Auutomated play with the Novag Citrine

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

Moderators: hgm, Rebel, chrisw

User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Auutomated play with the Novag Citrine

Post by sje »

I've been experimenting with automated play between Symbolic and a Novag Citrine as a prelude to connecting a Citrine to a Banjo ghess server. In a fifty round, five minutes per game test match, the score was: Symbolic forty nine, Citrine one. Considering that the Citrine's 20 MHz CPU chip can be had for US$5 vs several hundred US$ for a 2.66 GHz Xeon Woodcrest, this is actually not so bad for the Citrine.

Here's the Citrine's victory:

Code: Select all

[Event "Symbolic vs Citrine test match"]
[Site "gail"]
[Date "2009.07.27"]
[Round "44"]
[White "Novag Citrine"]
[Black "Symbolic 2009.07.27"]
[Result "1-0"]

1 d4 Nf6 2 c4 g6 3 Nc3 Bg7 4 e4 d6 5 f4 O-O 6 Nf3 c5 7 d5 e6 8 dxe6 fxe6 9 Be2
Nc6 10 O-O b6 11 Qd2 Bb7 12 Rd1 Nd4 13 Qd3 Qe7 14 Bd2 Ng4 15 Rf1 Nxe2+ 16 Qxe2
Bxc3 17 bxc3 e5 18 Ng5 Nf6 19 f5 h6 20 Ne6 Rfc8 21 Bxh6 Bxe4 22 fxg6 Qxe6 23
Rxf6 Qxf6 24 Qxe4 Rab8 25 Rf1 Qe6 26 Rf7 b5 27 Rg7+ Kh8 28 Rh7+ Kg8 29 Rg7+ Kf8
30 Qf3+ Ke8 31 Rh7 Qg8 32 Qf7+ Qxf7 33 gxf7+ Kd7 34 f8=Q+ Kc6 35 cxb5+ Kxb5 36
Qxd6 Rd8 37 Qxe5 a6 38 Rf7 Rdc8 39 Qe2+ Ka5 40 Bf4 Rb1+ 41 Kf2 Rc6 42 Bc7+ Rxc7
43 Rxc7 Kb6 44 Rc8 Kb7 45 Qe4+ Kxc8 46 Qxb1 Kc7 47 Qg6 a5 48 Qa6 a4 49 Qxa4 Kd6
50 Qc4 Kc6 51 Qe4+ Kd6 52 a4 Kc7 53 Qd5 Kc8 54 Qxc5+ Kd7 55 Qd4+ Ke6 56 Ke3 Kf7
57 Qa7+ Ke6 58 Ke4 Kd6 59 Qd4+ Kc6 60 a5 Kc7 61 Qc5+ Kd7 62 Qd5+ Ke8 63 Qc6+
Ke7 64 a6 Kf7 65 Qd5+ Kg6 66 a7 Kh7 67 a8=Q 1-0
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Match games uploaded

Post by sje »

For the PGN of the 50 game match, see:

http://idisk.mac.com/chessnotation-Public?view=web

PGN -> Citrine -> Blitz1.pgn

Symbolic's fastest win:

Code: Select all

[Event "Symbolic vs Citrine test match"]
[Site "gail"]
[Date "2009.07.27"]
[Round "5"]
[White "Symbolic 2009.07.27"]
[Black "Novag Citrine"]
[Result "1-0"]

1 e4 e5 2 Nf3 Nc6 3 Bb5 a6 4 Ba4 Nf6 5 O-O Nxe4 6 Re1 Nc5 7 Nxe5 Be7 8 Nc3 Nxe5
9 Rxe5 Nxa4 10 Nd5 O-O 11 Nxe7+ Kh8 12 Qh5 h6 13 d3 Kh7 14 Qf5+ Kh8 15 Bxh6 Rg8
16 Qh3 d6 17 Bc1+ Bxh3 18 Rh5# 1-0
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

A repetition draw

Post by sje »

In a second batch of five minute games, the Citrine has managed a repetition draw in one of them:

