Discussion of anything and everything relating to chess playing software and machines.
Moderators: hgm , Rebel , chrisw
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Sun Mar 24, 2019 10:37 am
M ANSARI wrote: ↑ Tue Mar 19, 2019 8:30 am
I guess if anything ... this shows the importance of bug free code in a chess program. I think the first to show that was Fruit. While it is nice to have many features and many minor little tweaks to get the program to play "proper" chess ... bugs are the Achilles Heel of any chess program. Each additional improvement needs to come without letting any bugs creep in.
Good job on your program and I must say it is very interesting to follow up on the progress!
Thanks for the comment ! although I am not sure at all Minic is bug free ... To be honest, just after you posted this message, I discovered that null-move was in fact not activated ... This some work to do !
Best regards
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Sat Mar 30, 2019 8:40 am
While Minic 0.45 is not released yet, here is a Minic no-book depth 30 opening line. What do you think of it ? the queen exchange seems to weaken white side
[pgn]
[Event "Import"]
[Site "https://lichess.org/jq3Ov8n1"]
[Date "2019.03.30"]
[Round "-"]
[White "?"]
[Black "?"]
[Result "*"]
[UTCDate "2019.03.30"]
[UTCTime "07:35:25"]
[WhiteElo "?"]
[BlackElo "?"]
[Variant "Standard"]
[TimeControl "-"]
[ECO "D00"]
[Opening "Queen's Pawn Game: Mason Attack"]
[Termination "Unterminated"]
[Annotator "lichess.org"]
1. d4 d5 2. Bf4 { D00 Queen's Pawn Game: Mason Attack } Bf5 3. e3 e6 4. Bd3 Bxd3 5. Qxd3 Ne7 6. Nc3 Nbc6 7. Nf3 Ng6 8. Bg3 Bd6 9. O-O O-O 10. Nb5 Bxg3 11. hxg3 a6 12. Nc3 Qf6 13. e4 Rad8 14. e5 Qf5 15. a3 Qxd3 16. cxd3 Rfe8 17. Rfe1 *
[/pgn]
abulmo2
Posts: 433 Joined: Fri Dec 16, 2016 11:04 am
Location: France
Full name: Richard Delorme
Post
by abulmo2 » Sat Mar 30, 2019 11:14 am
xr_a_y wrote: ↑ Sat Mar 30, 2019 8:40 am
While Minic 0.45 is not released yet, here is a Minic no-book depth 30 opening line. What do you think of it ? the queen exchange seems to weaken white side
[pgn]
[Event "Import"]
[Site "https://lichess.org/jq3Ov8n1"]
[Date "2019.03.30"]
[Round "-"]
[White "?"]
[Black "?"]
[Result "*"]
[UTCDate "2019.03.30"]
[UTCTime "07:35:25"]
[WhiteElo "?"]
[BlackElo "?"]
[Variant "Standard"]
[TimeControl "-"]
[ECO "D00"]
[Opening "Queen's Pawn Game: Mason Attack"]
[Termination "Unterminated"]
[Annotator "lichess.org"]
1. d4 d5 2. Bf4 { D00 Queen's Pawn Game: Mason Attack } Bf5 3. e3 e6 4. Bd3 Bxd3 5. Qxd3 Ne7 6. Nc3 Nbc6 7. Nf3 Ng6 8. Bg3 Bd6 9. O-O O-O 10. Nb5 Bxg3 11. hxg3 a6 12. Nc3 Qf6 13. e4 Rad8 14. e5 Qf5 15. a3 Qxd3 16. cxd3 Rfe8 17. Rfe1 *
[/pgn]
I do not think much by myself, but 15.a3 look like a weak move; and anticipating the exchange with 15.Qxf5 exf5 looks better. Here is a retro-analysis by Stockfish 10 of this opening, which seems to disagree on almost every move:
Code: Select all
$ postmortem -f pgn/minic-0.45-opening.pgn -c 4 -d 30 -w 80 -e stockfish-10 -H 1024
Stockfish 10 64 POPCNT used for analysis
reading pgn/minic-0.45-opening.pgn
read 1 games & 33 moves in 0.0004698 s.
