possible hets, in this case a 50% poss het is the result of breeding a heterozygous animal to a normal. From such a pairing, 50% of the offpsring have a chance of carrying the trait, the other 50% won't. This is statistically based, so with a large sample set you should roughly see a 50/50 outcome. The only problem is that ALL the offpsring from such a cross will phenotypically appear normal. Another way of looking at it is that each individual egg has a 50/50 chance of carrying the trait.
Furthermore, when dealing with recessive trait and you cross a heterozygous animal to another heterozygous animal the result is a 1:2:1 ratio genotypically and a 3:1 ratio phenotypically. So out of 4 offspring, 1 will be completely normal, 2 will be heterozygous, and 1 will be homozygous recessive from a genotype standpoint. Phenotypically, you'll have 3 normals and 1 recessive. This gives rise to the term 66% hets because 2 of the 3 normals have the ability to potentially carry the recessive trait.
In all honesty, the whole possible het thing is a crap shoot, but it does at least tell you that somewhere within the lineage a particular trait was introduced. Otherwise, it's pretty much just a marketing term to get you to pay more for something that odds wise will most likely just be a normal unless you buy the entire clutch to one by one prove out who is and isn't het.
Otherwise, it's pretty much just a marketing term to get you to pay more for something that odds wise will most likely just be a normal unless you buy the entire clutch to one by one prove out who is and isn't het.