You completely missed the point. Stats (3d6 vs 4d6) are going to average slightly higher with 4d6, and 4d6 was common back in 1E/2E also.
The point was that you could get away with lower stats in 1E/2E much easier because the bonuses were almost all smaller (except 18__ damage bonuses) and saves were on the chart, not stat-based. Only a few subclasses were MAD rather than all. The game math, while clunky, wasn't as dependent on high attributes. In other words, unless you were a stickler for encumberance, there was basically no difference between a 9-strength fighter and a 16-strength fighter.
As for Level Limits, which has no bearing on this point anyway, who actually used those? They were probably the single most ignored rule in the game.
It seems that you've at least one of my points, too: when comparing systems, RAW, the rules do matter, even when houseruling was common.
RAW, it was much more difficult to create the character you wanted in 1E/2E than it is in 5E.
Stat generation was defaulted to 3d6 in order in both 1E and 2E. Regardless of whether it was common to use one of the optional generation methods, or house-rule other options, that was the system rule, as seen in tournament play and otherwise. And I acknowledged in my prior response that yes, the 4d6 generation did offer a higher yield of scores, by approximately 1.75 points. The issue is completely with the default generation of 3d6, and the lack of genuine impact of higher or lower scores within the spread (excepting extreme ends of the spectrum), except in terms of limiting character options. With the exception of multiclassing, 5e doesn't have score requirements for characters - and sure, you may be less effective, depending on score placement, but standard assignable array is the default, followed by assignable 4d6 generation -- players have a high level of control over how effective the characters are, and a few low scores aren't going to prevent a character option outright.
Sure there was little difference between a 9 Strength and a 16 Strength at face value, when looking at hit and damage, but Ability Scores also arbitrarily meant different things in different situations (encumbrance aside). They were the metric against which non-combat tasks were attempted - direct d20 against the score - someone with a 16 Strength had a 35% higher likelihood of achieving the same task a 9 Strength PC attempted - in this instance it was way more swingy and the higher score held much more meaning. Sure, the 16 didn't have much more impact on to hit and damage, but it did impact other things in combat, especially where level limits are concerned. A PC elf fighter stuck at level 5, with either a 9 or 16 Strength, while the rest of the party had hit level 9 or higher, had, on average, 22 less hit points than than any other fighter in the party. One could argue that the elf player wasn't playing a character suitable for the group, to consider appropriate levels for a given challenge, but then, that player wouldn't be playing the character he or she preferred, either.
And in regards to subclasses and ability scores - with the exception of the core four classes, all the other classes, considered subclasses in 1E, required multiple scores in excess of 12, with some requiring 15 or higher.
Don't get me wrong - I loved playing the earlier editions, and would still play them, given the opportunity. I just feel that more choices are available and the rules don't outright bar most choices, and bounded accuracy ensures that all characters are effective, though some more than others.