.Nxc6 Qxc6 8.h4 Bd7
9.h5 Kg7 10.hxg6 fxg6 11.Rh6 Rf5
12.Rgh1 Rg8 13.Qh2 Kf7 14.Rh8 Rg7
15.Kb1 Be8 16.Qh6 Bf8 17.Rd1 Qc8
18.Rd4 Bc6 (2.292.442.831) 4714
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

Sven
Posts: 3877
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: Testposition King Safety

Post by Sven » Sun Oct 09, 2016 10:28 pm

After 5...g6, maybe 6.Qh4 with the idea Rg1-g3-h3 looks a bit more promising to me than 6.Qe2. Black needs to bring his bishop to g7. After that, White could plan to move Rh3-c3 and/or f4-f5.

User avatar
Eelco de Groot
Posts: 4265
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Testposition King Safety

Post by Eelco de Groot » Sun Oct 09, 2016 10:42 pm

Hi Sven, I think White has lost too much tempo in both your lines. After 6. Qh4. Black is already winning it seems.

[pgn][Event "?"]
[Site "?"]
[Date "2016.10.10"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[SetUp "1"]
[FEN "rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ -"]

1. Bxh7+ Kxh7 2. Qh5+ Kg8 3. O-O-O Ne4 4. Nxe4 dxe4 5. Rdg1 g6 8. Qh4 Rd8 *[/pgn]


[D]rnb2rk1/1pq1bp2/p3p1p1/4P1P1/3NpP1Q/4B3/PPP4P/2K3RR b - -

31/62 7:06 -1.41 6...Rd8 7.Rg3 Bf8 8.Bf2 Bg7 9.Rc3 Qe7
10.Nb3 e3 11.Bxe3 b6 12.Re1 Nd7
13.Rc7 Rb8 14.Qf2 Qe8 15.Qf3 Bf8
16.Qc6 Nc5 17.Bxc5 Qxc6 18.Rxc6 bxc5
19.Nxc5 Bxc5 (1.972.512.818) 4626
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

kgburcham
Posts: 2016
Joined: Sun Feb 17, 2008 3:19 pm

Re: Testposition King Safety

Post by kgburcham » Sun Oct 09, 2016 11:49 pm

[D]rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ -

1. Bxh7+ Kxh7 2. Qh5+ Kg8 3. O-O-O g6 4. Qg4 b5 5. h4 b4
6. Nce2 Kg7 7. h5 Rg8 8. hxg6 fxg6 9. Rh6 Kf8 10. Qh3 Ke8
11. f5 gxf5 12. Nxf5 exf5 13. Qh5+ Kd8 14. Rxd5+ Bd7
15. Rh8 Rf8 16. Rxf8+ Bxf8 17. Qf7 Kc8 18. Qxf8+ Kb7
19. Qxc5 a5 20. Qxc7+ Kxc7 21. Nf4 b3 22. axb3 Na6 23. Rd6
Bc6 24. Ne6+ Kb7 25. g6 Nc7 26. Nc5+ Kb6 27. g7 Rg8
28. Ne6+ Kb7 29. Nd8+ Ka8 30. Nxc6 Kb7 31. Nxa5+ Kc8
32. Nc6 Kb7 33. Rd8 Ne8 34. Ne7 Rxg7 35. Rxe8 f4 36. Bxf4
Rg1+ 37. Kd2 Rg2+ 38. Kd3 Rf2 39. Nd5 Kc6 40. Re6+ Kd7
41. Re7+ Kc8 42. e6 Rf3+ 43. Ke4 Rf2 44. Rc7+ Kb8 45. e7
Re2+ 46. Be3 Rxe3+ 47. Kxe3 Ka8 48. e8=Q# *
no chess program was born totally from one mind. all chess programs have many ideas from many minds.

Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 10:41 am

Re: Testposition King Safety

Post by Lyudmil Tsvetkov » Mon Oct 10, 2016 5:55 am

kgburcham wrote:[D]rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ -

1. Bxh7+ Kxh7 2. Qh5+ Kg8 3. O-O-O g6 4. Qg4 b5 5. h4 b4
6. Nce2 Kg7 7. h5 Rg8 8. hxg6 fxg6 9. Rh6 Kf8 10. Qh3 Ke8
11. f5 gxf5 12. Nxf5 exf5 13. Qh5+ Kd8 14. Rxd5+ Bd7
15. Rh8 Rf8 16. Rxf8+ Bxf8 17. Qf7 Kc8 18. Qxf8+ Kb7
19. Qxc5 a5 20. Qxc7+ Kxc7 21. Nf4 b3 22. axb3 Na6 23. Rd6
Bc6 24. Ne6+ Kb7 25. g6 Nc7 26. Nc5+ Kb6 27. g7 Rg8
28. Ne6+ Kb7 29. Nd8+ Ka8 30. Nxc6 Kb7 31. Nxa5+ Kc8
32. Nc6 Kb7 33. Rd8 Ne8 34. Ne7 Rxg7 35. Rxe8 f4 36. Bxf4
Rg1+ 37. Kd2 Rg2+ 38. Kd3 Rf2 39. Nd5 Kc6 40. Re6+ Kd7
41. Re7+ Kc8 42. e6 Rf3+ 43. Ke4 Rf2 44. Rc7+ Kb8 45. e7
Re2+ 46. Be3 Rxe3+ 47. Kxe3 Ka8 48. e8=Q# *
It is either that I am kokoo, but white does not have to castle and mates much faster with 3.Rhg1, and then Rg1-g3-h3. I do not see how you prevent mate in a couple of moves.

This was definitely a French advanced variation.

My root position static eval above would be at least +200cps white advantage.

But it is also quite possible that I am a bit lagaa at this moment.

Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 10:41 am

Re: Testposition King Safety

Post by Lyudmil Tsvetkov » Mon Oct 10, 2016 6:20 am

It proves I am partly kokoo, and partly lagaa.

3.Rg1 is still the best move, but mating sequence is a bit longer.

You need to play Qh3, place one rook in front of the white queen along the h file, play Rh7, sacrifice the rook for the defending g7 bishop, and only then bring the other, queen-side rook, mating.

A bit long for normal (unkokoo) engines.

Here how I mated Andscacs with the above sequence(actually stopped before mate, as it is boring, but Andscacs shows +400cps; Nf8 instead of Nc5 is met by the same Rg7+ sacrifice):

[pgn][Event "Blitz 2m+2s"]
[Site "Microsoft"]
[Date "2016.10.10"]
[Round "?"]
[White "Lyudmil Tsvetkov, owner"]
[Black "Andscacs 0.86134"]
[Result "1-0"]
[Annotator "owner"]
[SetUp "1"]
[FEN "rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ - 0 1"]
[PlyCount "22"]
[TimeControl "120+2"]

{512MB, OWNER-PC} 1. Bxh7+ {0} Kxh7 {-1.72/18 2} 2. Qh5+ {2} Kg8 {-2.71/1 0} 3.
Rg1 {3} g6 {-0.97/23 12} 4. Qh3 {3} Re8 {-0.98/23 8} 5. Rg4 {3} Bf8 {-1.27/23 3
} 6. Rh4 {2} Bg7 {-1.28/22 6} 7. Rh7 {15} Nbd7 {-0.94/19 0} 8. O-O-O {3} Ne4 {
-0.58/19 7} 9. Nxe4 {4} dxe4 {-1.49/20 0} 10. Rg1 {5} Nc5 {-1.00/16 3} 11. Rg4
{14} Nd3+ {3.78/24 10} 1-0

[/pgn]

Sven
Posts: 3877
Joined: Thu May 15, 2008 7:57 pm
Location: Berlin, Germany
Full name: Sven Schüle
Contact:

Re: Testposition King Safety

Post by Sven » Mon Oct 10, 2016 9:29 am

Lyudmil Tsvetkov wrote:It proves I am partly kokoo, and partly lagaa.

3.Rg1 is still the best move, but mating sequence is a bit longer.

You need to play Qh3, place one rook in front of the white queen along the h file, play Rh7, sacrifice the rook for the defending g7 bishop, and only then bring the other, queen-side rook, mating.

A bit long for normal (unkokoo) engines.

