Many breeders only accept the idea of 66% chance hets (the normals from recessive het X het) and 50% chance hets (normal from recessive het X normal). However, if you bred a large number of 66% hets to normals on average 33% of the offspring would be hets and likewise if you bred a large number of 50% hets to normals on average 25% of there offspring would be hets. The problem is that considering any individual 66% or 50% it either is a het or it isn't so it's offspring with a normal are either 50% hets or 0% hets so some don't accept the 33% or 25% labels. I've got several 25% chance het girls from 50% chance het males I've not yet proven and I feel that is as good a way as any to keep track of the chance that they are hets.