Discussion of computer chess matches and engine tournaments.
Moderator: Ras
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Thu Feb 25, 2021 11:50 am
HW: AMD Phenom II x4 (1 core per engine)
OS: Win10 64bits
TC: 10'+2"
GUI: Winboard AA
3rd Division
Code: Select all
Score Cr Sj Fe Sj Ne Sj Pu Ts Cr KK Py Py
----------------------------------------------------------------------------
1: CrazyWa 0.1.4 18.5 / 22 XX 00 10 11 11 11 11 11 1= 11 11 11 (+18 -3 =1)
2: Sjaak II 1.0 18.0 / 22 11 XX 10 10 10 01 11 11 11 11 11 11 (+18 -4 =0)
3: Feuerstein 0.461 16.0 / 22 01 01 XX 00 10 10 11 11 11 11 11 11 (+16 -6 =0)
4: Sjeng 11.2 16.0 / 22 00 01 11 XX 00 10 11 11 11 11 11 11 (+16 -6 =0)
5: Nebiyu 1.45a 14.5 / 22 00 01 01 11 XX 01 10 11 10 11 11 =1 (+14 -7 =1)
6: Sjaak II 1.3.1a 14.0 / 22 00 10 01 01 10 XX 00 11 11 11 11 11 (+14 -8 =0)
7: Pulsar 2009-9b 9.0 / 22 00 00 00 00 01 11 XX 00 11 10 10 11 (+9 -13 =0)
8: Tscp crazyhouse 1.1 8.0 / 22 00 00 00 00 00 00 11 XX 01 10 11 11 (+8 -14 =0)
9: Crazybot 6.5 / 22 0= 00 00 00 01 00 00 10 XX 01 11 01 (+6 -15 =1)
10: KKFChess 2.6.7b 6.0 / 22 00 00 00 00 00 00 01 01 10 XX 10 11 (+6 -16 =0)
11: PyChess 20181209 3.0 / 22 00 00 00 00 00 00 01 00 00 01 XX 10 (+3 -19 =0)
12: PyChess 20161204 pypy2 2.5 / 22 00 00 00 00 =0 00 00 00 10 00 01 XX (+2 -19 =1)
----------------------------------------------------------------------------
132 games: +72 -58 =2
lkaufman
Posts: 6284 Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA
Full name: Larry Kaufman
Post
by lkaufman » Fri Feb 26, 2021 4:35 am
Carlos777 wrote: ↑ Thu Feb 25, 2021 11:50 am
HW: AMD Phenom II x4 (1 core per engine)
OS: Win10 64bits
TC: 10'+2"
GUI: Winboard AA
3rd Division
Code: Select all
Score Cr Sj Fe Sj Ne Sj Pu Ts Cr KK Py Py
----------------------------------------------------------------------------
1: CrazyWa 0.1.4 18.5 / 22 XX 00 10 11 11 11 11 11 1= 11 11 11 (+18 -3 =1)
2: Sjaak II 1.0 18.0 / 22 11 XX 10 10 10 01 11 11 11 11 11 11 (+18 -4 =0)
3: Feuerstein 0.461 16.0 / 22 01 01 XX 00 10 10 11 11 11 11 11 11 (+16 -6 =0)
4: Sjeng 11.2 16.0 / 22 00 01 11 XX 00 10 11 11 11 11 11 11 (+16 -6 =0)
5: Nebiyu 1.45a 14.5 / 22 00 01 01 11 XX 01 10 11 10 11 11 =1 (+14 -7 =1)
6: Sjaak II 1.3.1a 14.0 / 22 00 10 01 01 10 XX 00 11 11 11 11 11 (+14 -8 =0)
7: Pulsar 2009-9b 9.0 / 22 00 00 00 00 01 11 XX 00 11 10 10 11 (+9 -13 =0)
8: Tscp crazyhouse 1.1 8.0 / 22 00 00 00 00 00 00 11 XX 01 10 11 11 (+8 -14 =0)
9: Crazybot 6.5 / 22 0= 00 00 00 01 00 00 10 XX 01 11 01 (+6 -15 =1)
10: KKFChess 2.6.7b 6.0 / 22 00 00 00 00 00 00 01 01 10 XX 10 11 (+6 -16 =0)
11: PyChess 20181209 3.0 / 22 00 00 00 00 00 00 01 00 00 01 XX 10 (+3 -19 =0)
12: PyChess 20161204 pypy2 2.5 / 22 00 00 00 00 =0 00 00 00 10 00 01 XX (+2 -19 =1)
----------------------------------------------------------------------------
132 games: +72 -58 =2
Interesting that not a single one of the 42 games among the top half ended in a draw. Clearly Crazyhouse does not suffer from a draw problem. But do you have the White vs. Black stats, for the whole tournament and (if possible) for the 42 games among the top half? From what I've read, the White vs. Black advantage in engine play is a huge problem especially among strong engines.
