Frank de Groot wrote:
It depends on the rules. If you want to know more about superko-issues and some possible solutions read my icga article and recent articles of Kishimoto.That may depend on what you're doing. GNU Go is doing fairly well without any superko checks. It has been on the TODO list for ages but nobody has found it important or fun enough to implement so far.
I thought that when doing MadLab- / GoTools-like L&D search, that SuperKo was important tot check for as pruning contraint? When generating the movetree, would you otherwise not generate a lot of superko situations?
Anyone knows?