game: ? vs ? - 2019.03.30
Played Move | Best Alternative
ply move score | score pv
----------------------+--------------------
17. Rfe1? -0.57 < +0.07 17.Ne2 h5 18.Nd2 Nge7 19.Rfc1 a5 20.f3 a4
21.Kf2 Nf5 22.g4 hxg4 23.fxg4 Nfxd4 24.Nxd4
Nxd4 25.Rxc7 Rc8 26.Rac1 Nc6 27.Rxc8 Rxc8
28.Nf3 Ra8 29.Rb1 Ra5 30.b4 axb3 31.Rxb3 Kf8
32.Rxb7 Rxa3 33.Ke3 Ke8 34.g5 Ra2 35.g3 Ra3
36.Rc7 Nb4 37.Ne1
Rfe8? -0.07 < +0.56 16... h5 17.b4 Nge7 18.Ne2 Nf5 19.Rac1 Rb8
20.Rc2 Kh7 21.Kh2 g6 22.Rfc1 Kg7 23.Kh3 Rbc8
24.Kh2 Ra8 25.Kh3 Rh8 26.Ng5 Rab8 27.Kh2 Rhf8
28.Nf3 Rfe8 29.Kh3 Rbc8 30.Kh2 Kh6 31.Kh3
16. cxd3 -0.56 > -13.49 16.Rac1 Qc4 17.Rfd1 a5 18.Rd3 a4 19.Kh2 h5
20.Rb1 Nge7 21.Rd2 Nf5 22.Ne2 g6 23.Kh3 Na5
24.Nf4 c5 25.dxc5 Qxc5 26.Nd3 Qc4 27.Kh2 Kg7
28.Re1 Qb5 29.Rdd1 Nc4 30.Ra1 Ra8 31.Ra2 Qc6
32.Rb1 Rad8 33.Kg1
Qxd3 +0.56 > -0.17 15... h6 16.Qxf5 exf5 17.Rfe1 Rfe8 18.Nd1 Nf8
19.Ne3 g6 20.b4 Ne6 21.c3 Kg7 22.Rab1 Ne7
23.Kf1 c6 24.Ke2 Nc8 25.a4 f6 26.Kd2 h5 27.Kd3
fxe5 28.dxe5 Nb6 29.a5 Nd7 30.Nc2 f4 31.Ncd4
fxg3
15. a3?? -0.56 < +0.59 15.Qxf5 exf5 16.Ne2 f6 17.exf6 Rxf6 18.Nc1 f4
19.g4 Re8 20.Nd3 Rfe6 21.c3 h6 22.Nc5 Re2
23.Rfe1 a5 24.Kf1 Rxe1+ 25.Rxe1 Rxe1+ 26.Kxe1
Nd8 27.Kd2 Kf7 28.b4 b6 29.Nd3 axb4 30.cxb4 b5
31.Nc5 Ne6 32.Kc3 Ke7 33.Nd2
Qf5? -0.59 < +0.00 14... Qe7 15.a3 Nb8 16.b4 Nd7 17.Nd1 c6 18.Re1
Ra8 19.c3 a5 20.Ne3 Rfc8 21.Nd2 b6 22.f4 axb4
23.axb4 c5 24.f5 Ngf8 25.Rxa8 Rxa8 26.fxe6 fxe6
27.dxc5 bxc5 28.Nxd5 Qd8 29.Nf4 cxb4 30.cxb4
14. e5? +0.00 < +0.68 14.Rfe1 h6 15.Rad1 Nb8 16.Ne2 dxe4 17.Qxe4 c6
18.Nf4 Nxf4 19.gxf4 Nd7 20.g3 Qe7 21.c3 Nf6
22.Qe2 Qc7 23.a3 c5 24.dxc5 Qxc5 25.Ne5 Qc7
26.Kh2 b5 27.a4 bxa4 28.Qxa6 Ra8 29.Qe2 a3
30.bxa3 Rxa3
Rad8? -0.68 < -0.24 13... Nb4 14.Qd2 Qe7 15.Rad1 Rfd8 16.Rfe1 c6
