Do queens really "wear out" after two years? Deciding in October which colonies get a new queen in May

Tyrol, eighteen years, 24 colonies. I'm writing my replacement list for next spring and I want to argue about the received wisdom, because I keep hearing "requeen every two years" repeated as a law of nature and my own records don't quite support it. What's real: a queen mates once, stores sperm, and that store is finite. As it runs down she lays more drone brood in worker cells and her pheromone output falls, which is when the colony supersedes her. Most queens are at their most productive in their first full season and many are noticeably worse in their third. What I think is oversold: the calendar. I've had four-year-old queens heading my best colonies and I've had first-year queens fail in August. Replacing on age alone means culling good queens on a schedule and keeping bad ones because they're young. What I actually use, decided in autumn: brood pattern in the last full inspection, any drone brood appearing in worker cells, temper, and whether the colony went into winter at weight without excessive feeding. Age is a tiebreaker, not a criterion.