Firstly, personal attacks are not necessary.
Second, the Merle gene is not binary. It is not an on/off gene. The best way I have heard it explained is like this (by someone much smarter than me): Think of it as a scale of 1-100. A dog who scores less then 30 appears to be solid. A dog who scores between 31 and 70 appears Merle, a dog who scores 71+ is a "double Merle". (making those numbers up, I don't know the actual number of repeats it takes)
Breed a 37 to a 20 (appearing Merle to appearing solid) = 57, regular Merle, or less, producing a dog who appears solid, but may carry 15
Breed a 50 to a 25 (appearing Merle to appearing solid) = 78, double Merle! Or less, producing blues and solids
Breed a 79 to a 15 (double Merle to solid) = in theory, you should get all Merle's, but if the dd doesn't pass enough Merle repeats, you can get a solid.
We know that the mutation is unstable, and we know that not every dog passes ALL it's Merle repeats to every offspring. But any dog with a Merle parent is theoretically likely to carry SOME Merle repeats. Which means it can pass those to its offspring.
Where you are correct is that once your solid dog produces like a Merle, it should be treated as such from then on.
Add to that that Merle is of varying expression. If its possible to have a solid appearing dog produce like a Merle (say it has 50 repeats, but little expression), then why would it not be possible for a dog with 75 repeats (technically a double merle in this example) to appear like a regular Merle, but produce more Merle's than not? See where that goes? It's one of the reasons I am not in favor of breeding the cryptic dogs, I want to KNOW I have a Merle, I don't like surprises.