17.a3 dxe4 18.Nxe4 Nd5 19.g4 Nf6 20.Nxf6+ Qxf6
21.g5 Qf5 22.Qd3 Qg4 23.Qe4 Qxe4 24.Rxe4 c5
25.c3 Rac8 26.Kh2 Rd5 27.g3 b5 28.Kg2 a5 29.Rc1
Ne7 30.Rh1 Rcd8 31.b3 c4 32.bxc4 bxc4
13. e4? +0.24 < +0.64 13.Rfe1 Rfd8 14.e4 Nge7 15.exd5 exd5 16.Rad1
Qg6 17.Qd2 Qf6 18.Ne2 Ng6 19.c3 Re8 20.Qc2 b6
21.g4 Qe6 22.Qf5 a5 23.Kf1 Qxf5 24.gxf5 Nge7
25.Nf4 a4 26.Ne5 Nxe5 27.dxe5 Nxf5 28.Nxd5
Qf6? -0.64 < +0.00 12... Nce7 13.e4 c6 14.Na4 Qc7 15.Rfe1 Rfd8
16.c3 b6 17.b3 h6 18.Rac1 dxe4 19.Qxe4 Nd5
20.c4 Nf6 21.Qe2 Rab8 22.Rcd1 Ne7 23.Qe5 Qb7
24.Qf4 Ng6 25.Qc1 Qc7 26.Qe3 Ne7
12. Nc3 +0.00 = +0.00 12.Na3 e5 13.dxe5 Ngxe5 14.Qb3 Nxf3+ 15.gxf3 d4
16.Rad1 Qf6 17.e4 b5 18.Kg2 Rfd8 19.f4 Na5
20.e5 Nxb3 21.exf6 Nc5 22.fxg7 Na4 23.Rb1 Kxg7
24.Rh1 Rd5 25.Rhe1 Rdd8
a6 +0.00 = +0.00 11... Nb4 12.Qe2 c6 13.Nc3 Na6 14.a4 h6 15.e4
Rc8 16.Rfe1 Qd7 17.Qd2 Rfd8 18.e5 Qe7 19.Ra3
Nb4 20.a5 Kh7 21.Rb3 Na6 22.Ra3 Nb4
11. hxg3 +0.00 = +0.00 11.fxg3 Qe7 12.Rad1 a6 13.Nc3 Rad8 14.e4 dxe4
15.Nxe4 e5 16.c3 exd4 17.cxd4 Rfe8 18.Rfe1 Qb4
19.Qc3 Qb6 20.Nc5 h6 21.Kf2 Rxe1 22.Rxe1 Qa7
23.g4 Nge7 24.Qc4 Na5 25.Qa4 Nec6 26.Qa3 b6
27.Nd3 Nxd4 28.Nxd4
Bxg3 +0.00 = +0.00 10... Nb4 11.Qd2 Bxg3 12.hxg3 Nc6 13.Qd3 Nb4
10. Nb5? +0.00 < +0.60 10.e4 Bxg3 11.hxg3 Nb4 12.Qe2 c6 13.Rfe1 Na6
14.a3 Nb8 15.g4 Nd7 16.g5 b5 17.g3 a5 18.Kg2
Qc7 19.exd5 exd5 20.Nb1 Rfd8 21.Nbd2 a4 22.Qe3
Nb6 23.Ne5 Nxe5 24.Qxe5 Qxe5 25.Rxe5
O-O? -0.60 < -0.05 9... Bxg3 10.fxg3 Qd7 11.e4 f6 12.Rad1 O-O-O
13.a3 Nge7 14.g4 Kb8 15.g5 Ng6 16.b4 Rhf8
17.gxf6 gxf6 18.Qd2 Nce7 19.Rfe1 a6 20.exd5
exd5 21.a4 Nf5 22.b5 Nd6 23.Ra1 Nc4 24.Qc1 Rfe8
25.bxa6 Rxe1+ 26.Qxe1 bxa6 27.Rb1+ Ka7
9. O-O? +0.05 < +0.78 9.a3 h5 10.O-O h4 11.Bxd6 cxd6 12.h3 O-O
13.Rfe1 Rc8 14.Rac1 Re8 15.a4 a6 16.Ne2 Qd7
