tag:blogger.com,1999:blog-4193297034681066259.post927638206101034981..comments2024-03-04T09:16:25.999+01:00Comments on Mediocre Chess: [Plan] A crude repetition detectionJonatan Petterssonhttp://www.blogger.com/profile/03152085169762414586noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-4193297034681066259.post-8990001663858404352007-01-05T13:59:00.000+01:002007-01-05T13:59:00.000+01:00There is no advantage of using FEN-strings instead...There is no advantage of using FEN-strings instead of Zobrist keys, except for easy implementation at this point since I already had them available..<br /><br />Not only can zobrist keys be compared quickly, and used in hash tables easily. They are very easy to maintain in the Board-class so a key is always available and you don't have to generate it every time.<br /><br />I only use FEN-strings temporarily at this point for development purposes. I will be implementing zobrist keys in the near future.<br /><br />About declaring any repetition a draw I am not sure yet. Most likely I will do as you say, I am not too worried about the repeats.Jonatan Petterssonhttps://www.blogger.com/profile/03152085169762414586noreply@blogger.comtag:blogger.com,1999:blog-4193297034681066259.post-76780041810769518202007-01-05T13:36:00.000+01:002007-01-05T13:36:00.000+01:001. Is there an advantage in using FEN strings inst...1. Is there an advantage in using FEN strings instead of the Zobrist keys? Zobrist keys can be compared quickly.<br />2. Draw detection prunes the search tree and speeds up the search; so I find it useful to internally declare any (not only the 2nd) repetition a draw. This could theoretically be abused by an opponent who knows the engine (if he blunders in a won position but still retains a positive score, he could reverse his move, and the engine would reverse its own move in hope for a draw), but I never saw that happen.Anonymousnoreply@blogger.com