Here how I mated Andscacs with the above sequence(actually stopped before mate, as it is boring, but Andscacs shows +400cps; Nf8 instead of Nc5 is met by the same Rg7+ sacrifice):

[pgn][Event "Blitz 2m+2s"]
[Site "Microsoft"]
[Date "2016.10.10"]
[Round "?"]
[White "Lyudmil Tsvetkov, owner"]
[Black "Andscacs 0.86134"]
[Result "1-0"]
[Annotator "owner"]
[SetUp "1"]
[FEN "rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ - 0 1"]
[PlyCount "22"]
[TimeControl "120+2"]

{512MB, OWNER-PC} 1. Bxh7+ {0} Kxh7 {-1.72/18 2} 2. Qh5+ {2} Kg8 {-2.71/1 0} 3.
Rg1 {3} g6 {-0.97/23 12} 4. Qh3 {3} Re8 {-0.98/23 8} 5. Rg4 {3} Bf8 {-1.27/23 3
} 6. Rh4 {2} Bg7 {-1.28/22 6} 7. Rh7 {15} Nbd7 {-0.94/19 0} 8. O-O-O {3} Ne4 {
-0.58/19 7} 9. Nxe4 {4} dxe4 {-1.49/20 0} 10. Rg1 {5} Nc5 {-1.00/16 3} 11. Rg4
{14} Nd3+ {3.78/24 10} 1-0

[/pgn]
From the line above:
8...Ne4?? +-
8...Nf8 9.Rxg7+ Kxg7 10.Qh7+ Kg8 11.Rg1 += looks promising for White (but no quick mate)
But: 7...Ne4 8.O-O-O =+ is not clear, seems to be better for Black

Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 10:41 am

Re: Testposition King Safety

Post by Lyudmil Tsvetkov » Mon Oct 10, 2016 9:48 am

Sven Schüle wrote:
Lyudmil Tsvetkov wrote:It proves I am partly kokoo, and partly lagaa.

3.Rg1 is still the best move, but mating sequence is a bit longer.

You need to play Qh3, place one rook in front of the white queen along the h file, play Rh7, sacrifice the rook for the defending g7 bishop, and only then bring the other, queen-side rook, mating.

A bit long for normal (unkokoo) engines.

Here how I mated Andscacs with the above sequence(actually stopped before mate, as it is boring, but Andscacs shows +400cps; Nf8 instead of Nc5 is met by the same Rg7+ sacrifice):

[pgn][Event "Blitz 2m+2s"]
[Site "Microsoft"]
[Date "2016.10.10"]
[Round "?"]
[White "Lyudmil Tsvetkov, owner"]
[Black "Andscacs 0.86134"]
[Result "1-0"]
[Annotator "owner"]
[SetUp "1"]
[FEN "rnb2rk1/1pq1bppp/p3p3/2npP1P1/3N1P2/2NBB3/PPP4P/R2QK2R w KQ - 0 1"]
[PlyCount "22"]
[TimeControl "120+2"]

{512MB, OWNER-PC} 1. Bxh7+ {0} Kxh7 {-1.72/18 2} 2. Qh5+ {2} Kg8 {-2.71/1 0} 3.
Rg1 {3} g6 {-0.97/23 12} 4. Qh3 {3} Re8 {-0.98/23 8} 5. Rg4 {3} Bf8 {-1.27/23 3
} 6. Rh4 {2} Bg7 {-1.28/22 6} 7. Rh7 {15} Nbd7 {-0.94/19 0} 8. O-O-O {3} Ne4 {
-0.58/19 7} 9. Nxe4 {4} dxe4 {-1.49/20 0} 10. Rg1 {5} Nc5 {-1.00/16 3} 11. Rg4
{14} Nd3+ {3.78/24 10} 1-0

[/pgn]
From the line above:
8...Ne4?? +-
8...Nf8 9.Rxg7+ Kxg7 10.Qh7+ Kg8 11.Rg1 += looks promising for White (but no quick mate)
But: 7...Ne4 8.O-O-O =+ is not clear, seems to be better for Black
both are very quick mates; check more carefully

I bet for both lines you mention SF, with a quick search, will show +700cps scores.

On 7...Ne4, Ne4 de4 0-0-0 mates rather quick, as for your 8...Nf8 line, certainly +700cps score is not just promising.

This is a standard mating pattern, I have played thousands of those.

I would also bet, Houdini 5, if available, would relatively quickly see this mate, as it attacks very well on the h file.

Post Reply