Sanny wrote:
I "face it now" by trading down to an elementary win,
only to have the game wrongly adjudicated as a draw
because somebody atGetClubdoesn't know much
aboutchess. :D

KBB vs. K is a forced win, just like KR vs. K is.

I know KBB vs K is forced win.

But when Computer has 2 Bishops extra and it will never be able to
find a win. So the game will never end.

If it is the computer with the extra material then it has to offer a
draw because it doesn't know how to win. But it cannot claim a draw
against a human opponent with that same extra material just because it
doesn't know how to win. You could always add tablebases to sort this
out onces and for all.

Or a few extra heuristics in the evaluation when the number of pieces on
the board are 5 or less and special cases for Kxx against lone K.
KBBk and KBNk are probably the most important gaps here. KNNNk although
a theoretical win requires a perverse set of circumstances to arise in
real play.

I has giving these cases as Draw.

KBB vs K

KNB vs K

But these are both mostly clear wins for a half decent human player
against any defence. They would be a routine win against Getclub because
it has no understanding of how the mating net is constructed.


KNN vs K

KB vs K

KN vs K

KN vs KB

KN vs KR

KB vs KR

In all above cases game will end in a draw.

Generally true except in the KR case where there may sometimes be a pin,
check or forced move that results in the capture of the N or B (which is
worth looking for before declaring it drawn). Reducing to KRk.

Regards,
Martin Brown
** Posted from http://www.teranews.com **