When I ran the World's Largest Dungeon in 3e, there was no way to purchase or craft magic items. Everything was 100% random - whatever they found in the WLD, that's the items they had. And, because I knew what they should have (at least in a ballpark way) at a given level, I could adjust encounters to suit the power of the party.
In 5e, I cannot predict that. If I give a group of 7th level PC's a Rare magic item each, what impact will that have on party power? Well, it ranges from nothing (5 potions of Superior Healing would have zero impact on this group) to massively overpowering the group - the fighter gets a Belt of Hill Giant Strength, the wizard gets a wand of Fireballs, the cleric gets a Staff of Healing, a Cloak of Displacement for the Rogue and a Flame Tongue sword for the Paladin.
Those two parties are now MILES apart in power level despite having what the game tells me are exactly the same value magic items.
1. By "value" here, you mean rarity; as by RAW items in 5e have no monetary value. And rarity is just that: how often these items are expected to appear in the game. It doesn't speak to either usefulness or monetary value, and it's left to the DM to decide - with, one assumes, a view to party power level - which item(s) of that rarity to place if-when something calls for it.
2. 1e-2e did fine without wealth-by-level mechanics and with a much more random selection of items showing up, in large part because the system didn't care as much whether a party was overpowered, underpowered, or somewhere in between. Why is that, and how can 5e get to the same result?
Yup. That’s largely what I’ve been talking about, from another angle.
I think that when a game isn’t open in that way, it becomes very hard for anyone else to design for. Whether that is a 3rd party publisher, a GM trying to homebrew something, or even another designer in-house!
The thread was started with the idea of thinking about the player experience, and I think to do that, you meed to be open about it.
“This is what we’re aiming for. This is how we do that. We made X this why so that Y would happen. If you change A, you’ll want to be careful of B.”
1e-2e did fine without wealth-by-level mechanics and with a much more random selection of items showing up, in large part because the system didn't care as much whether a party was overpowered, underpowered, or somewhere in between. Why is that, and how can 5e get to the same result?
I think this is rather revisionist! The AD&D DMG is full of admonitions against giving away overly powerful items. The Monty Haul game was a real thing.
Carelessness with magic items was absolutely a way to break the game.
I see the disconnect now. RPG design has a lot in common with non-RPG design, but there are areas of nuance that separate them. I think its best to view RPGs in two categories (there can certainly be more). The first is a general application. D&D has firmly moved into this category as a general fantasy RPG. What that means is you get a base kit ruleset that is designed to be loose so it can accommodate multiple experiences. The second, is a bespoke application. This aim is for a specific experience, for example, the Bladerunner RPG that leans into the themes of the film. Bladerunner is not a kit for running whatever sci-fi/cyberpunk/noir experience you want. There are pros and cons to each approach.
Except D&D is still a bespoke game. It's just a bespoke game which avoids telling you what it's bespoke for, because that's apparently upsetting to people...even if they're already doing the things D&D is bespoke for!
Each edition has, of course, colored the overall D&D direction or concept in its own way, pushing it toward one interpretation or another, but by and large it has been intended for one of two general bespoke applications, which I think we can positively entitle "Gygaxian" and "Lancian" (or, if you prefer, "Hickmannish/Weisian"). The "Gygaxian" experience is quite specific, and often captured by what is referred to as "hexcrawl" play today, though actually traversing hex-gridded space is not strictly required for it. The "Lancian" experience is looser, still committed to particular things.
The "Gygaxian" experience is logistics-focused, gritty, casual about story (and often spoken RP) but methodical and precise about tracking time/weight/hirelings/etc. and Exact Words, relatively high lethality but (generally) low volatility*, treating balance as something which occurs only at the level of multiple adventures or entire campaigns (so unusually powerful actions or even sessions are normal), obscurantist rule design (so players who can pierce the obscure presentation to discover the truth behind it feel good and are rewarded for their understanding), and heavy reliance on DM skill in numerous ways. Ironically, despite being rather obscurantist about player-facing rules and such, text regarding how to DM for the "Gygaxian" application was generally pretty up front about a lot of this. The infamous claim, "YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT." Magic items, outside of (deprecated) "Monty Haul" campaigns, are extremely rare and precious things, if they are present at all.