Komodo rules!
Guenther
Posts: 4718 Joined: Wed Oct 01, 2008 6:33 am
Location: Regensburg, Germany
Full name: Guenther Simon
Post
by Guenther » Fri Feb 26, 2021 9:02 am
lkaufman wrote: ↑ Fri Feb 26, 2021 4:35 am
Carlos777 wrote: ↑ Thu Feb 25, 2021 11:50 am
HW: AMD Phenom II x4 (1 core per engine)
OS: Win10 64bits
TC: 10'+2"
GUI: Winboard AA
3rd Division
Code: Select all
Score Cr Sj Fe Sj Ne Sj Pu Ts Cr KK Py Py
----------------------------------------------------------------------------
1: CrazyWa 0.1.4 18.5 / 22 XX 00 10 11 11 11 11 11 1= 11 11 11 (+18 -3 =1)
2: Sjaak II 1.0 18.0 / 22 11 XX 10 10 10 01 11 11 11 11 11 11 (+18 -4 =0)
3: Feuerstein 0.461 16.0 / 22 01 01 XX 00 10 10 11 11 11 11 11 11 (+16 -6 =0)
4: Sjeng 11.2 16.0 / 22 00 01 11 XX 00 10 11 11 11 11 11 11 (+16 -6 =0)
5: Nebiyu 1.45a 14.5 / 22 00 01 01 11 XX 01 10 11 10 11 11 =1 (+14 -7 =1)
6: Sjaak II 1.3.1a 14.0 / 22 00 10 01 01 10 XX 00 11 11 11 11 11 (+14 -8 =0)
7: Pulsar 2009-9b 9.0 / 22 00 00 00 00 01 11 XX 00 11 10 10 11 (+9 -13 =0)
8: Tscp crazyhouse 1.1 8.0 / 22 00 00 00 00 00 00 11 XX 01 10 11 11 (+8 -14 =0)
9: Crazybot 6.5 / 22 0= 00 00 00 01 00 00 10 XX 01 11 01 (+6 -15 =1)
10: KKFChess 2.6.7b 6.0 / 22 00 00 00 00 00 00 01 01 10 XX 10 11 (+6 -16 =0)
11: PyChess 20181209 3.0 / 22 00 00 00 00 00 00 01 00 00 01 XX 10 (+3 -19 =0)
12: PyChess 20161204 pypy2 2.5 / 22 00 00 00 00 =0 00 00 00 10 00 01 XX (+2 -19 =1)
----------------------------------------------------------------------------
132 games: +72 -58 =2
Interesting that not a single one of the 42 games among the top half ended in a draw. Clearly Crazyhouse does not suffer from a draw problem. But do you have the White vs. Black stats, for the whole tournament and (if possible) for the 42 games among the top half? From what I've read, the White vs. Black advantage in engine play is a huge problem especially among strong engines.
You know that there are (small) ZH books available too? There was a huge ZH boom a few years ago and Ferdinand (et al) created several tournaments and also added freely available ZH start positions for download. So I think the White advantage was reduced with start positions.
Just one of the main threads in that time
http://talkchess.com/forum3/viewtopic.p ... crazyhouse
I don't know though if Carlos had used one of those books, I guess not.
Since that time the XB/UCI chronology has a ZH section too.
