After a discussion with Larry Kaufman on the Rybka forum over material values,
http://www.rybkaforum.net/cgi-bin/rybka ... l?tid=1264
I decided to come up with a CM9000 setting and a CM10th setting based on his values.
I based the material values on pawn=1.1 for CM9000 and pawn=1.2 for CM10th and went from there.
The reason for not using pawn=1.0 as a base is my observation that CM will give up a pawn too easily based on positional considerations that usually fail to bear fruit.
Larry appeared to agree that this was sound reasoning.
Valuing the pawn more highly of course means that the right balance of positional settings must also be found so that the overall package works in unison (material and positional settings together).
Because CM GUI settings only allow values to one decimal place, I needed to take that into account because I've always made it a rule with my settings that they must be based on true CM GUI values.
Also, because my impression is that CM doesn't value the bishop pair highly enough in any hidden settings, I made a slight alteration to compensate.
I gave the bishop a slightly higher value than the knight, then increased the mobility importance and decreased the pawn weakness values to try and make up for Larry's half pawn advantage for the bishop pair.
Because the bishop pair comes into its own in open positions, too much consideration for pawn chains through pawn weaknesses being given too high a priority would lead to blocked positions that would obviously favour the knights.
So here are the two settings I came up with, based on the highly successful CM9000 Enforcer and CM10th Paralyse settings with the necessary modifications:
CM10th Guardian (Graham Banks)
opp=120 opn=128 opb=132 opr=120 opq=130 myp=120 myn=128 myb=132 myr=120 myq=130 cc=120 mob=125 ks=150 pp=120 pw=110 cfd=0 sop=100 avd=0 rnd=0 sel=14 md=99
Material values under CM10th GUI would be:
P=1.2 N=3.8 B=3.9 R=6.0 Q=11.7
CM9000 Tribute (Graham Banks)
opp=110 opn=117 opb=121 opr=110 opq=119 myp=110 myn=117 myb=121 myr=110 myq=119 mycc=110 mymob=115 myks=160 mypp=115 mypw=110 opcc=110 opmob=115 opks=160 oppp=115 oppw=110 avd=0 cfd=0 sop=100 rnd=0 sel=12 md=99
Material values under CM9000 GUI would be:
P=1.1 N=3.5 B=3.6 R=5.5 Q=10.7
I was hoping that some CM devotees could test these settings out and let me know their thoughts.
For any still keen Chessmaster enthusiasts
Moderators: hgm, Rebel, chrisw
-
- Posts: 41457
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
-
- Posts: 900
- Joined: Wed Mar 08, 2006 9:06 pm
Re: For any still keen Chessmaster enthusiasts
Hi Graham
Where can i find the original settings for CM9000 Enforcer and CM10th Paralyse?
Thanks
Dave
Where can i find the original settings for CM9000 Enforcer and CM10th Paralyse?
Thanks
Dave
-
- Posts: 41457
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
Re: For any still keen Chessmaster enthusiasts
CM9000 Enforcer (Graham Banks)David Dahlem wrote:Hi Graham
Where can i find the original settings for CM9000 Enforcer and CM10th Paralyse?
Thanks
Dave
opp=110 opn=113 opb=113 opr=116 opq=118 myp=110 myn=113 myb=113 myr=116 myq=118 mycc=110 mymob=110 myks=160 mypp=115 mypw=115 opcc=110 opmob=110 opks=160 oppp=115 oppw=115 avd=0 cfd=0 sop=100 rnd=0 sel=12 md=99
CM10th Paralyse (Graham Banks)
opp=120 opn=123 opb=123 opr=126 opq=128 myp=120 myn=123 myb=123 myr=126 myq=128 cc=120 mob=120 ks=150 pp=120 pw=120 cfd=0 sop=100 avd=0 rnd=0 sel=14 md=99
Were you wanting the material values under the CM GUIs as well David?
Regards, Graham.
-
- Posts: 900
- Joined: Wed Mar 08, 2006 9:06 pm
Re: For any still keen Chessmaster enthusiasts
Thanks Graham. No, i don't need the CM GUI values. I'm using the Arena interface.Graham Banks wrote:CM9000 Enforcer (Graham Banks)David Dahlem wrote:Hi Graham
Where can i find the original settings for CM9000 Enforcer and CM10th Paralyse?
Thanks
Dave
opp=110 opn=113 opb=113 opr=116 opq=118 myp=110 myn=113 myb=113 myr=116 myq=118 mycc=110 mymob=110 myks=160 mypp=115 mypw=115 opcc=110 opmob=110 opks=160 oppp=115 oppw=115 avd=0 cfd=0 sop=100 rnd=0 sel=12 md=99
CM10th Paralyse (Graham Banks)
opp=120 opn=123 opb=123 opr=126 opq=128 myp=120 myn=123 myb=123 myr=126 myq=128 cc=120 mob=120 ks=150 pp=120 pw=120 cfd=0 sop=100 avd=0 rnd=0 sel=14 md=99
Were you wanting the material values under the CM GUIs as well David?
Regards, Graham.
Regards
Dave
-
- Posts: 910
- Joined: Wed Mar 08, 2006 9:46 pm
- Location: Plovdiv, Bulgaria
Re: For any still keen Chessmaster enthusiasts
hi Graham
is there any software that can convert CM GUI values to these used in wb2uci?
is there any software that can convert CM GUI values to these used in wb2uci?
take it easy
-
- Posts: 41457
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
Re: For any still keen Chessmaster enthusiasts
I believe that it can be done using Crafty, but it's easy enough to do anyway (except if you change the material/positional parameter - MP, then you must use the Crafty log file output).GenoM wrote:hi Graham
is there any software that can convert CM GUI values to these used in wb2uci?
The positional values remain the same.
For material values, divide your value by the default, multiply by 100 and always round down where rounding is necessary.
Example - if your N=3.3 then 3.3 divided by 3 = 1.1 times 100 = 110
Hope this helps,
Regards, Graham.
-
- Posts: 910
- Joined: Wed Mar 08, 2006 9:46 pm
- Location: Plovdiv, Bulgaria
Re: For any still keen Chessmaster enthusiasts
Hi Graham
Thanks for your answer.
And i want to ask you about selective search in CM -- did anyone made any research about values under defaults? How they affect the results in blitz and standard games? Any tests?
regards,
Geno
Thanks for your answer.
And i want to ask you about selective search in CM -- did anyone made any research about values under defaults? How they affect the results in blitz and standard games? Any tests?
regards,
Geno
take it easy
-
- Posts: 41457
- Joined: Sun Feb 26, 2006 10:52 am
- Location: Auckland, NZ
Re: For any still keen Chessmaster enthusiasts
Extensive tests were carried out by a number of testers although I haven't got the links.GenoM wrote:Hi Graham
Thanks for your answer.
And i want to ask you about selective search in CM -- did anyone made any research about values under defaults? How they affect the results in blitz and standard games? Any tests?
regards,
Geno
For CM9000, selectivity 12 was widely regarded as best.
For CM10th, selectivity 14 got the nod.
These seemed to hold generally true for both blitz and longer time controls as can be evidenced in CM blitz testing.
Regards, Graham.