[Retros] Add pieces

Noam Elkies elkies at math.harvard.edu
Tue Mar 8 23:13:20 EST 2011


Kevin Begley <kevinjbegley at gmail.com> writes:


> [.....] I'm not certain why you care to distinguish between problems

> in which a royal unit is/isn't under attack. [...]


I'm too inexperienced in this genre to have much to say about most of
these topics, but this one seems clear enough: check is such a strong
constraint on the previous move (at least if the checking side has
more than one unit with an otherwise viable retro-move...) that it makes
the constructor's task much easier, so it makes sense to separate
"in check" from "not in check" tasks. Same idea as placing a higher
value on (say) a captured promotee if the capture wasn't necessary
just to account for the pawn structure.

NDE



More information about the Retros mailing list