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Fri Feb 26, 2021 11:04 am
lkaufman wrote: ↑ Fri Feb 26, 2021 4:35 am
Interesting that not a single one of the 42 games among the top half ended in a draw. Clearly Crazyhouse does not suffer from a draw problem. But do you have the White vs. Black stats, for the whole tournament and (if possible) for the 42 games among the top half? From what I've read, the White vs. Black advantage in engine play is a huge problem especially among strong engines.
White 72 wins
Black 58 wins
It's at the bottom. I don't have the statistics of the top engines, but here are the games:
http://www.mediafire.com/file/3qqhzkbqo ... 3.pgn/file
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Fri Feb 26, 2021 11:28 am
Guenther wrote: ↑ Fri Feb 26, 2021 9:02 am
I don't know though if Carlos had used one of those books, I guess not.
No this time, they played without books. I have my own suite (some positions I borrowed from Ferdinand's first book) that consists of 54 opening positions.
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Fri Feb 26, 2021 11:30 am
Division 2
Code: Select all
Score Su Im Su Im Su Cr Sj Cr Sj Sj
-------------------------------------------------------------
1: Sunsetter 9 15.0 / 18 XX 10 10 01 11 11 11 11 11 11 (+15 -3 =0)
2: Imortal 2.0 13.0 / 18 01 XX 01 11 10 10 11 01 11 11 (+13 -5 =0)
3: Sunsetter 7g 12.0 / 18 01 10 XX 00 11 11 00 11 11 11 (+12 -6 =0)
4: Imortal 1.0 10.0 / 18 10 00 11 XX 01 00 10 01 11 11 (+10 -8 =0)
5: Sunsetter 7e 10.0 / 18 00 01 00 10 XX 00 11 11 11 11 (+10 -8 =0)
6: CrazyWa 1.0.1 9.0 / 18 00 01 00 11 11 XX 10 00 11 10 (+9 -9 =0)
7: Sjeng 11.3 7.0 / 18 00 00 11 01 00 01 XX 11 00 10 (+7 -11 =0)
8: CrazyWa 0.1.4 7.0 / 18 00 10 00 10 00 11 00 XX 11 10 (+7 -11 =0)
9: SjaakII 1.4.1 4.0 / 18 00 00 00 00 00 00 11 00 XX 11 (+4 -14 =0)
10: Sjaak II 1.0 3.0 / 18 00 00 00 00 00 01 01 01 00 XX (+3 -15 =0)
-------------------------------------------------------------
90 games: +55 -35 =0
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Sat Feb 27, 2021 8:47 pm
Division 1
Code: Select all
Score Fa St Fa Fa Su Im Im St Im TJ
-------------------------------------------------------------------------
1: FairyStockfish 8x8 120221 17.0 / 18 XX 11 11 11 10 11 11 11 11 11 (+17 -1 =0)
2: Stockfish 020717 15.5 / 18 00 XX =1 11 11 11 11 11 11 11 (+15 -2 =1)
3: FairyStockfish LB 120221 12.5 / 18 00 =0 XX 01 11 11 10 11 11 11 (+12 -5 =1)
4: FairyStockfish LB 11.2 12.0 / 18 00 00 10 XX 11 11 10 11 11 11 (+12 -6 =0)
5: Sunsetter 9 8.0 / 18 01 00 00 00 XX 10 11 01 01 11 (+8 -10 =0)
6: Imortal 3.04 7.5 / 18 00 00 00 00 01 XX 11 11 01 1= (+7 -10 =1)
7: Imortal 2.9001 7.0 / 18 00 00 01 01 00 00 XX 10 11 11 (+7 -11 =0)
8: Stockfish 021116 5.0 / 18 00 00 00 00 10 00 01 XX 11 10 (+5 -13 =0)
9: Imortal 2.0 4.0 / 18 00 00 00 00 10 10 00 00 XX 11 (+4 -14 =0)
10: TJchess 1.3 1.5 / 18 00 00 00 00 00 0= 00 01 00 XX (+1 -16 =1)
