Discussion of chess software programming and technical issues.
Moderators: hgm , Rebel , chrisw
Daniel Anulliero
Posts: 759Joined: Fri Jan 04, 2013 4:55 pmLocation: Nice
Post by Daniel Anulliero Tue May 26, 2015 5:38 pm
Small update , after 282 games out of 1000 :
Code: Select all
TEST LMR 1
DANY-PC, 2015.05.24 - 2015.05.26
--------------------------------------------------------
1: ISA lmr aggressif 68.0 / 113 (+52 -28 =32)
2: ISA lmr depth - 2 60.5 / 113 (+48 -40 =25)
3: JARS_175 59.0 / 112 (+48 -42 =22)
4: ISA lmr a la senpai 50.5 / 114 (+34 -46 =33)
5: ISA pas de lmr 43.0 / 112 (+32 -58 =22)
---------------------------------------------------------
282 games: +104 -110 =67 *1
Ferdy
Posts: 4833Joined: Sun Aug 10, 2008 3:15 pmLocation: Philippines
Post by Ferdy Tue May 26, 2015 5:48 pm
Daniel Anulliero wrote: Small update , after 282 games out of 1000 :
Code: Select all
TEST LMR 1
DANY-PC, 2015.05.24 - 2015.05.26
--------------------------------------------------------
1: ISA lmr aggressif 68.0 / 113 (+52 -28 =32)
2: ISA lmr depth - 2 60.5 / 113 (+48 -40 =25)
3: JARS_175 59.0 / 112 (+48 -42 =22)
4: ISA lmr a la senpai 50.5 / 114 (+34 -46 =33)
5: ISA pas de lmr 43.0 / 112 (+32 -58 =22)
---------------------------------------------------------
282 games: +104 -110 =67 *1
Also try to run ordo.
PK
Posts: 893Joined: Mon Jan 15, 2007 11:23 amLocation: Warsza
Post by PK Tue May 26, 2015 9:30 pm
Senpai method is not exactly what you describe. IIRC it does not reduce first three moves, moves 4-6 get reduced by 1 ply, remaining moves by depth / 3.
Daniel Anulliero
Posts: 759Joined: Fri Jan 04, 2013 4:55 pmLocation: Nice
Post by Daniel Anulliero Tue May 26, 2015 10:16 pm
PK wrote: Senpai method is not exactly what you describe. IIRC it does not reduce first three moves, moves 4-6 get reduced by 1 ply, remaining moves by depth / 3.
Yes true it's not really the senpai method in my engine , it's an inspiration 😊
Daniel Anulliero
Posts: 759Joined: Fri Jan 04, 2013 4:55 pmLocation: Nice
Post by Daniel Anulliero Fri May 29, 2015 8:00 pm
Update after 534 games :
Code: Select all
TEST LMR 1
DANY-PC, 2015.05.24 - 2015.05.29
Score
-------------------------------------------------------
1: ISA lmr aggressif 123.0 / 214 (+94 -61 =58)
2: ISA lmr depth - 2 111.5 / 213 (+88 -78 =47)
3: JARS_175 106.0 / 213 (+87 -87 =38)
4: ISA lmr a la senpai 106.0 / 215 (+78 -79 =56)
5: ISA pas de lmr 85.5 / 213 (+65 -107 =41)
-------------------------------------------------------
534 games: +196 -216 =120 *2
ok ... It seems Isa is not much stronger than my old Jars ..
Some work (and patience
) ) to do again ...
may be the "lmr depth - 2 " version is much promising than the "aggressive " version ? :
Code: Select all
ISA lmr aggressif - JARS_175 : +21-18=14
ISA lmr depth - 2 - JARS_175 : +28-17=8
) ) to do again ...