If "fair" marking were "dominant" over "good" marking, then it might be possible for two "fair" parents to produce a "good". Ff X Ff = 75% fair, 25% good. FF X Ff = 0% good.
However, if "good" marking is recessive to "fair", then a pair of "good" parents could NEVER produce a "fair" marker
With marking, as with other traits it is very likely that the mode of inheritance is NOT that simple. It could be multi-genetic or involve modifier genes. Just like we can get dysplastic hips from two "excellent" parents.
In both cases, our problem is that we can only go by "phenotype" in making our choices (not to mention the subjectivity & emotional that creeps into it). We are NOT able to see the "genotype" ... actual genetic makeup of the dog.
Obviously, we see some generational relationship between good markers and not-so-good. You are asking: "Why is that so?"
Eyesight in humans differs. Why not in dogs? Maybe marking has more to do with physical eyesight than we realize?
Maybe a smarter dog can adjust for what he lacks in eyesight? So, marking could be a combination of eyesight capability and intelligence level? However, if eyesight were so bad that the dog literally cannot see the bird at long distance, all the intelligence in the world will not be able change that.
Memory of some sort plays a role in long marks. When a dog runs through cover & change of terrain on its way to a mark, how does it remember where it's going? I could not do what some (a lot!) of these dogs do! They must have some kind of internal GPS to stay on line in such situations?
How much does "personality" influence the scenario? Some dogs (and humans) seem to be inherently more "intense" in everything they do.
Just food for thought ... as I've also wondered what magical combination of genes it takes to make some dogs so good at this; and others not very good at it.