-------------------------------------------------------------------------
90 games: +55 -33 =2
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Mon Mar 01, 2021 3:13 am
Premier
2 openings were used: 1. e4 and 1. d4
Code: Select all
Score Stoc Stoc Stoc Stoc Stoc Fair Stoc Stoc Fair
----------------------------------------------------------------------------------------
1: Stockfish 081018 20.5 / 32 XXXX 0001 0011 1001 1011 0011 1111 11=0 1111 (+20 -11 =1)
2: Stockfish 260719 19.5 / 32 1110 XXXX 0010 0101 1100 0111 0101 111= 1110 (+19 -12 =1)
3: Stockfish 130620 18.5 / 32 1100 1101 XXXX 0001 01=1 1000 0101 1111 1011 (+18 -13 =1)
4: Stockfish 210418 18.0 / 32 0110 1010 1110 XXXX 01=0 1011 010= 0011 0111 (+17 -13 =2)
5: Stockfish 160221 18.0 / 32 0100 0011 10=0 10=1 XXXX 1110 1011 0111 1001 (+17 -13 =2)
6: FairyStockfish 8x8 120221 17.0 / 32 1100 1000 0111 0100 0001 XXXX 0111 0101 1111 (+17 -15 =0)
7: Stockfish 020717 12.0 / 32 0000 1010 1010 101= 0100 1000 XXXX 0101 =010 (+11 -19 =2)
8: Stockfish 310720 12.0 / 32 00=1 000= 0000 1100 1000 1010 1010 XXXX 0111 (+11 -19 =2)
9: FairyStockfish 8x8 11.2 8.5 / 32 0000 0001 0100 1000 0110 0000 =101 1000 XXXX (+8 -23 =1)
----------------------------------------------------------------------------------------
144 games: +91 -47 =6
lkaufman
Posts: 6284 Joined: Sun Jan 10, 2010 6:15 am
Location: Maryland USA
Full name: Larry Kaufman
Post
by lkaufman » Mon Mar 01, 2021 6:42 pm
Carlos777 wrote: ↑ Mon Mar 01, 2021 3:13 am
Premier
2 openings were used: 1. e4 and 1. d4
Code: Select all
Score Stoc Stoc Stoc Stoc Stoc Fair Stoc Stoc Fair
----------------------------------------------------------------------------------------
1: Stockfish 081018 20.5 / 32 XXXX 0001 0011 1001 1011 0011 1111 11=0 1111 (+20 -11 =1)
2: Stockfish 260719 19.5 / 32 1110 XXXX 0010 0101 1100 0111 0101 111= 1110 (+19 -12 =1)
3: Stockfish 130620 18.5 / 32 1100 1101 XXXX 0001 01=1 1000 0101 1111 1011 (+18 -13 =1)
4: Stockfish 210418 18.0 / 32 0110 1010 1110 XXXX 01=0 1011 010= 0011 0111 (+17 -13 =2)
5: Stockfish 160221 18.0 / 32 0100 0011 10=0 10=1 XXXX 1110 1011 0111 1001 (+17 -13 =2)
6: FairyStockfish 8x8 120221 17.0 / 32 1100 1000 0111 0100 0001 XXXX 0111 0101 1111 (+17 -15 =0)
7: Stockfish 020717 12.0 / 32 0000 1010 1010 101= 0100 1000 XXXX 0101 =010 (+11 -19 =2)
8: Stockfish 310720 12.0 / 32 00=1 000= 0000 1100 1000 1010 1010 XXXX 0111 (+11 -19 =2)
9: FairyStockfish 8x8 11.2 8.5 / 32 0000 0001 0100 1000 0110 0000 =101 1000 XXXX (+8 -23 =1)
----------------------------------------------------------------------------------------
144 games: +91 -47 =6
Was the White to Black win ratio much different between the 1e4 games and the 1.d4 games?
Komodo rules!
Carlos777
Posts: 1977 Joined: Sun Dec 13, 2009 6:09 pm
Post
by Carlos777 » Tue Mar 02, 2021 12:20 pm
lkaufman wrote: ↑ Mon Mar 01, 2021 6:42 pm
Was the White to Black win ratio much different between the 1e4 games and the 1.d4 games?
Yes.
1. e4 => 72.9% for white
1. d4 => 57.6% for white