[Computer-go] God Go Rank
Erik van der Werf
erikvanderwerf at gmail.com
Tue Oct 12 13:48:55 PDT 2010
On Tue, Oct 12, 2010 at 10:40 PM, Mark Boon <tesujisoftware at gmail.com> wrote:
> On Tue, Oct 12, 2010 at 10:31 AM, Nick Wedd <nick at maproom.co.uk> wrote:
>>> Under Chinese rules, the minimum error is more or less systematically 2
>> This is only true if you use a very pedantic measure of the size of an
>> error, assuming perfect play for all subsequent moves. More realistically,
>> a move can give away say 0.37 of a point, because it makes it that much less
>> likely that the player will play subsequent moves correctly.
> Indeed. Smaller errors are possible, they just may not affect the
> final outcome with further perfect play. Take for example a simple ko
> situation. If you have three of them they're worth one point. It makes
> the value of this ko 1/3rd of a point. (Despite the fact it's commonly
> called 1/2-point ko.) If I make a mistake where compared with perfect
> play I leave one stone in atari in a simple ko, I obviously made a
> 1/3rd point error, even if this didn't affect the final result. This
> is just an example, I'm pretty sure much smaller fractions are
> actually possible.
This only applies when approximating the game with combinatorial
theory. A God-like player may simply infer the minimax value, which
will take integer steps in the score.
More information about the Computer-go