I have honestly no idea what your argument is in this post, but let me update my 6-year-old post and see if I can clarify things.
First of all, like I said, whether an allele is dominant or recessive has nothing whatsoever to do with how common it is. Nothing. They’re completely independent properties. There are two forces that can alter the prevalence of an allele (that is, how common it is): selection and genetic drift.
The easiest way to think of this is as a probability question. You can Google “Hardy-Weinberg equilibrium” if you want a full treatment, but I’ll give a summary. Let’s say we have a gene pool. That is, some number of individuals who each have two copies of a gene. Let’s represent each copy as a marble in a bag. Let’s say the polydactyly allele are represented by red balls, and the “normal” allele is represented by green balls. Each individual can have two red balls, two green balls, or one of each, but only passes one on to each child.
So, to model our current human population, you’d have lots and lots of marbles, and almost all of them are green. That’s the parental generation. Now, let’s make the next generation, which will have the same number of individuals as the parental. To do this, for each individual, we reach into our bag of marbles and grab out two to determine the individual’s genotype. We then put the marbles back in the bag (selection with replacement), because each parent can have more than one child. If you do the probability math, you’ll find that you would expect the second generation to have very close to the same ratio of red to green balls. If the parental generation had 2% red, the second generation should have very close to 2% red, as long as the population sizes are large. This should be obvious, because each ball is equally likely to be picked.
Now, just due to random chance, there could be a slight deviation from the parental ratio. This is genetic drift, and it’s far more prevalent in small populations. If you’re only picking 100 balls, you’re much more likely to deviate from the expected 2% than if you picked 10 million balls.
Note that none of this is affected at ALL by whether green or red is dominant or recessive. The only difference that has at all is on the appearance of heterozygotes. In other words, if an individual has one red ball and one green ball, does he look red or green? That’s it. As long as people that look red are just as likely to have kids as people that look green (which is generally the case for polydactyly in particular), the ratio of red to green will stay unchanged (or very very close to unchanged) generation after generation.
Now, if you want to get into selection, that’s a different matter. Let’s say red people are twice as likely as green people to reproduce. You’d model this by doubling all the balls that go into the bag from all of the red people in the population, while leaving the green peoples’ balls the same. If you do the math on that, you’ll find that within not very many generations, your population will be entirely red (if red is recessive) or almost entirely red (if red is dominant).