Code: Select all

[Event "Symbolic vs Citrine test match"]
[Site "gail"]
[Date "2009.07.27"]
[Round "28"]
[White "Novag Citrine"]
[Black "Symbolic 2009.07.27"]
[Result "1/2-1/2"]

1 e4 c5 2 Nf3 d6 3 d4 cxd4 4 Nxd4 Nf6 5 Nc3 a6 6 Bg5 Nbd7 7 Bc4 h6 8 Be3 Ne5 9
Be2 Neg4 10 Bc1 Qb6 11 O-O Bd7 12 a3 Rc8 13 Nd5 Nxd5 14 Bxg4 Nf6 15 Bxd7+ Nxd7
16 Rb1 e5 17 Nf5 Qc6 18 f3 Qxc2 19 Be3 Qxd1 20 Rfxd1 Rc6 21 Rbc1 Rxc1 22 Bxc1
Kd8 23 Be3 Kc7 24 Rc1+ Kb8 25 Rd1 Kc7 26 Rc1+ Kb8 27 Rd1 Kc7 1/2-1/2
The interface code to the Citrine seems to be working better than before, so I've increased the match length to one hundred games. I'll post them as they become available.

Sometimes the Citrine will resign a lost game, but other times it will play until it's checkmated. I'm not sure why this is the case, but it does complicate the interface slightly. Perhaps there is special anti-human swindle code that assumes that a human opponent might not see a tricky mate.
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Another repetition draw

Post by sje »

A bit longer this time:

Code: Select all

[Event "Symbolic vs Citrine test match"]
[Site "gail"]
[Date "2009.07.28"]
[Round "71"]
[White "Symbolic 2009.07.27"]
[Black "Novag Citrine"]
[Result "1/2-1/2"]

1 d4 Nf6 2 c4 e6 3 Nc3 Bb4 4 Qc2 O-O 5 a3 Bxc3+ 6 Qxc3 d6 7 Nf3 Ne4 8 Qc2 Nf6 9
g3 Bd7 10 Bg2 Bc6 11 O-O Be4 12 Qb3 Nc6 13 Bg5 Bxf3 14 exf3 Nxd4 15 Qxb7 Ne2+
16 Kh1 Rb8 17 Qxa7 Rxb2 18 Bxf6 gxf6 19 Qe3 c5 20 Rab1 Qb6 21 Qd3 Nd4 22 Qe4 f5
23 Rxb2 Qxb2 24 Qf4 Qb6 25 Qg5+ Kh8 26 Re1 Ra8 27 Qf6+ Kg8 28 Qg5+ Kf8 29 Qh6+
Ke8 30 Qxh7 Rxa3 31 Qh8+ Ke7 32 Qh4+ Ke8 33 Qh8+ Ke7 34 Qh4+ Ke8 35 Qh8+
1/2-1/2
Why does Symbolic allow avoidable draws with players ranked far lower than itself?

1) I am too lazy to implement a contempt factor and a nonzero draw value.

2) I think that a contempt factor and a nonzero draw, while good in some cases for better score results, somehow negatively affects the basic symmetry of the game. As the saying goes, "Play the board, not the man".
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: Match games uploaded

Post by sje »

For the second batch:

PGN-> Citrine -> Blitz2.pgn

Score: Symbolic 98 wins, 0 losses, 2 draws
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Re: Match games uploaded

Post by sje »

Ten more games, this time at 15 minutes/game:

See: PGN -> Citrine -> g15min.pgn

Score: Symbolic 10 wins, 0 losses, 0 draws

A new batch of ten games at one hour/game is underway.

The Citrine is advertised by Our Sponsor to have a 2330 elo, but I'll stick with my original estimate of 2000-2050 elo.
User avatar
sje
Posts: 4675
Joined: Mon Mar 13, 2006 7:43 pm

Last one for now

Post by sje »

A ten game match with 60 minutes per game:

http://idisk.mac.com/chessnotation-Public?view=web

See: PGN -> Citrine -> g60min.pgn

Score: Symbolic 10 wins, 0 losses, 0 draws

If you own a Citrine and have a hard time beating the machine, these game scores might be useful.