17.Qb3 b5 18.axb5 axb5 19.Nf4 Nxf4 20.exf4 Qd8
21.c3 b4 22.f5 bxc3 23.fxe6 fxe6 24.bxc3 Qf6
25.Qc2
Bd6? -0.78 < -0.28 8... Qd7 9.O-O Bd6 10.Ne2 O-O 11.c4 Nce7
12.Rac1 Nf5 13.Bxd6 Nxd6 14.c5 Ne4 15.Nc3 Nf6
16.Ne5 Qe7 17.f4 c6 18.b4 Rac8 19.a4 Nd7
20.Nxd7 Qxd7 21.b5 Rfe8 22.Qc2 h6 23.Rb1 Ne7
8. Bg3 +0.28 ~ +0.18 8.g3 Bb4 9.O-O-O Bxc3 10.Qxc3 Qd7 11.h4 h5
12.Kb1 f6 13.Qb3 O-O-O 14.Ne1 b6 15.Nd3 Na5
16.Qb4 Nc4 17.b3 a5 18.Qe1 Nxf4 19.Nxf4 Nd6
20.a4 Kb7 21.f3 g5 22.hxg5 fxg5 23.Nxh5 Qf7
Ng6 -0.28 ~ -0.19 7... a6 8.O-O Qd7 9.a3 Ng6 10.Bg3 Bd6 11.Rfe1
Bxg3 12.hxg3 Nce7 13.Rad1 c6 14.Ne2 O-O 15.Nf4
h6 16.c4 b5 17.cxd5 cxd5 18.g4 Qd6 19.Nxg6 Nxg6
20.Rc1 Rac8 21.g5 hxg5 22.Nxg5 e5 23.dxe5 Qxe5
7. Nf3? +0.19 < +0.60 7.O-O-O a6 8.h4 Qd7 9.h5 h6 10.Kb1 O-O-O 11.Na4
Nf5 12.Ne2 Na5 13.Nec3 Qc6 14.g4 Nd6 15.b3 b5
16.Bxd6 Qxd6 17.Nc5 Qc6 18.e4 dxe4 19.N3xe4 Nb7
20.b4 Nxc5 21.bxc5 Kb7 22.Qf3 Be7 23.Qxf7 Qxe4
24.Qxe7 Rxd4
Nbc6? -0.60 < +0.09 6... Ng6 7.Bg3 c6 8.Nf3 Be7 9.O-O O-O 10.Ne2
Nd7 11.c4 Nf6 12.Nc3 Re8 13.Rfd1 Bb4 14.Ne2 Ne4
15.Rac1 Ba5 16.a3 Bc7 17.Qc2 h6 18.Bxc7 Qxc7
19.Nc3 Nf6 20.Qd3 Rad8 21.c5
6. Nc3? -0.09 < +0.76 6.Nf3 Nbc6 7.O-O Ng6 8.Bg3 h5 9.h3 Bd6 10.Bxd6
cxd6 11.Nbd2 O-O 12.c4 dxc4 13.Qxc4 Rc8 14.Qe2
Nge7 15.Rfc1 d5 16.Nb3 Nf5 17.Rc2 Qe7 18.Rac1
f6 19.Nfd2 g6 20.a3 Rfd8 21.Nf3 e5
Ne7? -0.76 < +0.21 5... c5 6.Nf3 cxd4 7.exd4 Bd6 8.Bxd6 Qxd6 9.c3
Nd7 10.Nbd2 Ne7 11.a4 Rc8 12.O-O O-O 13.Rfe1 h6
14.Rac1 a6 15.h3 Nf6 16.Nb3 Qb6 17.Nc5 Nf5
18.Qe2 Nd6 19.Ne5 Qc7 20.Ncd3 Nfe4 21.Kh1
5. Qxd3? -0.21 < +0.29 5.cxd3 Bd6 6.Bxd6 cxd6 7.Nc3 Nc6 8.a3 Nf6 9.Nf3
O-O 10.O-O a6 11.b4 b5 12.Qb3 Qb6 13.Qb2 h6
14.h3 Rfc8 15.Rfc1 Rc7 16.Ne2 Ne7 17.Rxc7 Qxc7
18.a4 bxa4 19.Rxa4 Qc6 20.Ra1
