[Retros] Variant Proof Games wanted

[andrew buchanan]

Has anyone got a Labelle-style database of exact openings that they can hit with the alpha-beta algorithm? Assign a non-extensible game 1/0 if White/Black is the last to move. Assign 0.5 if the game is on the frontier of what has been computed, and so its extensibility is not known. Et voila! With a bit of luck the value of the game will be determined to be 0 or 1 - if not, we should know which areas to explore more deeply.



----- Original Message ----
From: "Nicolas.Dupont at math.univ-lille1.fr" <Nicolas.Dupont at math.univ-lille1.fr>
To: The Retrograde Analysis Mailing List <retros at janko.at>
Cc: andrew at anselan.com; The Retrograde Analysis Mailing List <retros at janko.at>
Sent: Wed, January 27, 2010 11:51:20 PM
Subject: Re: [Retros] Variant Proof Games wanted

Dear all,

Andrey said :


> I suggested that

> it might be of interest to compose the longest proof game featuring an

> official game debut


A couple of years ago, I had a related idea : playing "exact chess". This
goes as ordinary chess, but the first player who makes a move such that
the resulting position is not exact looses the game !

Best,

Nicolas.



_______________________________________________
Retros mailing list
Retros at janko.at
http://www.pairlist.net/mailman/listinfo/retros