Author: Eelco de Groot
Date: 13:22:20 01/25/01
Go up one level in this thread
On January 23, 2001 at 20:38:08, Howard Exner wrote:
>[Event "?"]
>[Site "?"]
>[Date "??.??.????"]
>[Round "?"]
>[White ""]
>[Black ""]
>[Result "*"]
>[WhiteElo "?"]
>[BlackElo "?"]
>[FEN "r4r1k/4bppb/2n1p2p/p1n1P3/1p1p1BNP/3P1NP1/qP2QPB1/2RR2K1 w - - 0 1"]
>
>1. Ng5 hxg5 2. hxg5 Rfc8 3. Nf6 gxf6 4. gxf6 Bf8 5. Qh5 Kg8 6. Bh6 Bg6
>7. Qh1 Nxe5 8. Bxf8 Qxb2 9. Rf1 Qd2 10. Bg7 Bh7 11. Bxa8 Qxd3 12. Qb7 Qxf1+
>13. Kxf1 Bd3+ 14. Kg2 Nxb7 15. Rxc8+ Kh7 16. Rh8+ Kg6 17. Rh6+ Kf5 18.
>Rh5+ Kg6 19. Rxe5 Nd6 20. g4 Ne8 21. Bf8 Nxf6 22. Rxa5 b3 23. Ba3 Nxg4
>24. Bb2 e5 25. Bf3 Bf5 26. Bxg4 Bxg4 27. Rxe5 *
>
>Here's a start at trying to uncover a winning line. There is most likely better
>moves to be found but posting a plausible line may clarify what to play or not
>to play. Having the game score would be usefull but it looks like it is
>difficult to find.
>
>I entered in the first 4 moves manually on a gut feeling and let Gambit Tiger
>wade through the rest. I doubt if this is the gamescore line but it wins barring
>any oversights.
Hello Howard, I think the line is looking good for White! I couldn't believe at
first that Gambit Tiger actually managed to beat itself here, as legendary "Q5T"
still thought White almost 4 points down after 3.Ng4-f6. But after 5.Qe2-h5! I
think it should be possible for programs to see that is not going to be all
roses for Black. If they search deep enough, that is.. In the position after 8.
..Qa2xb2 there is still an 11 ply search needed but the score is almost +2 by
then, a difference of almost 6 points! I didn't let the program look at all
possibilities for Black, six engines simultaneously is about the max but I think
taking back the Knight with 1. ..hxg5 may indeed not be such a good idea!
Maybe programs with position learning could work backwards from 8. ..Qxb2 to see
if Black can do a little better somewhere, between 1. ..hxg5 and 8. ..Qa2xb2
After 8. ..Qa2xb2 in your main line:
Last Move : 8... Qxb2 (White to play)
00:00 02.18 -1.00 9.Bg7 Bh5 10.Bxa8 Bxd1 11.Rxd1 Ncxd3
00:00 03.00 -1.32 9.Rb1 Qe2 10.Bxc5 Ng4 11.Bxd4
00:00 04.00 -0.79 9.Bg7 Bh7 10.Rxc5 Rxc5 11.Bxa8 Nxd3
00:02 05.00 -1.28 9.Bg7 Bh7 10.Rb1 Qe2 11.Bxa8 Nexd3
