Tough Tactical Test 1

[Andrew]

Only hit on the unsolved Ng6+ position with google was here in German

https://forum.computerschach.de/cgi-bin ... ?tid=11642

Have you solved it with any other engines?


[Event "Teststellung Festung bauen"]
[Site ""]
[Date ""]
[Round ""]
[White "TTT1.023"]
[Black ""]
[Result "1/2-1/2"]
[FEN "3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1"]
[SetUp "1"]
[PlyCount "1"]

{In dieser schwierigen Situation findet Weiß eine erstaunliche Rettung mit} 1.
Ng6+ {!!} 1... hxg6 2. Qxg7+ (2. Qf6 exd3+ 3. Kd2 (3. Kd1 Kh7) 3... d4 4. Qxg6
dxe3+ 5. Kxe3 Nf8+ {- +}) 2... Nxg7 3. Bxg7+ Kxg7 4. d4 {und Weiß hat eine
uneinnehmbare Festung errichtet und die Partie ist Remis.} 1/2-1/2

[peter]

Only hit on the unsolved Ng6+ position with google was here in German

https://forum.computerschach.de/cgi-bin ... ?tid=11642

Have you solved it with any other engines?


[Event "Teststellung Festung bauen"]
[Site ""]
[Date ""]
[Round ""]
[White "TTT1.023"]
[Black ""]
[Result "1/2-1/2"]
[FEN "3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1"]
[SetUp "1"]
[PlyCount "1"]

{In dieser schwierigen Situation findet Weiß eine erstaunliche Rettung mit} 1.
Ng6+ {!!} 1... hxg6 2. Qxg7+ (2. Qf6 exd3+ 3. Kd2 (3. Kd1 Kh7) 3... d4 4. Qxg6
dxe3+ 5. Kxe3 Nf8+ {- +}) 2... Nxg7 3. Bxg7+ Kxg7 4. d4 {und Weiß hat eine
uneinnehmbare Festung errichtet und die Partie ist Remis.} 1/2-1/2

No problem for Crystallite


[fen]3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1[/fen]


Analysis by Crystallite 1.1a-bmi2:

1.Db2 d4 2.Sxc6 exd3+ 3.Kc1 Lf7 4.Sxd8 Txd8 5.Kd2 dxe3+ 6.Kxe3 d2
Black is better: -/+ (-0.89) Depth: 10/17 00:00:00
...
1.Db2 d4 2.Dc1 dxe3 3.Dxe3 Sf8 4.Dd2 Lb3+ 5.Kb1 exd3 6.Kc1 Kg8 7.Sxc6 Td5 8.Se5 Sge6 9.Kb1 Sxf4 10.g6 S4xg6 11.Sxg6 Sxg6 12.Dh6 Te1+ 13.Kb2 d2
Black is clearly winning: -+ (-4.56) Depth: 16/33 00:00:01 23241kN
1.Sg6+ hxg6 2.Dxg7+ Sxg7 3.Lxg7+ Kxg7 4.d4 Tc8 5.Kd1 Le6 6.Kc2
The position is equal: = (0.00) Depth: 17/33 00:00:01 33689kN
...
1.Sg6+ hxg6 2.Dxg7+ Sxg7 3.Lxg7+ Kxg7 4.d4 Tc8 5.Kd1 Le6 6.Kc2
The position is equal: = (0.00) Depth: 40/15 00:00:05 254MN
1146: Tough,Tactical - Test,Toughest 1/2-1/2 0.0, CCC
3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1
...
1.Sg6+ hxg6 2.Dxg7+ Sxg7 3.Lxg7+ Kxg7 4.d4 Le6 5.Kd2 Kg8 6.Kd1 Kg7
The position is equal: = (0.00) Depth: 48/14 00:01:17 3263MN

All options default, except Syzgyy Probe Limit 0, 30 threads of 16x3.5Ghz Ryzen- CPU, 8G hash.
3 trials all were as fast as the pasted one.

Crystallite is a modification of Crystal from user dorsz of Open Chess Forum:

https://www.open-chess.org/viewtopic.php?p=25642#p25642

LC0 with 3070ti:

3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1

Analysis by Lc0 v0.30.0-dev+git.77b5659:

1.Qb2 d4 2.Nxc6
Schwarz hat entscheidenden Vorteil: -+ (-3.21) Tiefe: 2/3 00:00:00
...
1.Qb2 d4 2.Nxc6 Rd7 3.Ne5 Rd5 4.Qc1 Nxc5 5.Bxd4 Nb3 6.Qg1 Rc8+ 7.Kb2 exd3 8.Nxd3 Nxd4 9.exd4 Rc4 10.Ne5 Rcxd4 11.g6 Rd2+ 12.Kc3 Rd6 13.Nf7+
Schwarz steht klar auf Gewinn: -+ (-5.62) Tiefe: 13/30 00:00:06 77kN
1.Ng6+ hxg6 2.Qxg7+ Nxg7 3.Bxg7+ Kxg7 4.d4 Re5 5.Kd2 Rd6 6.Kd1 Be6 7.Kd2 Bd7 8.Kd1 Kf7 9.Kd2 Bc8 10.Ke1 Rd8 11.Kd2 Rde8 12.Ke2
Schwarz steht deutlich besser: -+ (-1.89) Tiefe: 13/30 00:00:10 173kN
...
1.Ng6+ hxg6 2.Qxg7+ Nxg7 3.Bxg7+ Kxg7 4.d4 Ra8 5.Kc1 Ra5 6.Kd2 Re5 7.Ke1 Be6 8.Kd2 Bc8 9.Kd1 Bd7 10.Kd2 Kf8 11.Kc2 Re8 12.bxa5
Schwarz steht deutlich besser: -+ (-1.87) Tiefe: 17/31 00:01:17 5802kN


BTW: an even harder one fortress test- position is this one out of Hard Talkchess- suite.


