Part 5 of a series:
As with reading Classic Traveller, this was a very interesting exercise.
A lot of things were just not covered at all. The turn structure in combat comes entirely from Chainmail; man-to-man combat gets a single paragraph which is just a reference. Which leads to another (terrible?) idea for D&D - secret, written orders for your character in D&D combat. Using Chainmail for combat does explain the origin of initiative by side rather than individual initiative, with movement, then shooting, then melee, then spellcasting, as seen in Basic - that's about what a Chainmail turn would look like under the non-written-orders option.
Recovering spells is also never really covered. It's perfectly possible given the text here that you might only need your spellbook to prepare / choose spells at the beginning of the adventure, and then you get them back every day, but to swap them out you need the book again. And that would also free wizards to go on naval adventures, or swimming, or out in the rain, without worrying about their precious tomes.
I like the idea that counter-spelling is an action that MUs can attempt at any level. I guess the flow for this under something like ACKS' initiative system might be "declare spells, declare counterspells, if spell-caster beats counter-speller's initiative then it goes off, otherwise counterspeller gets a roll adjusted by level to stop the spell from going off." This would provide low-level parties with another way to stop sleep by an NPC caster from being a TPK. Apply modifiers for relative power of the two casters, whether the counter-speller knows the spell being countered, maybe even make high-level spells easier to counter (they are higher "complexity", which sounds like they'd be easier to throw wrenches into).
I like Chainmail's roll-to-cast with the possibility that the spell goes off on the next round, too. That would also offer a bigger window to counterspell or get stabbed. I'm not sure whether counterspelling should cause you to lose the spell for the day / the slot, or just to prevent you from casting it that round (or for the rest of that fight / 10 minutes).
It's weird that the "alternate" combat system won out so completely. I guess if you didn't have Chainmail, then the alternate was the only option available to you. I'm really surprised that fighters gain THAC0 so infrequently given that they do gain capability every level under the Chainmail combat system.
I almost want to see a TSR D&D variant where THAC0 doesn't improve as you level - you just get more attacks at a rate that approximates the fighter's ability to fight massed troops in Chainmail (and clerics and MUs get more attacks at a slower rate). It isn't even crazy for a 10th level fighter to make 10 attacks in a 1-minute combat round. The problem would be building an initiative system where fights between equally-matched opponents didn't just come down to "whoever won init" (although I guess wizard-fights are already like that). Speed in play would also suffer.
It's odd to think that Codex Martialis for 3e was sort of on this track back to Chainmail.
I guess the fact that non-Chainmail combat won out might also explain the disappearance of counterspelling. And morale only being in mass combat part of the Chainmail book, not in the LBBs, might have been the beginning of morale's status as "that rule everyone forgets".
As others have noted, the implied setting here is suuuper low population density. Securing a fortress is sufficient to control a 20-mile radius around it - about 540 square miles. In that 540 square miles you get 2d4 villages, each of 1d4x100 villagers, so a total population of 1250 on average. Spread across 540 square miles, that's a population density of about 2.3 people per square mile, which is too dense for hunter-gatherer populations for also about 5% of the density of medieval England. Having that population per 5-mile hex would bring population density up to 60 people per square mile, which is in line with ancient Mediterranean population densities but still well short of the population density of medieval France or the Holy Roman Empire. I almost wonder if the conventional wisdom, that population density in OD&D is really low, is the result of the textual ambiguity around "territory", which could refer to either the entire barony containing 5 villages, or to each hex containing 5 villages. Another possibility for making the numbers more reasonable would be to rule that the village population is "urban" and supported by a larger rural population; with 9 rural per urban, you'd end up at 23 people per square mile which is still thin but not crazy thin.