00:05 05.01 -0.89 9.Rb1 Qe2 10.Bxc5 Rxc5 11.Qh6 Qxf2+ 12.Kxf2
00:07 05.14 -0.89 9.Rf1
00:07 05.14 0.14 9.Rf1 Ncxd3 10.Rxc8 Rxc8 11.Qh6 Nf3+ 12.Bxf3
00:09 06.00 -0.31 9.Rf1 Qxc1 10.Rxc1 Ncxd3 11.Rxc8 Rxc8 12.Bd6
00:20 07.00 0.89 9.Rf1 Qxc1 10.Rxc1 Ncxd3 11.Rxc8 Rxc8
00:42 08.00 0.86 9.Rf1 Qd2 10.Bxa8 Bxd3 11.Rxc5 Rxf8 12.Rxe5+ Bxf1
01:34 09.00 0.47 9.Rf1 Qd2 10.Bg7 Bh7 11.Bxa8 Qxd3 12.Rfd1 Qf5
03:15 10.00 0.60 9.Rf1 Qd2 10.Bg7 Bh7 11.Bxa8 Qxd3 12.Rfd1 Qf5
06:41 11.00 1.94 9.Rf1 Qd2 10.Bg7 Bh7 11.Rxc5 Qxd3 12.Rxe5 Ra7
43:32 12.00 3.41 9.Rf1 Qd2 10.Bg7 Bh7 11.Rxc5 Qxd3 12.Rxe5 Ra7
03:23 13.00 2.28 9.Rf1 Qd2 10.Bg7 Bh7 11.Rxc5 Rxc5 12.Be4 Nf3+
05:25 14.00 2.43 9.Rf1 Qd2 10.Bg7 Bh7 11.Rxc5 Rxc5 12.Be4 Nf3+ 13.Qxf3
3 plies earlier it only senses trouble 12 plies deep. Q5T wants to play 7.
..Qa2xb2 now but that is only a transposition:
Last Move : 7.Qh1 (Black to play)
00:00 01.03 4.35 7..Nxd3 8.Bxf8 Kxf8 9.Bxc6 Nxc1 10.Rxc1
00:01 05.00 4.78 7..Nxd3 8.Bxf8 Nxc1 9.Qh6 Ne2+ 10.Kf1
00:04 06.00 0.15 7..Nxd3 8.Bxf8 Ne7 9.Rxc8 Rxc8 10.Bxe7 Qxb2
00:19 06.06 1.71 7..Qxb2 8.Bxf8 Nxe5 9.Rf1 Ncxd3
00:41 07.00 1.55 7..Qxb2 8.Bxf8 Nxe5 9.Bxc5 Nxd3 10.Qh6 Qxf2+
01:03 08.00 1.27 7..Qxb2 8.Bxf8 Nxe5 9.Bxc5 Nxd3 10.Qh6 Qxf2+
01:36 09.00 1.27 7..Qxb2 8.Bxf8 Nxe5 9.Bxc5 Nxd3 10.Qh6 Qxf2+
03:22 10.00 1.36 7..Qxb2 8.Bxf8 Nxe5 9.Rf1 Qd2 10.Bg7
15:17 11.00 1.10 7..Qxb2 8.Bxf8 Nxe5 9.Rf1 Qd2 10.Bg7
39:29 12.00 -0.19 7..Qxb2 8.Bxf8 Nxe5 9.Rf1 Qd2 10.Bg7 Bh7
08:23 13.00 -0.81 7..Qxb2 8.Bxf8 Nxe5 9.Rf1 Qd2 10.Bg7
After 5.Qe2-h5! which may be a key-move in itself and may be difficult to find
for some programs? At least Q5T, searching 4 plies earlier again, still thinks
Black is well ahead. I let this one simmer for 24 hours but it doesn't see any
White advantage yet. At least at 15 ply it does still follow part of the main
line, but chooses 7.Qh5-h4 again. If other programs also show a plusscore for
Black here I think it must mean they do not apply good enough King Safety
knowledge for this position? So it might make a good testposition, especially if
they do not play 5.Qe2-h5 one ply earlier. This seems to be the breaking point
in "my" analysis; Q5T has the right continuation, at least what we think is best
for Black, but the wrong eval.
