If you accounted for it, then the rule in fact came up in play, and your statement that it didn't come up in play, that tables could simply ignore the existence of those rules, and that the repeated 20s only apply in weird edge cases is incorrect.
I ran 1e AD&D tables for like 20 years, and I can't remember a single time I was consulting that part of the table.
First, a quote from the DMG. Bottom of page 73:
"Armor class
below 10 is not possible except through cursed items. Armor class
above 2 is easily possible due to magical bonuses and dexterity bonuses." - Gary Gygax
So for all you claiming that I've gotten the terminology backwards, I am using exactly the same conventions used by Gygax in the DMG. It seems I'm the one that remembers how to speak in proper 1e terminology after all.
That is the statement of yours that I initially disagreed with, and all of the dancing around is irrelevant to this minor fact about first edition hit tables, and has involved far more effort than it's worth. If you don't want to think that accounting for the '6 20s' part of the table means that you aren't ignoring that part of the table, nothing I post will change your mind anyway.
And there goes OverlordOcelot with the goal posts on his back zipping right out of the stadium.
By accounting for it, I mean that I said that if someone used THAC0 as their procedure of play, then all they'd need to do was set the THAC0 to 21 and remember that a natural 20 still hit in the unlikely event that their first level thief was suicidally attacking something with 0 or better AC. It's the 1st level thief attacking something with 0 AC that I consider the weird edge case. To consider that statement incorrect, you have to show that opponents with 0 AC were common foes for low level characters. Looked at the MM lately? They weren't. The first time you encounter a 0 AC is likely a Stone Giant or an Anklyosaurus - not things you'll be facing early on.
There was nothing like the uniformly higher progression of AC you'd see in 3.X or later editions. It wasn't that unusual to face things with AC 4 or AC 6 well into high level play. The first time you might encounter an opponent with AC 0, you might already be 5th level, at which point the weirdness with the repeating 20's had been put off to an AC of -2. AC -2 in a monster was fantastically rare in normal play. We're talking things like Xorn, Gold Dragons, and Fiend Lords - stuff you probably aren't going to be fighting at 5th level. Your 10th level thief might face such foes, but now the repeating 20 problem has been pushed off until an AC of -5. THAC0 15 now works perfectly fine except for a few edge cases. Only a few of the harder to hit fiends and Will-o-wisps have an AC that high. You could easily get your PC up to name level without never once throwing a to hit roll against an AC in the repeating 20 range. That goes doubly true if you had a fighter classed character.
Compounding this, I ran - like many tables - that a '1' always missed and a '20' always hit. This slightly changes how I used the table in play, as the number '20' on the table meant in effect - "Either roll a natural 20 or else, roll any number which when modified equals 20 or higher." Likewise a 22 on the table mean, "Either roll a natural 20 or else, roll any number which when modified equals 22 or higher." Which means that if you needed a 20 (or more) to hit, the repeating 20's in the table only matter if your modifier to hit was greater than +0. And since monsters rarely had positive modifiers to hit, this meant that pretty much anything in that part of the table was counting on a natural 20 to hit.
But even this adjustment to the rules doesn't effect the chance of hitting much, because it really only effects attacking AC's beyond the six repeating 20's - which frankly never came up ever. Even to the extent that it was more common for a monster to need a 20 to hit a PC than a PC to hit a monster, fundamentally all the shenanigans with the six repeating 20's ever meant for me was the monster needed a natural 20 to hit. Even when I ran a PC, I played with a different group that also used 'nat 20 always hits' completely without any influence from me, and though my high level PC had a -4 AC, not even a Kobold would have needed a 21 to hit me and so failed with a natural 20 roll. So here my house rule and the RAW produced the same results. And bugbears or ogres, more typical mooks facing my character, could have happily had a -5 adjustment to hit and still hit with a natural 20. So this 'house rule' actually impacted the real chances of success basically never in actual play.
What it does do is simplifies the procedures of play and quite frankly is more coherent than the RAW. For example, this procedure meant that the "Important Note" on page 70 was something I could basically ignore and just use natural math because "it is still possible for opponents to roll natural 20's and thus score hits". I also note that the way the combat rules in this section of the DMG are written, they largely neglect the existence of the "Armor to hit adjustments" on page 34 of the PH, something I wasn't doing.
Finally, again, I didn't actually ever use THAC0 routinely in play because I went straight from 1e to 3e, and because I adopted PC specific tables to speed play (as outlined above). I was familiar with the term. If someone told me that they'd used THAC0 starting in '82, it wouldn't surprise me. Very rarely would have it made any difference compared to consulting the tables.