I don't have much of an opinion on this, I think the question you're asking is really more along the lines of at what point is the body of work sufficient from a handicapper's perspective and includes enough data points to draw reliable conclusions from it.
In that case I think you really need X number of races in the record rather than focusing on age, a five year old with three lifetime starts probably doesn't tell you much in the way of preferences yet.
The other way to look at the question is when does the horse generally reach let's say a career top and possibly even at what age is the average horse maybe most consistent in its level of effort from race to race.
I could research it with the class ratings but I would probably be looking to only include the subset of horses with very long careers. In such a study then I would have no use for any horse that retired at 3 or 4 for example because including those horses wouldn't tell me anything at all about the optimum age for a racehorse to compete at its best.
In any case I'm not sure a study like that would tell me much, if (pick a number here) age 7 is 'best' but a horse has many nagging injuries then 7 could be best for the mind but not this particular horse's body. It's too much of an individual thing IMO, but apparently Benter cared about age in his model so maybe you're on to something. Good luck.