[Event ""]
[Site ""]
[Date ""]
[Round ""]
[White ""]
[Black ""]
[Result ""]
[FEN "r4r1k/4bppb/2n1p2p/p1n1P3/1p1p1BNP/3P1NP1/qP2QPB1/2RR2K1 w - - 0 1"]
1.Ng5 hxg5 2.hxg5 Rfc8 3.Nf6 gxf6 4.gxf6 Bf8 5.Qh5 *
[D]r1r2b1k/5p1b/2n1pP2/p1n1P2Q/1p1p1B2/3P2P1/qP3PB1/2RR2K1 w - -
Last Move : 5.Qh5 (Black to play)
00:00 04.00 2.30 5..Qxb2 6.Rxc5 Nd8 7.Be4
00:01 04.02 2.48 5..Nxe5 6.Bxe5 Qxb2 7.Rxc5 Rxc5 8.Be4 Bh6 9.Qxh6
00:01 04.06 4.72 5..Kg8 6.Bxc6 Rxc6 7.Rc2 Ra7
00:02 05.00 4.26 5..Kg8 6.Rxc5 Bxc5 7.Qg5+ Bg6 8.Rc1+
00:05 06.00 4.84 5..Kg8 6.Rd2 Ra7 7.Bxc6 Rxc6 8.Rdc2 Bxd3
00:09 07.00 4.38 5..Kg8 6.Rxc5 Bxc5 7.Be4 Bxe4 8.dxe4 Nd8 9.Bc1
00:17 08.00 4.16 5..Kg8 6.Rxc5 Bxc5 7.Be4 Bxe4 8.dxe4 Nd8 9.Bc1
00:49 09.00 4.32 5..Kg8 6.Bh6 Bg6 7.Qg5 Nxd3 8.Rxd3 Qxb2 9.Bxf8 Nxe5
01:29 10.00 0.66 5..Kg8 6.Bh6 Bg6 7.Qh1 Nxe5 8.Bxf8 Qxb2 9.Bg7 Bh7 10.Bxa8
03:30 10.01 1.61 5..Nxe5 6.Bxe5 Kg8 7.Bxa8 Bg6 8.Qh4 Rxa8 9.Bxd4 Nb3 10.Rc7
08:29 11.00 1.61 5..Nxe5 6.Bxe5 Kg8 7.Bxa8 Bg6 8.Qh4 Rxa8 9.Bxd4 Nb3 10.Rc7
Nxd4
08:45 11.01 1.61 5..Kg8
09:03 11.01 1.66 5..Kg8 6.Bh6 Bg6 7.Qh1 Nxe5 8.Bxf8 Qxb2 9.Bg7
11:41 12.00 0.66 5..Kg8 6.Bh6 Bg6 7.Qh1 Nxe5 8.Bxf8 Qxb2 9.Bg7
58:15 12.01 1.45 5..Nxe5 6.Bxe5 Kg8 7.Bxa8 Rxa8 8.Bxd4 Bg6 9.Qh1 Rc8 10.Rxc5
46:04 13.00 1.37 5..Nxe5 6.Bxe5 Kg8 7.Bxa8 Rxa8 8.Bxd4 Nd7 9.Rc7 Rd8
50:42 13.01 1.37 5..Kg8
54:37 13.01 1.52 5..Kg8 6.Bh6 Bg6 7.Qh4 Nxe5 8.Bxf8 Qxb2
42:22 14.00 1.43 5..Kg8 6.Bh6 Bg6 7.Qh1
28:20 14.01 1.55 5..Nxe5 6.Bxe5 Kg8 7.Bxa8
38:04 15.00 1.45 5..Nxe5 6.Bxe5 Kg8 7.Bxa8 Rxa8 8.Bxd4 Na4 9.Rc7 Bg6
07:06 15.01 1.52 5..Kg8 6.Bh6 Bg6 7.Qh4 Nxe5
-------
The rest of the analysis is necessarily a bit unreliable and not too
interesting, as the program doesn't see the White advantage in the main line
yet.
