Re: [computer-go] Pattern matching - example play

["Frank de Groot" <frank@xxxxxxxxxxxxx>]

From: "Erik van der Werf" <E.vanderWerf@xxxxxxxxxxxxxxxxx>
Subject: Re: [computer-go] Pattern matching - example play


> Frank de Groot wrote:
> > What makes you think you can enlighten me about these issues?
>
> Your posts.


If you do not understand my posts it doesn't mean you can "explain" things
to me.
I have seen the same thing here when I said something about my pattern
system.
Religious rantings devoid of a shred of logic.
You better ensure that you know all there is to know (and I mane waaay
beyond Einstein-level) before you offer to lecture others about quantum
mechanics and time and stuff like that.

It's always funny to see how it's *always* and *only* the Dutch who think
they are Gods :)
(Tip: Read the "expatica" site to put your own position into perspective.)


> If for example optimal play is a deep narrow line there is a very good
> chance that such statistics derived from weak players are completely
> meaningless.

Well human statistics are even much more meaningless as players are never
equal like they are in a program.
Even a perceived "disadvantage" of starting at Tengen can ruin the
significance of the result.

> So far, the chess programmers have failed to make a decent 9x9 Go program.

Because there has been put vastly less resources into 9x9 Go compared to
Chess.
That is the only obvious difference, not the branching factors.
There are a few practical issues with the efficient representation of Go
positions and that kind of stuff.
It takes years of hard work to reach a comparable level as they have in
chess.
In chess, people build upon the works of millions of man-hours, in Go it's
pathetic what has been done and published.

> Personally I think that at the current level of computer Go 9x9 is very
> interesting. After all beginners start on 9x9, not on 19x19.

Well that's their choice.
I never did that and I managed to win from people who played Go for years.
So it's not a prerequisite.

> If we
> cannot even build a decent 9x9 player how can we ever hope to succeed on
> 19x19?


I advocate working on the actual problem, not on some pantsy sub-problem.
What works on 9x9 Go will never work on 19x19 Go and what is acceptable
performance (speed-wise) on 9x9 Go will prove sub-par on 19x19 Go so I don't
see any reason why I should bother with this kiddy stuff first. Life is
short, you know..

I can make a program that plays very well a 4x4 game but I'd have to scrap
it for 9x9.
Same with 9x9 vs. 19x19.

_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/