I'd like to call everyone's attention to the on-going study on aging pilots are Stanford. It would seem that the top pilots keep on trucking, while the bottom pilots really lose their ability.
There have been numerous studies of supposed cognitive decline with age, and they have generally shown that people who are bright in their youth tend to remain bright for a lifetime, whereas people who are marginal in their youth deteriorate significantly with age.
So a good pilot is likely to remain a good pilot throughout his life, even in old age. But someone who is just barely squeaking by in his early flying days may deteriorate to well below the minimum competence level required to fly as he gets old.
To some extent the evolution of cognitive ability with age seems to depend on how much it is used … the more you use your brain, the better the shape it will stay in as you get older (so if you fly and you want to continue flying, be sure that you fly a lot!). However, the ability one starts out with also seems to predict its evolution over time. For example, if the threshold of acceptable performance is 100, then a pilot at 300 at age 30 might still be at 250 by age 80, but a pilot who is at 105 at age 30 might be at only 20 by age 80.
Another factor concerns reflexes substituting for brains. A young person might count on fast reflexes to help him escape the consequences of his mistakes; but as he gets older, the reflexes slow and eventually his mistakes catch up with him. A young person who makes few mistakes to begin with, however, doesn't need reflexes to compensate for them, and will still be avoiding mistakes in old age.
What all of this means is that age is not a valid criterion of pilot evaluation at the individual level. If you look at large groups of pilots, you may see a slight decline in performance with age, but that has absolutely no predictive value when applied to individuals. For individuals, the only way to measure performance is to test it, irrespective of age. That's because individual variation in this case is vastly greater than age-related changes.