This is after 3.Ng4-f6, again 4 ply back,which is also a nice-looking move! I
concentrated on moves for Black naturally. Here Q5T wants to play 4. ..Nc5xd3
instead of pulling back the bishop (4. ..Be7-f8 in the main line) Giving back
material and the black king has a little more space now? Out of the blue maybe
4. ..Be7xf6 is also worth looking at? I didn't investigate that much further
yet, if these are better than 4. ..Be7-f8. After 4. ..Be7-f8 I think White
certainly wins.
Last Move : 3.Nf6 (Black to play)
00:00 03.00 3.72 3..gxf6 4.exf6 Nxd3 5.Rxc6 Qxb2
00:00 04.00 4.51 3..gxf6 4.exf6 Bf8 5.Bxc6 Rxc6 6.Qh5+ Qxb2
00:01 05.00 4.84 3..gxf6 4.gxf6 Bf8 5.Rd2 Ra6 6.Bg5
00:03 06.00 4.43 3..gxf6 4.exf6 Bf8 5.Qh5 Kg8 6.Rc2 e5
00:08 07.00 4.68 3..gxf6 4.exf6 Bf8 5.Qh5 Kg8 6.Rd2 Ra7 7.Re2 Nxd3
00:26 08.00 4.22 3..gxf6 4.gxf6 Bf8 5.Qh5 Kg8 6.Rxc5 Bxc5 7.Qg5 Bg6 8.Rc1+
01:22 09.00 4.73 3..gxf6 4.gxf6 Bf8 5.Qh5 Kg8 6.Rc2 Qb3 7.Bxc6
02:53 10.00 3.76 3..gxf6 4.gxf6 Bf8 5.Rxc5 Bxc5 6.Rc1 Bf8 7.Bxc6 Ra6 8.Bb5
Rxc1+
05:30 11.00 4.26 3..gxf6 4.gxf6 Bf8 5.Rxc5 Bxc5 6.Rc1 Nxe5 7.Bxe5 Bb6 8.Re1
Ra7
25:03 12.00 3.89 3..gxf6 4.gxf6 Bf8 5.Rxc5 Bxc5 6.Qg4 Rg8 7.Qh5 Rac8 8.Be4
Rg6
34:24 13.00 2.28 3..gxf6 4.gxf6 Nxd3 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8+ Rxc8
8.Bxe5 Bf8 9.Qd1
And finally, for the starting position of Nolot #9:
[D]r4r1k/4bppb/2n1p2p/p1n1P3/1p1p1BNP/3P1NP1/qP2QPB1/2RR2K1 w - -
00:00 02.24 -0.65 1.Nxh6 gxh6 2.Bxh6 Rg8 3.h5
00:00 04.00 -0.91 1.Ne1 Qa4 2.Nxh6 gxh6 3.Bxh6 Rg8 4.Qh5 Rg6
00:00 04.01 -0.82 1.Nxh6 Rac8 2.Ne1 gxh6 3.Bxh6 Rg8 4.Ra1 Rgd8
00:03 05.00 -0.18 1.Nxh6 a4 2.Bg5 Nxd3 3.Rxd3 Bxd3 4.Qxd3 Qxb2 5.Rxc6 gxh6
6.Bxe7
00:07 06.00 -0.26 1.Nxh6 a4 2.Ng5 Bxg5 3.Rxc5 Bxf4 4.gxf4 gxh6 5.Rxc6
00:09 06.05 -0.24 1.Rxc5 Bxc5 2.Rc1 Rac8 3.Rxc5 Qb1+ 4.Bf1 a4 5.Rc1
00:13 07.00 -0.21 1.Rxc5 Bxc5 2.Rc1 Ba7 3.Rxc6 Qb1+ 4.Rc1 Qxd3 5.Qxd3 Bxd3
6.Rc7
00:22 07.01 -0.14 1.Nxh6 a4 2.Ng5 Bxg5 3.Bxg5 Nxd3 4.Bxc6 Qxb2 5.Rc2 Nf4
00:25 07.08 -0.02 1.Nf6 Rac8 2.Nxh7 Kxh7 3.Rc4 Qb3 4.h5 Rfd8
