How many wishes are available in your setting/game? And how many of those are used to increase stats? IME the answers are, in sequence, not many and none.
sigh I knew when I posted that you'd obstinately miss the point. Yes, there aren't so many wishes and in general they aren't used for that. The more salient observation is the system is designed to produce that outcome. This is Gygax saying "no" while pretending to say "yes".
But the point wasn't whether this was a functional and well used means of increasing your stats.
If you spilt out percentile strength into integers then at least it becomes consistent: 10 wishes per point above 16. It also makes the Cavalier's percentile-increment system work consistently with itself and not have most Cavaliers hitting 18.00 by mid-high level, and also allows one to port that same system out to every class rather than just Cavaliers (a move we made instantly, once we saw how the percentile-increment system worked).
I know how it works. Part of my write up of the revised thief was to give them the Cavalier ability to raise attributes, in that case for DEX. I'm likely to do this with a revision of every class that needs an 18 to unlock its full abilities, and really the thief is the least concern there. (And again, this would actually be dragging the game in the direction of 3e regardless of my intention.)
But a more salient point would be that in the roughly 10 years we've been talking about 1st AD&D rules and design, you've never once been able to respond to me without referencing at least one house ruling, and often two or three. You are further diverged from 1e AD&D than I am in these revisions, and less like 1e AD&D in many respects than 2e AD&D is. You've got your own whole system going, and that's fine, I do love a good homebrew, but please seeing as you've wholesale changed almost the entire game over those 45 years you've been playing, don't make the claim then that it "just works for you". Because if it did, you'd be playing something closer to RAW and not something that is as big of a fork as "Chivalry & Sorcery" or other early fantasy heartbreakers. Yes, obviously your aesthetics are different than Ed Simbalist and you are fixing different "problems" to get things where you want them to be, but otherwise you are your own game system that is 1e AD&D only in the inspiration.
I have no idea how you're dealing with the surprise problem, but the very fact that you thought the problem I had with surprise was that it was too lethal is just so telling. As things like the 1988 Dragon article shows, I'm not making up any new post-WotC idea that the system has problems that can be resolved by a recourse to the published rules. This didn't come to me from my lack of experience of "old school" play. This understanding comes from actually having run 1e AD&D which, to be honest, I'm not entirely sure you have ever done.