As Mabel Adams once posted, you rule out the presence of anti-K1 every time you do an antibody screen and there is no agglutination of the screen cell that is K+k+!

On the original discussion, the whole question becomes moot if you allow rule outs with the same criteria as you do with your negative antibody screens--one homozygous cell for most, 1 hetero for K.  The rule of 3 came about for improving statistical accuracy when identifying antibodies, not for ruling them out.