[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [computer-go] Super Ko on KGS ignores player to move

On 8/13/05, john tromp <John.Tromp@xxxxxxxxxxxxxxxxx> wrote:
> That's right. Chinese rules use the simpler positional superko,

No they don't. The official Chinese rules do *not* use superko. (*1)

The examples shown at http://www.cs.cmu.edu/~wjh/go/rules/Chinese.html
(Chapter 3. Section 20) clearly indicate that balanced (*2) long
cycles are declared drawn. The only exception to this is moonshine
life (which is always considered dead).

I'm not sure what exactly has been implemented on KGS, but if it is
true that positional superko is used then the claim that KGS has "the
most accurate interpretation possible" (see
http://kgs.kiseido.com/en_US/help/ruleSets.html) is definitely false.


*1 The term superko could be used if the definition would be extended
to allow for direct draws under certain special conditions. However as
far as I know all established superko rules have the property that
they can only declare repetition illegal (or equivalently, a direct
maximal loss for the player that made the last move that created the

*2 A cycle is balanced if after one cycle the player to move is the
same and the number of stones placed on the board (or equivalently the
number of passes) is equal for both sides.
The alternative, unbalanced cycles, have the property that they either
cannot constitute legal play (with players taking turns consecutively)
or are the result of clear abusive plays by one side to stay in the
cycle (such as may happen with a send-two-return-one in the final
computer-go mailing list