Bxd3? -0.29 < +0.00 4... Bb4+ 5.c3 Bxd3 6.Qxd3 Bd6 7.Ne2 Nf6 8.O-O
c5 9.dxc5 Bxc5 10.c4 Nc6 11.a3 dxc4 12.Qxc4 Be7
13.Nbc3 Qb6 14.Na4 Qa5 15.Nac3 Qb6
4. Bd3? +0.00 < +0.40 4.Nf3 Nf6 5.Bd3 Bd6 6.Bxf5 exf5 7.O-O Bxf4
8.exf4 O-O 9.c3 c6 10.Re1 Nbd7 11.Nbd2 Ne4
12.Ne5 Qc7 13.Ndf3 Rfe8 14.Qc2 Ndf6 15.a4 g6
16.a5 a6 17.b4 Kg7 18.Nd3 Qd6 19.h3 Rad8
e6? -0.40 < +0.00 3... Nf6 4.Bd3 Bxd3 5.cxd3 Nbd7 6.Nc3 c6 7.e4
e6 8.Nf3 h6 9.O-O Bb4 10.Qc2 Nf8 11.Be5 N8d7
3. e3? +0.00 < +0.62 3.c4 dxc4 4.Nc3 Nf6 5.e3 e6 6.Bxc4 Nbd7 7.Nf3
Bd6 8.Ne5 O-O 9.O-O Nb6 10.Bd3 Nbd5 11.Nxd5
Nxd5 12.Bg3 Bxe5 13.Bxe5 Bxd3 14.Qxd3 f6 15.Bg3
f5 16.Rfd1 Qd7 17.Rac1 c6 18.h3 Rad8 19.Qe2 Qe7
Bf5? -0.62 < +0.12 2... c5 3.e3 Nc6 4.Nf3 Nf6 5.Nbd2 e6 6.c3 Bd6
7.Bxd6 Qxd6 8.Bb5 O-O 9.O-O b6 10.a4 Bb7
11.Bxc6 Bxc6 12.Ne5 Nd7 13.Ndf3 f6 14.Nxc6 Qxc6
15.b4 cxb4 16.cxb4 Rac8 17.a5 Rfd8 18.Qe2 e5
19.axb6 Qxb6 20.Rfb1 e4
2. Bf4? -0.12 < +0.58 2.Nf3 Nf6 3.c4 e6 4.Nc3 c6 5.e3 Nbd7 6.Bd3 dxc4
7.Bxc4 b5 8.Be2 Bb7 9.O-O Be7 10.e4 b4 11.e5
bxc3 12.exf6 Bxf6 13.bxc3 c5 14.Rb1 Bd5 15.c4
Be4 16.Bd3 Bxd3 17.Qxd3 cxd4 18.Ba3 Qa5
d5? -0.58 < +0.00 1... e6 2.c4 Nf6 3.Nf3 b6 4.g3 Ba6 5.b3 d5
6.Bg2 Bb4+ 7.Bd2 Be7 8.O-O Nbd7 9.Ne5 Bb7
10.Nc3 c5 11.Bf4 O-O 12.dxc5 Nxc5 13.Qc2 Ncd7
14.Nxd7 Qxd7 15.Rad1 Rad8 16.cxd5 exd5 17.e3
Rc8
1. d4? +0.00 < +0.37 1.Nf3 e6 2.e3 Nf6 3.Be2 b6 4.b3 d5 5.c4 Bb7
6.O-O Bd6 7.cxd5 exd5 8.Nd4 O-O 9.Nf5 c5
10.Nxd6 Qxd6 11.Bb2 Nc6 12.d4 cxd4 13.Bxd4 Nxd4
14.Qxd4 Rfe8 15.Nc3 Ne4 16.Nxe4 Rxe4
Richard Delorme
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Tue Apr 02, 2019 8:01 pm
Code: Select all
Score of Minic vs Topple_0.4.0_nehalem: 172 - 271 - 130 [0.414]
Elo difference: -60.64 +/- 25.29
573 of 1000 games finished.