The "Lancian" experience is somewhat tactics-focused, heroic (often disparaged as "superheroic" by fans of grittier experiences), serious about story and RP but generally more casual about "bookkeeping," unreliable for both lethality and volatility* (see the controversy about rules regarding "important" character deaths), treating balance as something both tactical and strategic, an odd contrast of obscurantist presentation but transparent design (meaning it isn't actually that hard to pierce the obscuring veil), and some but generally less reliance on DM skill in favor of already-functional systems. It's best exemplified by 3e.
5e aims to offer something sort of like "Gygaxian" experience, but informed by "Lancian"--there's a very good reason so many people refer to 5e as "AD&D3" without having any particular thing they can point to to explain why. The problem, of course, is that much of its rules structure is in fact actively opposed to the "Gygaxian" experience: it has far too much powerful magic, too much quick healing, too many ways to ignore or even subvert resource tracking, little if anything at the broad, whole-adventure strategic layer to support interesting play, and very weak design with regard to the "adventuring day" (as demonstrated by the need for 5.5e.) It preserves, however, the obscurantism, massive dependence on DM skill, and high lethality and low volatility* of the "Gygaxian" experience. Conversely, it brings in the tactics-focused gameplay, some of the heroic bent (too much for anyone who likes grit, but often not enough for folks who like Big Damn Heroes), and emphasis on personal story and growth over timekeeping/logistics/SOPs/etc. It tries to have its cake and eat it too with regard to magic items, but has pretty clearly ended up on the "Gygaxian" side, where magic items are extremely rare, with the "magic items are optional" concept having been interpreted as "magic items shouldn't be given out in most cases."
All three--"Gygaxian," "Lancian," and 5e--are very heavy on combat vs. all other modes of interacting with the world and, naturally, feature the classic D&D tropes like six stats, chunky levels, unrealistic training methods or processes, the quirk that defense protects you from attack, not from damage (hence the common "AC becomes DR" rules hacks/homebrew/etc.), and various other mechanical elements that have been present in every edition of D&D, even 4e.
5e is rather poor at supporting anything outside of the above, unless the DM effectively designs a new bolt-on ruleset to support it. This is both a function of the rules themselves (they're already somewhat rickety and need frequent DM attention just to function smoothly), and of the extreme dearth of useful, productive DM guidance in the book allegedly there to "guide" them. 5e was a system written for old hands, not for helping people get into it.
*The lethality/volatility distinction came from discussing various editions on another form. TL;DR: lethality = PC death chance, volatility = chance of dramatic status change.
"Lethality" is a loose description of how likely actual character deaths are, while "volatility" describes how much character status changes from one round or even one turn to the next. 4e is a mid-to-low lethality, high-volatility game basically the whole way through. 3e starts off relatively low on both, but becomes relatively high in both at high levels, what is often called "rocket tag." 5e is often considered more "old school" than other WotC editions because it is relatively high in lethality (especially for the first few levels) but relatively low in volatilty.
But...that has essentially nothing to do with the rules of the game. You could say exactly the same thing about any system: design good adventures of different types, and suddenly any game can offer a "general experience."
Adventures need to make use of the game's rules, but adventure design as a whole is genuinely a separate topic from system design. Hence why using a very well-made system cannot protect you from making $#¡+ modules (just look at 4e's Keep on the Shadowfell or Pyramid of Shadows, which were both absolutely awful), while a system generally recognized for extremely flawed design does not prevent you from making excellent adventures (just look at 3e's Red Hand of Doom, or the numerous beloved APs for PF1e, which even its own creators admitted was too broken to continue developing for.)
If D&D is bouyed by adventure design, all that that says is that it's had authors who could write good adventures--and enough market share to make writing such adventures actually worthwhile.
And most likely not even that! Because what makes magic items feel wondrous often has absolutely nothing to do with how powerful or uncommon they are in-game. It has everything to do with three things: (1) how fresh and new they are to the player, (2) how valuable the magic item is to them personally (e.g. utility, symbolism, aesthetics, marketability, etc.), and (3) the manner in which the item was earned/acquired.
You make magic items feel magical again, not by making boring +N items essentially unknown, but by not using boring items. Anyone can make a boring +2 Flaming Sword of Swiftness. It takes an actual DM, using the most important part of DM skill, caring about player enthusiasm, to create something like...
