In another instance of recording bright things people have said in the ACKS discord, Arbrethil had some thoughts in response to Surviving First Level: The Heist Hypothesis and since he still hasn't started a blog (hint hint), I guess I'll record it here.
My players often break your assumptions (hiring lots of henchmen) [ed:to be fair, this is Moldvay's position and I'm just trying to solve it], but I've also definitely seen heist type play happen. And I think when it does, it's often in the wilderness, where treasure hauls are bigger and random encounters are checked less often. Certainly the odds of winning a wilderness fight are worse, but evasion lets you get away most of the time if it's not a fight you want. And if you can find a good lair - stupid ogres that you can distract and bamboozle, or a lone dragon that can't both pursue pesky adventurers and guard its lair - that can be enough to level the thief on its own. The other piece of the puzzle that stands out to me is the utility of a massed spear charge. Even if only three of your five characters can make a spear charge, if you win initiative you've got solid odds of cleaving through a dungeon-sized band of beastmen. Once you can hire a few additional characters, a high AC PC-led dungeon phalanx trumps most any random-encounter sized band of beastmen on dungeon levels 1-2 even without magic support.Emphasis mine. I think the points about spear charges and high-AC phalanxes are good ones in ACKS specifically, though spear-charges can also work against the players if they encounter organized and appropriately-armed opponents. The really interesting point (and one which cuts across both ACKS and OSE/B/X) is that that low-wilderness-level play is often very heist-like. My Bjornaborg game was very treasure-map-centric once it got into the wilderness levels; get to the treasure, kill whatever's guarding it (if anything), and get back to town with it, avoiding encounters whenever possible. And intuitively, it seems like there should be some parallels between low-dungeoneering-level and low-wilderness-level play, in that you don't have all the tools for either yet, under a theory of spell/class design where new abilities tend to be appropriate for the phase of the game where they become available. To gather intel, you don't have Invisibility at 1st, and you don't have Wizard Eye at 5th. To slow pursuit, you don't have Web at 1st, and you don't have Wall of Fire at 5th. So it sort of makes sense that a style of play appropriate to the early-dungeon phase might re-appear in the early-wilderness phase, because you're in a similar lacking-tools situation.
So I think your analysis of them going down to 2nd level pretty much checks out, and wilderness heists are like that but in every way moreso.
This might also have something to do with my players' frustration with low-wilderness play. I have never run 1st level before. They've never played 1st-level OSR games before. So we've never had to "solve" 1st level. Hence, having to learn 1st level's lessons at 5th level instead. But by 5th, you're more invested and the stakes are much higher; if you're learning heist play at 1st and you mess up, oh well, new characters are easy.
I wonder if this is the whole root of the problems I've been having with running wilderness game for literally a decade at this point. It would be pretty funny if for all my theorizing about wilderness as dungeons and hiding maps and resource models and microsandboxes, the real answer was "make sure your players have had to survive 1st level." A simple, practical, culture-of-play thing with unexpected consequences being the answer would be so perfectly on-brand for the OSR that I wouldn't even be mad.
But since, we're here and theorizing - there are some important differences in the resource model between 1st in the dungeon and 5th in the wilderness. Fireball is pretty analogous to sleep - except that often in the wilderness you can regain fireball most days. So that's a tremendous difference in your ability to deal with repeated encounters in a single expedition. On the other hand, there's still some similarity in tackling lairs; one fireball isn't going to win a wilderness humanoid lair fight any more than a single sleep is going to win a dungeon lair fight. And this is what drives the back to the heist dynamic - it's really about inability to take lairs head-on, since that's where the treasure (hence XP) is. On the other hand, the ability to replenish fireballs daily opens up Fabian options for gradual lair reduction during a single expedition not available in the dungeon at 1st level.
The mercenaries-and-hirelings situation also bears examination through this lens. Mercenary troops in the low wilderness levels probably serve about the same function are hirelings at 1st level; you're not going to be able to acquire and finance enough of them of high quality to rely on them alone to take out humanoid lairs, but they can even your odds against humanoid encounters, and at least help hold the line, pin the humanoids, and help prevent you from being overrun. I'm really curious whether Moldvay would push against the use of mercenaries in the low wilderness levels in the same way he pushes against hirelings right out of the gate, so that players learn to do without them rather than using them as a crutch. It seems worth considering to me; we certainly had fights in low-wilderness that my players "cheesed" with massed troops and then I got butthurt and wrote a long post about it. I guess maybe I should go read Expert and see if Moldvay expresses an opinion on this.
In any case - thanks again Arb for pointing this parallel out!