To my way of thinking either they get the single yellow belly gene or they don't, and only the ones that get it could produce ivory. It's just that some of the het ivory may not always do a good job of showing it.
I feel pretty confident that there are some "recessive" mutations with co-dominant tendencies - labyrinth and granite in Burmese pythons and piebald in ball pythons. Some hets look perfectly normal but some seem to show an intermediate appearance between normal and the homozygous. I don't know what controls when the hets show and when they don't.
Similarly, there may be some "co-dominant" mutations with recessive tendencies. Perhaps some small percentage of genetic yellow bellies don't show all they classic traits.
Such a situation would be inconvenient and confusing but nature doesn't owe us nice easy cut and dry situations where every mutation fits into the textbook definition 100% of the time.
Has anyone produced classic yellow belies from the really questionable looking offspring of a yellow belly? Such a case would tend to support het ivories not always showing all the characteristics. However, it would take a long time without such a case to feel pretty confident that all het ivories show the full classic trait combination of a yellow belly.