Chessmaster Blitz - first CM11 setting

[Graham Banks]

Hi,

Just wait and see....
Werner and i has now played around 400-450 games with the CM XI Sauron settings. Let's see in the next CEGT list what it gives compared to The King 3.50 default.
It could be a surprise

Best
Johan

Hi Johan,

Ray is testing many of the new CM11 personalities at blitz.
After 200 games, Sauron is 10 ELO higher than Default. Still a long way to go though!
Looking forward to seeing your results.
It could well be that selectivity 21 hurts at blitz, but is good at longer time controls.

Regards, Graham.

[Graham Banks]

ah, so what do you have left to test the cm personalities. don't tell me you use a slide rule and abacus?

I usually use just one GUI for CM experimenting.

[ozziejoe]

ah, than we are on even playing field. I have a chance.

As for your previous post about cm selectivity not being good for blitz, I am in the middle of a large test where I only changed selectivity and time controls are very fast (1/1). and I am finding that selectivity of 21 is making a huge improvement. I can not believe my eyes. I'll post later, but right now i am guessing that cm goes from 40% (default) to 50% (sel =21) winning against my particular field of competitors.

If you try to modify the parameters within the cm gui, it does not look like you can even set selectivity that high. I don't think de kooning did a carefull test of these parameters. My guess is he improved the program in some fundmental way (e.g., by adding knowledge as you suggested earlier) and did not tweek the parameters much.

[Graham Banks]

As for your previous post about cm selectivity not being good for blitz, I am in the middle of a large test where I only changed selectivity and time controls are very fast (1/1). and I am finding that selectivity of 21 is making a huge improvement. I can not believe my eyes. I'll post later, but right now i am guessing that cm goes from 40% (default) to 50% (sel =21) winning against my particular field of competitors.

If you go to the cm gui, it does not look like you can even set selectivity that high. I don't think de kooning did a carefull test of these parameters. My guess is he improved the program in some fundmental way (e.g., by adding knowledge as you suggested earlier) and did not tweek the parameters much.

That's really interesting about the higher selectivity not hurting at blitz!
Yes - from memory, you can only go up to selectivity 16 in the GUI?