Say there is some actual adaptive advantage like upper body strength. Genes that cause greater upper body strength tend to survive, and because of this genes in

Say there is some actual adaptive advantage like upper body strength. Genes that cause greater upper body strength tend to survive, and because of this genes in the opposite sex that can detect upper body strength in a potential mate tend to survive. The population of individuals of the opposite sex, tend towards a preference for upper body strength. What is the limit? Well the answer is very little, while more upper body strength may have little additional value above some level there is no environmental influence that selects against it until it becomes so pronounced as to become an impairment. If there are mutations in genes of the opposite sex that prefer evidence of upper body strength it is unlikely they will be removed, while genes that disregard upper body strength will occasionally fail to distinguish good quality mates and will therefore become less prevalent. Further, the appearance of upper body strength may become so preferred that fake upper body strength is also a more successful trait at least to the degree that it doesn't impair survivability. Because of all of this, sexual selection is a mechanism in which a small astatic for actual survivability traits becomes a large exaggeration of traits that suggest survivability but may or may not actually improve survivability. In the case of the peacock tail, and birds in general it probably stems from the fact that sick birds often have dirty tails, therefor females that prefer male birds with not dirty tails (or better) are selected for.