"As you turn the finely-wrought lacquer-and-nacre scabbard in your hand, you realize that this must be an Arkhosian High Blademaster's sword. Drawing it, you see the curved blade within, austere in its perfection. The handle bears gold and mythril embroidery of cranes, the cross-guard lilies in crimson and gold, implying this sword once belonged to someone of a cadet branch of the royal family. After a few minutes of trial and error, an Arkhosian word that loosely translates as 'for blood and honor' sets the sword alight, the color and intensity clearly marking it as dragon-fire. And that leads to a second realization: though its greatest magic has faded slightly, it retains some final mote of the Golden One's flame, still burning after all these centuries. Karthedaas of clan Nyax, you hold in your hands one of the few treasures of Lost Arkhosia that remain, dimmed but not dulled by the turning of the ages--and now it is yours, to bring honor to your name and your clan. Or perhaps that unquenched spark may bring forth Arkhosia's light again, at your hand? You are not the first to have such thoughts...but none had the strength, or tools, or allies you have at your disposal."
That is how you make a magic item wondrous. You make it matter. You give it weight and meaning.
Something as simple as a leather duster that acts like armor, or a hat of disguise, can be a beloved and precious thing, to the point that losing it marks a ruthlessly determined quest to get it back, whatever the cost. Something incredibly powerful but nondescript and flatly presented can be a nothingburger largely ignored.
The magic of the story is in the telling, not in the numbers.
Fully agreed. Sadly, opaque design is quite popular, in part because people so readily confuse "X is designed to do Y well" with "X cannot do anything other than Y," and thence transform that into, "Because X is designed to do Y well, if you like stuff that isn't Y, you aren't welcome here." Thus we get intentional obscurantism; if you avoid ever explaining yourself, no one can (almost always mistakenly) feel like you've told them your game excludes them.
I think this is rather revisionist! The AD&D DMG is full of admonitions against giving away overly powerful items. The Monty Haul game was a real thing.
If one went nuts, sure. But the game didn't break if your party had no* or nearly-no magic, nor did it break if your party had gone through a few classic modules and scooped all the rather-abundant loot from those. There was a wide "what/ how many magic items does the party have?" window within which the game worked fine; significantly wider than 3e I think.
Magic items were also much more easy come easy go in 1e, given that some bad luck with AoE saves could strip you clean in a hurry. 3e largely did away with this (I think if you rolled a nat. 1 vs AoE then one of your items had to save, or something like that), and 4e and 5e have no item destruction that I can recall.
* - unless the DM was cruel and threw in opponents that required magic weapons to hit.
Each edition has, of course, colored the overall D&D direction or concept in its own way, pushing it toward one interpretation or another, but by and large it has been intended for one of two general bespoke applications, which I think we can positively entitle "Gygaxian" and "Lancian" (or, if you prefer, "Hickmannish/Weisian"). The "Gygaxian" experience is quite specific, and often captured by what is referred to as "hexcrawl" play today, though actually traversing hex-gridded space is not strictly required for it. The "Lancian" experience is looser, still committed to particular things.
The "Gygaxian" experience is logistics-focused, gritty, casual about story (and often spoken RP) but methodical and precise about tracking time/weight/hirelings/etc. and Exact Words, relatively high lethality but (generally) low volatility*, treating balance as something which occurs only at the level of multiple adventures or entire campaigns (so unusually powerful actions or even sessions are normal), obscurantist rule design (so players who can pierce the obscure presentation to discover the truth behind it feel good and are rewarded for their understanding), and heavy reliance on DM skill in numerous ways. Ironically, despite being rather obscurantist about player-facing rules and such, text regarding how to DM for the "Gygaxian" application was generally pretty up front about a lot of this. The infamous claim, "YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT." Magic items, outside of (deprecated) "Monty Haul" campaigns, are extremely rare and precious things, if they are present at all.
The "Lancian" experience is somewhat tactics-focused, heroic (often disparaged as "superheroic" by fans of grittier experiences), serious about story and RP but generally more casual about "bookkeeping," unreliable for both lethality and volatility* (see the controversy about rules regarding "important" character deaths), treating balance as something both tactical and strategic, an odd contrast of obscurantist presentation but transparent design (meaning it isn't actually that hard to pierce the obscuring veil), and some but generally less reliance on DM skill in favor of already-functional systems. It's best exemplified by 3e.
