Yes, this is an interesting problem.
66% is the raw statistical probability of a Het assuming basic genetic assortment and no additional data.
As soon as you produce a clutch: that 66% value *technically* should be adjusted based on statistical results.
If you produce no Homozygous (HO) offspring in a 66% Het vs. HO mating: then that Het probability percentage goes *downward* proportionally to the number of non-HO offspring produced--Likewise a single HO offspring adjusts the value to a proven 100%.
In addition, you may have a 66% Het that seemingly produces no proven offspring from one cross, but, there are several potential factors that could have prevented the Homozygous form from being viable (with no noticeable evidence). If you cross it to a different specimen then you may have very different results.
Anyhow, my![]()