We know, for example, S1, was an early tournament module, was used in Gygax's home game, and was around for a few years before publication. Clearly it gave time to refine in an ever nastier and more creative piece of work. Hickman's early work has similar histories and for Moldvay, some of that has to be coming from his own games...or he just had a very fertile mind.
It probably goes without saying that many latter modules have lacked that time and thought.
That's an interesting point and probably has some merit, though I'm skeptical that that plays a big role because further attempts at refinement have usually failed. For me, adventures like 'Return to the Tomb of Horrors' and 'Expedition to Castle Ravenloft' show that even with a lot of time to reflect, and even with some very talented designers, it's not easy to improve on the originals. Indeed, I'd argue in many cases they show evidence of failing to understand why the original is so good or how to replicate it - a problem I think is more pervasive in the industry than even time crunches and deadlines.
S1 is one of my favorite modules, and I'm very familiar with the text and also with the genera it spawned. There have been many attempts to capture the essentials of S1 and improve on it, but mostly they have failed pretty badly - though Mud Sorcerer's Tomb did come really close and seemed to understand S1 the best. I think it is possible to improve on S1, so I don't actually believe that we've reached the height of the RPG module as art, but it remains to some future designer to initiate that new renaissance.
One of the reasons its hard to improve on S1 is that Gygax really throws the kitchen sink into the module, challenging the player with just about every obvious sort of trap. Gygax loves the kitchen sink approach, and it fails for him elsewhere, but it really paid off here. Unfortunately, that means that he's left very little for anyone else to work with. All the low hanging fruit is in Tomb of Horrors. Mud Sorcerer's Tomb (and in less of an obvious knockoff, The Hidden Shrine of Tamoachan, again proving how good the '1's were in AD&D) hit upon one big area that Gygax overlooked - fluids. But I'm not convinced that Gygax didn't avoid fluids on purpose because he realized how hard situations with fluids are to adjudicate and how poorly the AD&D rules set copes with them. Further, while MST does really capture the fair but lethal aspect of ToH in a way that things like RttToH don't, S1 continues to excel it in that ToH remains the only module I can think of that usually wipes out whole parties of 9th and 10th level characters while at the same time mere 1st level characters could reasonably loot the Tomb successfully through careful play. In fact, the only horrendously unfair encounter of the module at all IMO is the demilich himself, and this mainly because their is no real way a player could be expected to know the abilities and weaknesses of the demilich without recourse to OOC knowledge. Assuming you've snuck a peek at the MM2, a party of quite low 1e explorers with the right equipment could beat the module without loss of life. Thus to me, S1 is an unparalleled test of player skill compared to character skill precisely because it lacks the unavoidable combat and active foes you'd normally expect of a module - especially a Gygaxian module. It's easy to imagine how easily S1 humbled his best players. It was radically and I think consciously like nothing they'd seen before. But it is also in some key ways like nothing since.
Decried as an unfair meat grinder, what actually strikes me about the text is how little effort it is putting into actually killing the characters. If Gygax just wanted to show that he could exercise his fiat as a DM and kill PC's, that would have been a lot easier than what he actually did. It would have been easy to have the mists and devil's mouths and various other features of the tomb leap out and grab anything coming close to them, resulting in a far more lethal module. Instead, it leaves everything fairly passive and then requires you to manipulate them without getting yourself killed. And even then, it isn't as horribly unfair as it could be. The solutions are far from obvious and sometimes counterintuitive, but its usually more than obvious what you shouldn't do even to a novice player. And yet, players eventually get frustrated and do it anyway. Thus, the usual experience of play is, "a) Observe you shouldn't do something, whether enter the mists, pull a lever, or touch something glowy, b) Get stuck, c) Do what you've already know you shouldn't do. d) Die horribly knowing that you just did something extremely boneheaded." Some players get mentally stuck at that, but I found it refreshing compared to dying because the DM rolled high or because you rolled low. One of the biggest flaws with RttToH, which isn't a bad module in its own right, is that it doesn't really get that, thinking instead that having things jump out and grab you improves the design when actually it only improves the lethality.
Or compare with the really terrible 'Grimtooth's Traps' that ToH seems to have inspired along with the mentality that goes with it. Instead of making it obvious what you shouldn't do, the Grimtooth works are all about making doing the obvious thing look entirely innocuous and then gleefully punishing the player for doing it. They seem to forget that any reasonable clever DM can always kill the PCs.
Gygax also manages to do something that would seem to be impossible - build a relationship between the antagonist and the protagonists when the antagonists is a pile of dust lying somewhere. Acererak becomes a reoccurring NPC in the module, always standing beside your shoulder and taunting you. By the time it is over, you know Acererak better than almost any other antagonist in any module - even though you may never encounter him. There are a ton of modern adventure writers that could learn from that. Even the best single module of modern times - Mona's 'Whispering Cairn' - has as one of its few flaws that the final fight is with a pair of nameless mooks that have played zero role in the story to this point. *yawn* (I kid, because the module is awesome, and I've stolen liberally from it.)
