Re: How long would it take a computer to completely "solve" chess?
From: Sean O'Leathlobhair (jwlawler_at_yahoo.com)
Date: 09/15/04
- Next message: *** T. Winter: "Re: The Star-Spangled Conjecture"
- Previous message: edA-qa mort-ora-y: "Re: Solve/Approx: th / (1 - cos th) = c"
- In reply to: Guy Macon: "Re: How long would it take a computer to completely "solve" chess?"
- Next in thread: Guy Macon: "Re: How long would it take a computer to completely "solve" chess?"
- Reply: Guy Macon: "Re: How long would it take a computer to completely "solve" chess?"
- Messages sorted by: [ date ] [ thread ]
Date: 15 Sep 2004 07:29:10 -0700
Guy Macon <http://www.guymacon.com> wrote in message news:<10ke628ibj2a099@news.supernews.com>...
> (I will be starting a hot project later this week and will be limiting
> my newsgroup usage to reading selected posts and an occasional reply.)
>
> Sean O'Leathlobhair <jwlawler@yahoo.com> says...
>
> >But if we do any pruning, can we be sure that we have not missed some
> >very obscure trick? We may have solved the game for all practical
> >purposes but if we have pruned anything, could we really say that we
> >have theoretically solved chess?
>
> Yes. Yourvobjection is valid for the normal runing that chess
> programs do, but there are other ways to prune which have zero
> chance of missing something subtle like throwing away two rooks
> and a queen in order to set up a forced mate in 49. Here is how:
>
> Imagine that you are a program searching the entire move tree,
> looking to find wins, losses and draws.
>
> Now add a modification; if you see a position that has a king and
> a rook against a king, search one ply farther to rule out stalemate
> or losing the rook, mark it as a win for the side with the rook,
> and stop searching that branch. Why? because there exists an
> algorithm that wins every time from that position. You just pruned
> without any chance of missing something subtle.
>
> Now imagine a thousand years of humans and computers searching
> for more algorithms that will reduce the size of the search...
>
> It's not just that RK vs K position that you can prune at either;
> yiu can prune at any position that has 6 men on the board (soon it
> will be 8); we know the result from every such position already.
Thanks.
I can see that this sort of pruning is safe. I did not intend to
question all sorts of pruning, just alpha-beta-pruning (I should have
been clearer).
Others have responded to my question on whether specifically alpha
beta pruning is safe. I am still digesting their comments.
Seán O'Leathlóbhair
- Next message: *** T. Winter: "Re: The Star-Spangled Conjecture"
- Previous message: edA-qa mort-ora-y: "Re: Solve/Approx: th / (1 - cos th) = c"
- In reply to: Guy Macon: "Re: How long would it take a computer to completely "solve" chess?"
- Next in thread: Guy Macon: "Re: How long would it take a computer to completely "solve" chess?"
- Reply: Guy Macon: "Re: How long would it take a computer to completely "solve" chess?"
- Messages sorted by: [ date ] [ thread ]