[fen]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1[/fen]

Rxh1!

Here even Crystallite and Crystal struggle, Sting gets it at once and LC0 within a minute at 3070ti, yet eval isn't any of =.

Pity I don't know the real source of any of the two positions, second one is a study, as far as I know, and I think, Arpad Rusz brought it to CCC long ago, but there's too much mentioning of it in too many threads about HTC- positions to find the one posting I think I found it first regards

[Andrew]

Thanks Peter! I can confirm it finds it very quickly.

I tested a few different engines and another SugaR AI ICCF 2.30
also finds it quickly, but doesn't show 0.0, instead around -1.4

Andrew

[Andrew]

This page also has a range of interesting fortress position, including the second one of
yours.

https://lczero.org/blog/2018/10/achille ... es-neural/

Andrew

[Ajedrecista]

Hello Peter:

[...]


[fen]3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1[/fen]


[...]


[fen]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1[/fen]


[...]

Pity I don't know the real source of any of the two positions, second one is a study, as far as I know, and I think, Arpad Rusz brought it to CCC long ago, but there's too much mentioning of it in too many threads about HTC- positions to find the one posting I think I found it first regards

The first mention of the second position that I found is a post by George Tsavdaris from 2013, just searching the FEN in the TalkChess search engine:

But to make up here is an extra very difficult position!!

Reading the thread, you asked about the source of the problem and George answered that he forgot where he found the problem. The second oldest menton on the Internet to the position that I found, other than the TalkChess thread of 2013, is a topic in an Italian forum few days later:

Un problema dubbio (la soluzione)

I have been unsuccessful in my search of the sources of the two problems, not even at the old CCC archive.

Regards from Spain.

Ajedrecista.

[peter]

Hello Peter:

[...]


[fen]3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1[/fen]


[...]


[fen]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1[/fen]


[...]

Pity I don't know the real source of any of the two positions, second one is a study, as far as I know, and I think, Arpad Rusz brought it to CCC long ago, but there's too much mentioning of it in too many threads about HTC- positions to find the one posting I think I found it first regards

The first mention of the second position that I found is a post by George Tsavdaris from 2013, just searching the FEN in the TalkChess search engine:

But to make up here is an extra very difficult position!!

Reading the thread, you asked about the source of the problem and George answered that he forgot where he found the problem. The second oldest menton on the Internet to the position that I found, other than the TalkChess thread of 2013, is a topic in an Italian forum few days later:

Un problema dubbio (la soluzione)

I have been unsuccessful in my search of the sources of the two problems, not even at the old CCC archive.

Thanks a lot for trying regards

[Dann Corbit]

For this one:
[fen]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - [/fen]-
I have this link:
https://lczero.org/blog/2018/10/achille ... es-neural/

[peter]

For this one:
[fen]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - [/fen]-
I have this link:
https://lczero.org/blog/2018/10/achille ... es-neural/

Thanks, Dann, but no origin given there neither

[Event “?"] [Site “?"] [Date “????.??.??"] [Round “?"] [White “New game”] [Black “?"] [Result “*"] [SetUp “1”] [FEN “1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1”] [PlyCount “31”] {White draws with only one move.} 1. Rxh1 $3 Nxh1 2. e6 $1 dxe6 3. Bc7+ Kb7 4. Bxg3 {[%cal Gg3h4,Gg3e1] White’s Bishop is immune as if it gets captured it’s stalemate. Stays in e1-h4 diagonal and captures everything.} Kc6 5. Bh4 Kd5 6. Bg3 e5 7. Bf2 e4 8. Bg3 Kd4 9. Bh4 e3 10. Be1 f4 11. Bh4 f3 12. Bf2 Kd3 13. Bg3 f2 14. Bxf2 e2 15. Bh4 Kd2 16. Bf2 $11 {Hopefully one day computers will find this move.} *

[peter]

Hello Peter:

[...]


[fen]3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - 0 1[/fen]


[...]


[fen]1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - 0 1[/fen]


[...]

Pity I don't know the real source of any of the two positions, second one is a study, as far as I know, and I think, Arpad Rusz brought it to CCC long ago, but there's too much mentioning of it in too many threads about HTC- positions to find the one posting I think I found it first regards

The first mention of the second position that I found is a post by George Tsavdaris from 2013, just searching the FEN in the TalkChess search engine:

But to make up here is an extra very difficult position!!

Reading the thread, you asked about the source of the problem and George answered that he forgot where he found the problem. The second oldest menton on the Internet to the position that I found, other than the TalkChess thread of 2013, is a topic in an Italian forum few days later:

Un problema dubbio (la soluzione)

I have been unsuccessful in my search of the sources of the two problems, not even at the old CCC archive.

From this one old thread,

forum3/viewtopic.php?p=606840#p606840

I got this one link to second position:

https://chessengine.blogspot.com/p/horizon-limit.html

[Dann Corbit]

This position is from a game by dorszcz:
[fen]3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - -[/fen]