Computer programs tournament

Topics about Checkers programming and other projects.
I hope to keep users informed of the ACF website projects.

Re: Computer programs tournament

Postby Alex_Moiseyev on Fri Aug 12, 2016 11:09 pm

Ed Trice wrote:2. Weakly solved means winning every winnable position, not losing from the starting position, and being able to hold a draw from any position that is a draw. This is what Chinook can do.
Stop. we've got a point! It is very possible that Chinook can win any winning position, but this was never proved scientifically.

It's like Last Fermant's Theorem in math: for 358 years since 1637 everyone felt like it was a right statement, but this legendary theorem was finally proved only in 1995. Same situation here with Chinook - "it is obvious, but it is not proved."

The only thing really was proved, that if both sides make all right moves - it is a drawn result. Nothing else was proved scientifically. An arguments like ... "please play with Chinook and see how good it is" not working very well. :D

BTW. WCC perfect 7 pieces database is awesome!
I am playing checkers, not chess.
User avatar
Alex_Moiseyev
 
Posts: 4082
Joined: Sat Nov 12, 2005 5:03 pm

Re: Computer programs tournament

Postby megamau on Sat Aug 13, 2016 3:49 am

Ed Trice wrote:Edit: I think I see the issue. We have been using slightly different versions of these terms.

Yes, I think we got to the bottom of the issue. Your definition of "weakly solved" actually correspond to mine of "strongly solved".
I don't have a definition of "winning in the least amount of moves" because a win is a win, while you call it "strongly solved".

However, please note that the Schaeffer paper in science uses my definition (actually I copied and pasted from them), so when they claim to have weakly solved they don't mean it can win any win position.
User avatar
megamau
 
Posts: 15
Joined: Sun Aug 07, 2016 8:08 pm

Re: Computer programs tournament

Postby Ed Trice on Sat Aug 13, 2016 12:08 pm

megamau wrote:Yes, I think we got to the bottom of the issue. Your definition of "weakly solved" actually correspond to mine of "strongly solved".
I don't have a definition of "winning in the least amount of moves" because a win is a win, while you call it "strongly solved".

However, please note that the Schaeffer paper in science uses my definition (actually I copied and pasted from them), so when they claim to have weakly solved they don't mean it can win any win position.


I have often wondered how they settled on "weakly," "ultra-weakly," etc. The prefixes are nebulous at best.

So if their definition of "strongly solved" does not require announcing the distance-to-win, then what is it called when it can be predicted? "Ultra-strongly solved?"

I will see if I can introduce better nomenclature in the next paper I publish.
Ed Trice
 
Posts: 64
Joined: Wed Aug 10, 2016 9:16 pm

Previous

Return to Projects/Programming

Who is online

Users browsing this forum: No registered users and 1 guest

class=