Re: Any pro mistakes found?

[mueller@xxxxxxxxx (Martin Mueller)]

Darren Cook <darren@xxxxxxxxxxxxxxxxx> wrote:
>Though interesting, mathematical endgame theory is not actually that useful
>in real games.
>
>However I was wondering if the theory has managed to find mistakes in any
>high level professional games. Ie. real-life games, rather than carefully
>constructed problem positions.
>

There was a recent game between Jujo Jiang and Naiwei Rui which is
currently being analyzed by Berlekamp and his crew. They found several
mistakes, which were confirmed by the pros. There should be a paper on it
soon.

That game was played using 'environmental' rules, which adds a large number
of simple endgame plays to a normal Go board.

Bill Spight is going over some of the Japanese endgame books and keeps
finding lots of inaccuracies and errors.


>For that matter, are there any examples of computer analysis of endgames
>that have found better moves than the pros actually played?
>

I don't know any program that can do this. I do know that it will be a lot
of work to create such a program. I have some of the required pieces but
nowhere near enough.

My Ph.D. thesis program could solve endgame puzzles that are too hard for
Go pros, but solvable for combinatorial game theory experts.

        Martin (who did better than his program at the US Go congress...)