# Computer Chess Club Archives

## Messages

### Subject: Re: Parameter Tuning

Author: jonathan Baxter

Date: 15:05:41 10/02/98

Go up one level in this thread

```On October 02, 1998 at 12:32:18, Don Beal wrote:

> It's fun to run these methods, and see "knowledge" appearing in the
>weights.  My first reaction was "wow, it's really learnt some chess
>knowledge - by itself".  It's only weight-tuning, of course, but it
>still impresses me that the weights match (roughly) elementary chess
>

I agree, its heaps of fun. I had attributed the piece/square weights
turning out as "standard" chess knowledge to the fact that my program
is learning against opponents with the standard chess knowledge built in.
But if you see it in self play then that can't be the whole story.....

> Don Beal wrote:
>>> The method works for any weights, not just piece values, provided
>>>that the evaluation consists of a sum of term*weight components.
>>
>>This isn't true. You can use the method for any evaluation function provided
>>it is  a differentiable function of its weights. This includes linear evaluation
>>functions, but is not restricted to them.
>>Cheers,
>>
>>Jonathan Baxter
>
>Erm, I think you mean:  not only is Don's statement true, but the method
>can also be applied to a wider range of evaluation functions, provided
>it is differentiable with respect to each weight.
>My statement was correct.  Your point is correct too.

Well, you did say "*provided* the evaluation consists of a sum of ...." which
would usually be taken to mean it is a *necessary* not just sufficient
condition.

>BTW, we applied the same method to Shogi, and learnt piece values there.
>That's directly useful, because Shogi doesn't have a standardised set of
>values that programmers can pick up and use.

I saw one of your papers on this (and liked it :). One question I have is: the
piece values drift
an awful lot, with the drift seeming to reduce only with the decrease in step
size. I was wondering, if you keep the step size constant (but small) and run
for 100's of 1000's of games, do the pieces swap their relative ordering a lot
or do they finally settle to a constant relative ordering?

Cheers,

Jonathan Baxter

```