Dave,
I think you need to rethink this one, you are pretty far off the mark on
this. Chess programs of today do not resemble even programs of the 80's,
not to mention programs of 1950 - even when adjusted for the amazing
hardware improvements.
Our understanding of how to build a strong chess program changes every year
- you only have to look to see that. Compare Fritz, for example, over the
last 20 years on the same hardware and I think you would be surprised.
Like everyone else Franz (the author of Fritz) was forced to constantly
upgrade his program.
This is WAY MORE than just alpha/beta and counting pieces.
Don
On Wed, Jun 2, 2010 at 12:43 AM, Dave Dyer <[email protected]> wrote:
> At 05:30 PM 6/1/2010, Greg Walton wrote:
>
> Hello,
>
> Could someone help me understand [in layman's terms] the different
> challenges involved in programming AI for Chess as opposed to Go?
>
>
> It's very simple. There are a few obvious things to do to create a Chess
> AI; and *they work.* ie; Count the wood. Give a little consideration to
> position, and use alpha-beta to prune the lookahead tree. All these things
> were well understood by about 1950 and there have been only incremental
> improvements
> since then. The main improvement has been that computers are thousands of
> times faster than when these techniques were developed.
>
> By contrast, there are no obvious evaluation metrics that work for Go, and
> the
> search tree is impossibly large. Advances in computer speed have not, and
> arguably never will make the same simple, brute force methods that work for
> Chess suitable for Go.
>
> _______________________________________________
> Computer-go mailing list
> [email protected]
> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
>
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
