Re: [gnugo-devel] regress endgame:603

[Arend Bayer]

Dan wrote:

> It may make sense to have problems for which we have no machinery
> yet to solve them.
>
> For example we've been discussing another problem century2002:180
> in which there is no method in the engine for recognizing that
> making or preventing seki is worth a few points.

But the drawback is that e.g. endgame:603 will do nothing but producing
accidental FAILs and PASSes for (I guess) at least another year or so.
(Until someone starts to write a CGT-like or lookahead endgame module.)

Arend