Re: What is "randomness" for a CM9k personality?

I assume that all moves do not receive the same adjustment. That is, the
best move if randomness is zero may be given a -10 if randomness is 10,
and the sixth best move if randomness is zero may be given a +6, say.

I like it. Given that we can design our own personalities, Chessmaster
has given us tremendous flexibility in creating new personalities that
don't always play the same move in a position, and we can't easily
predict what move it will make if we give a randomness above zero.

God, I love this program. Creating opponents is my favorite pastime.

BTW, when I said that Toni with randomness beat Toni without randomness,
I had gotten confused. It wasn't randomness I adjusted, it was the
evaluation of pawns. So disregard.

Wilma


JVMerlino@xxxxxxx wrote in
news:1133921223.157838.214870@xxxxxxxxxxxxxxxxxxxxxxxxxxxx:

> The randomness value basically adds or subtracts a number range to
> each legal move's evaluation. So, let's say that this number is 10;
> then each legal move's evaluation can be adjusted by any number
> between -10 and +10.
>
> However, I DON'T think that it's a 1-to-1 equivalent between the
> actual randomness value in the personality and the number that is used
> to define the potential range of adjustment. In other words, if you
> have a randomness value of 10, I don't think that means that each
> score will be modified by a range of -10 to +10, or even -0.10 to
> +0.10. I think it's a bit more complex than that.
>
> But that's basically how it works. And, of course, as Johan told me,
> any randomness value greater than zero will result in overall lower
> quality play.
>
> jm
>
>

.

Relevant Pages