00:29 08.00 -0.14 1.Nf6 Bg6 2.Nd7 Nxd7 3.Rxc6 a4 4.Rdc1 Qb3
01:03 08.01 -0.14 1.Nxh6
01:06 08.01 0.10 1.Nxh6 a4 2.Ng5 Bxg5 3.Bxg5 Bxd3 4.Rxd3 Nxd3 5.Qxd3 Qxb2
03:04 09.00 0.36 1.Nxh6 Rac8 2.Ng5 Bxg5 3.Bxg5 Nxd3 4.Rxc6+ Rxc6 5.Bxc6 Nxb2
6.Rxd4 gxh6
05:10 10.00 0.36 1.Nxh6 Rac8 2.Ng5 Bxg5 3.Bxg5 Nxd3 4.Rxc6
21:47 11.00 0.23 1.Nxh6 Na4 2.Rd2 Rac8 3.Ng5 Bxg5 4.Bxg5 Na7 5.Rxc8 Nxc8
6.Qh5 Nxb2
39:44 12.00 0.18 1.Nxh6 Na4 2.Rd2 Rac8 3.Kh2 Bg6 4.Ne1
Sideline: after 4. ..Nc5xd3 instead of pulling back the bishop (4. ..Be7-f8 in
the main line), I'm not sure if taking with the Queen on d3 is better then
giving check with it on h5. the score is positive for White now but not yet
winning.
[Event ""]
[Site ""]
[Date ""]
[Round ""]
[White ""]
[Black ""]
[Result ""]
[FEN "r4r1k/4bppb/2n1p2p/p1n1P3/1p1p1BNP/3P1NP1/qP2QPB1/2RR2K1 w - - 0 1"]
1.Ng5 hxg5 2.hxg5 Rfc8 3.Nf6 gxf6 4.gxf6 Nxd3 *
[D]r1r4k/4bp1b/2n1pP2/p3P3/1p1p1B2/3n2P1/qP2QPB1/2RR2K1 w - -
Last Move : 4... Nxd3 (White to play)
00:00 02.01 -2.05 5.Rxd3 Bxd3 6.Qxd3 Qxb2 7.Bxc6
00:00 03.00 -1.47 5.Rxd3 Bxd3 6.Qh5+ Kg8 7.Bxc6 Rxc6 8.Rxc6
00:02 04.00 -1.73 5.Rxd3 Bxd3 6.Qh5+ Kg8 7.Bxc6 Bg6 8.Qh6
00:06 05.00 -1.86 5.Rxd3 Bxd3 6.Qxd3 Qxb2 7.Bxc6 Qxc1+ 8.Bxc1 Rxc6 9.Bh6
00:19 06.00 -2.08 5.Rxd3 Bxd3 6.Qxd3 Qxb2 7.Qd1 a4 8.Qh5 Kg8 9.Qg5+
01:07 07.00 -0.96 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8+ Rxc8 8.Bxe5 Rc1+ 9.Kh2 Bc5
10.Bxd4 Bxd4
01:56 08.00 -1.15 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8+ Rxc8 8.Bxe5 Rc1+ 9.Kh2 Bc5
10.Qe2 Qb1
02:37 09.00 -1.36 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8+ Rxc8 8.Bxe5 Bf8 9.Qd1 Bh6
07:10 10.00 -1.09 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8 Rxc8+ 8.Bxe5 Bf8 9.Qd1 Bh6
14:12 11.00 -0.74 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Qxd4 Rxc1+ 8.Bxc1 Bxf6 9.Qf4 Kg8
10.Bxa8 Qb3
14:12 12.00 -0.64 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8+ Rxc8 8.Bxe5 Bf8 9.Qd1 Bh6
27:46 13.00 -0.69 5.Rxd3 Bxd3 6.Qxd3 Nxe5 7.Rxc8+ Rxc8 8.Bxe5 Qa1+
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.