On the road to 2700elo ... will start to work on evaluation soon ...
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Thu Apr 04, 2019 7:04 am
Good progress since 0.43 : +60elo
MikeB
Posts: 4889 Joined: Thu Mar 09, 2006 6:34 am
Location: Pen Argyl, Pennsylvania
Post
by MikeB » Fri Apr 05, 2019 6:16 pm
xr_a_y wrote: ↑ Tue Apr 02, 2019 8:01 pm
Code: Select all
Score of Minic vs Topple_0.4.0_nehalem: 172 - 271 - 130 [0.414]
Elo difference: -60.64 +/- 25.29
573 of 1000 games finished.
On the road to 2700elo ... will start to work on evaluation soon ...
Still under 4000 lines - that's impressive for the strength it has. Nice work!
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Fri Apr 05, 2019 6:29 pm
Thanks ! I'd really like to get under 3000slocs ...
Here's some progress table
Code: Select all
Rank Name Elo +/- Games Score Draws
1 minic_dev_linux_x64 142 16 1280 69.4% 37.0%
2 minic_0.46_linux_x64_avx2_bmi2 132 15 1280 68.1% 38.4%
3 minic_0.45_linux_x64_avx2_bmi2 109 15 1280 65.2% 40.5%
4 minic_0.43_linux_x64_see4.2 80 14 1280 61.3% 42.8%
5 minic_0.40_linux_x64_see4.2 58 14 1281 58.3% 45.5%
6 minic_0.37_linux_x64_see4.2 8 14 1281 51.1% 48.7%
7 minic_0.34_linux_x64_see4.2 -5 14 1280 49.2% 45.2%
8 minic_0.30_linux_x64_see4.2 -11 14 1280 48.5% 42.7%
9 minic_0.32_linux_x64_see4.2 -16 14 1282 47.7% 46.0%
10 minic_0.39_linux_x64_see4.2 -20 14 1280 47.1% 44.4%
11 minic_0.28_linux_x64_see4.2 -145 16 1281 30.2% 31.9%
12 minic_0.26_linux_x64_see4.2 -169 17 1280 27.4% 31.6%
13 minic_0.20_linux_x64_see4.2 -178 18 1281 26.5% 25.8%
8323 of 78000 games finished.
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Fri Apr 19, 2019 7:25 am
While I am now working on evaluation (mostly king safety and pawn structure) I ran a 40/3 match Minic 0.47 against Fruit 2.1. Here is the results, Minic is just behind.
xr_a_y
Posts: 1871 Joined: Sat Nov 25, 2017 2:28 pm
Location: France
Post
by xr_a_y » Sun Apr 28, 2019 10:36 am
Minic 0.50 is now available. Should be 2700 on CCRL scale.
I started to add some evaluation features, I was hoping for some easy elo but that is not the case at all ! I have to be very careful and test and tune each new feature. For now I am only getting a +50 from release 0.47.
Other new features are :
* new books based on Carlen and Anand style (thanks to Jonathan Cremers for his usefull opening lines). More to come.
* Windows 32 bits executables. You can try them, but for some reason I don't get yet there are 200elo weaker (and very slow) ... So please never "test" Minic using those, only for playing purpose.
* New executables for both Linux and Windows without any optimization activated (no sse, no avx, no bmi)
I will now work on rook eval, end-game eval, some more pawn eval features, test the syzygy fathom, try to reduce the number of lines, and try again texel tuning (until now I never succeeded)
Werner Taelemans
Posts: 119 Joined: Mon Feb 03, 2014 11:57 am
Location: Belgium
Full name: Werner Taelemans
Post
by Werner Taelemans » Fri May 03, 2019 2:25 pm
xr_a_y wrote: ↑ Sun Apr 28, 2019 10:36 am
Minic 0.50 is now available. Should be 2700 on CCRL scale.
Thanks Vivien.
Although one remark: Minic doesn't respond to the 'protover 2' command. My own GUI (galjoengui) tries to autodetect the protocol and thinks that Minic follows the 'xboard version 1' protocol. And then, nothing works of course...