Genetic Inheritance Patterns

[Agent Smith]

It's been a while since I used my ancient biology files. I hope they aren't so outdated that what I'm about to post and what I'm about to ask isn't just plain nonsense. Well, here goes ...   

 

Genes come in pairs, for every trait there is, let's say for trait T, there's a recessive gene a and a dominant gene A. Offspring inherit combinations of these genes from their parents. Using a Punit square we can see that we have for trait T, 4 possible genotypes: AA, Aa, aA, aa. I recall being taught that dominant genes fully express themselves in the offspring with genotypes AA, Aa, aA. That's 3 out of 4 children that'll manifest the phenotype encoded by the dominant gene, here A. In only 1 out of 4 children will the recessive gene express itself. 

My query is is this safe/good for evolution? What's the evolutionary advantage of this type of inheritance pattern? My own guess is dominant genes have passed the test of natural selection (they have a history of being evolutionarily successful) and so it's better that they're expressed in the offspring, increasing the odds of survival.

The recessive gene (here a) is our backup plan B, a just in case thing, a space for mutation to create novel traits, which might turn out to be a life-saver. This would require mutations to be more frequent in recessive genes than dominant ones. There are downsides though; not all mutations are beneficial. I suggest reading sickle-cell anemia.