Hi, I'm afraid banksia doesn't keep time well.
The original time control is 3'+3"
Coiled 1.2 plays white.
As you can see in the .log, coiled 1.2 has 3 seconds left plus 3 incremental seconds.
Coiled is supposed to have: wtime 3232 + winc 3000 = 6232.
Because banksia gives him Timeleft.
log
10.23:32:00 Coiled 1.2 x64< position startpos moves e2e4 e7e5 g1f3 b8c6 f1b5 a7a6 b5a4 g8f6 e1g1 f8e7 f1e1 b7b5 a4b3 d7d6 c2c3 e8g8 h2h3 c6a5 b3c2 c7c5 d2d4 d8c7 d4d5 f6d7 b2b3 d7b6 c1e3 c8d7 b1d2 c5c4 d1e2 f8c8 b3b4 a5b7 f3h2 a6a5 a2a3 a8a6 e1b1 c8a8 e2d1 e7d8 d2f1 c7c8 f1g3 a5b4 a3b4 a6a2 g3e2 f7f5 f2f4 f5e4 a1a2 a8a2 b1c1 b6a4 c2b1 a2b2 b1e4 d8f6 f4e5 f6e5 e2d4 c8e8 d1f3 h7h5 g1h1 b7d8 d4c6 d8c6 d5c6 d7e6 e3f4 e5f4 f3f4 d6d5 e4f3 e8c6 f3h5 d5d4 h5f3 c6d7 c3d4 d7f7 f4b8 f7f8 b8b5 b2b4 b5e5 f8f5 e5e1 b4b6 f3e4 f5h5 c1a1 e6d7 h2f3 b6f6 h1h2 a4b6 a1a5 h5e8 a5g5 f6h6 f3e5 e8e7 e1g3 e7f6 e4d5 g8h7 d5f7 d7f5 g3f4 f5b1 f4e3 b1d3 g5g3 b6c8 e5c4 f6f7 e3d3 h7g8 c4e5 f7e6 e5f3 c8d6 f3g5 e6e7 d3b3 d6f7 g3f3 e7g5 b3f7 g8h7 f7f5 g5f5 f3f5 h6a6 h2g3 a6a2 f5f2 a2a4 f2d2 h7g6 g3f4 g6f6 h3h4 a4a8 d2c2 a8h8 g2g3 g7g6 c2c6 f6f7
10.23:32:00 Coiled 1.2 x64< go wtime 3232 btime 3887 winc 3000 binc 3000
10.23:32:00 Coiled 1.2 x64> info depth 1 seldepth 1 time 0 nodes 3 nps 0 hashfull 1000 score cp 872 pv c6c5
10.23:32:00 Coiled 1.2 x64> info depth 2 seldepth 2 time 0 nodes 9 nps 0 hashfull 1000 score cp 872 pv c6c5 f7f6
...
10.23:32:01 Coiled 1.2 x64> info depth 24 seldepth 38 time 938 nodes 903759 nps 963000 hashfull 1000 score cp 879 pv c6c5 f7f6 c5c6 f6f7 c6c7 f7e6 c7a7 h8h5 a7a3 e6f6 a3c3 f6e6 c3b3 e6f6 b3b6 f6f7 b6b7 f7f6 b7b3 f6g7 b3c3 h5f5 f4g4 f5d5 c3c4 g7f6 g4f4 d5d8 c4a4 f6g7 f4g4 g7f6 g4f3
10.23:32:01 Coiled 1.2 x64> info depth 25 seldepth 42 time 1250 nodes 1181529 nps 945000 hashfull 1000 score cp 868 pv c6c5 f7f6 c5c6 f6f7 c6c7 f7f6 c7c5 h8d8 c5c6 f6f7 f4e4 d8e8 e4f3 e8a8 c6c3 a8d8 f3e3 d8h8 c3a3 h8h5 e3f4 f7g7 a3c3 h5f5 f4g4 f5d5 c3c4 g7f6 g4f4 d5d8 c4a4
10.23:32:03 banksia> Timeleft for Coiled 1.2 x64: 3.23, used: 3.23, Arasan 23.3 64-bit: 3.89
10.23:32:03 Coiled 1.2 x64< stop
10.23:32:03 Coiled 1.2 x64< stop
10.23:32:03 Coiled 1.2 x64> info time 3235 nodes 2807922 nps 867000 hashfull 1000
10.23:32:03 Coiled 1.2 x64> bestmove c6c5 ponder f7f6
10.23:32:06 Arasan 23.3 64-bit< quit
10.23:32:06 Coiled 1.2 x64< quit
Time control error. Timeleft. Version 0.54 beta
Moderators: AlexChess, TedSummers
Re: Time control error. Timeleft. Version 0.54 beta
wtime is the total time the white side can use to compute a move. If an engine uses more time than that value, it will lose by time.
winc is the amount of adding time that will be added to the time left AFTER making that move.
I think engines typically consider only wtime/btime and ignore winc/binc.
winc is the amount of adding time that will be added to the time left AFTER making that move.
I think engines typically consider only wtime/btime and ignore winc/binc.
Re: Time control error. Timeleft. Version 0.54 beta
Hi!
Fixed bug by wtime.
Now it doesn't fail by Timeleft
Thanks.
Fixed bug by wtime.
Now it doesn't fail by Timeleft
Thanks.