Not bad until the very last word - did you mean 2e? 3e doesn't follow this pattern in that it's quite reliably lethal, fairly gritty, and does rely on a certain amount of bookkeeping etc. (though you're recording different things than in 1e).
*The lethality/volatility distinction came from discussing various editions on another form. TL;DR: lethality = PC death chance, volatility = chance of dramatic status change.
"Lethality" is a loose description of how likely actual character deaths are, while "volatility" describes how much character status changes from one round or even one turn to the next. 4e is a mid-to-low lethality, high-volatility game basically the whole way through. 3e starts off relatively low on both, but becomes relatively high in both at high levels, what is often called "rocket tag." 5e is often considered more "old school" than other WotC editions because it is relatively high in lethality (especially for the first few levels) but relatively low in volatilty.
I disagree with the level of volatility you suggest in both 1e and 3e, if in "character status change" you include non-death bad things. Both 1e and 3e (sort of) had level drain, poison, and a host of other effects that 5e either lumps under hit point damage or eschews completely; and in this way I'd say both are quite volatile. The one difference with 5e (and 4e?) is that most of those effects can be shrugged off much faster than in 3e or earlier.
Also, I'd say 3e is mid-to-high lethality all the way through, right from 1st level.
1. By "value" here, you mean rarity; as by RAW items in 5e have no monetary value. And rarity is just that: how often these items are expected to appear in the game. It doesn't speak to either usefulness or monetary value, and it's left to the DM to decide - with, one assumes, a view to party power level - which item(s) of that rarity to place if-when something calls for it.
2. 1e-2e did fine without wealth-by-level mechanics and with a much more random selection of items showing up, in large part because the system didn't care as much whether a party was overpowered, underpowered, or somewhere in between. Why is that, and how can 5e get to the same result?
5E gives a gp value range for rarity. The Downtime rules in Xanathar's uses those values for buying magic items. It's not like there's no GP values in 5e magic items.
And, I would argue that 1e absolutely did have wealth by level. But, it was largely opaque. You were expected to be absolutely dripping in magic items at fairly low levels, compared to later editions. The paladin being limited to a mere 10 magic items was meant to be a heavy restriction. In a group of 6-8 PC's, you were expected to have close to a hundred magic items, fairly quickly. And the modules certainly supported this as did the random treasure generation tables where you could easily get 4-6 magic items from a single lair.
But, as @pemerton rightly points out, the whole "Monte Haul" thing in the DMG and many, many sources throughout 1e and 2e meant that the system absolutely DID care if the party was overpowered.
And most likely not even that! Because what makes magic items feel wondrous often has absolutely nothing to do with how powerful or uncommon they are in-game. It has everything to do with three things: (1) how fresh and new they are to the player, (2) how valuable the magic item is to them personally (e.g. utility, symbolism, aesthetics, marketability, etc.), and (3) the manner in which the item was earned/acquired.
You make magic items feel magical again, not by making boring +N items essentially unknown, but by not using boring items. Anyone can make a boring +2 Flaming Sword of Swiftness. It takes an actual DM, using the most important part of DM skill, caring about player enthusiasm, to create something like...
"As you turn the finely-wrought lacquer-and-nacre scabbard in your hand, you realize that this must be an Arkhosian High Blademaster's sword. Drawing it, you see the curved blade within, austere in its perfection. The handle bears gold and mythril embroidery of cranes, the cross-guard lilies in crimson and gold, implying this sword once belonged to someone of a cadet branch of the royal family. After a few minutes of trial and error, an Arkhosian word that loosely translates as 'for blood and honor' sets the sword alight, the color and intensity clearly marking it as dragon-fire. And that leads to a second realization: though its greatest magic has faded slightly, it retains some final mote of the Golden One's flame, still burning after all these centuries. Karthedaas of clan Nyax, you hold in your hands one of the few treasures of Lost Arkhosia that remain, dimmed but not dulled by the turning of the ages--and now it is yours, to bring honor to your name and your clan. Or perhaps that unquenched spark may bring forth Arkhosia's light again, at your hand? You are not the first to have such thoughts...but none had the strength, or tools, or allies you have at your disposal."
That is how you make a magic item wondrous. You make it matter. You give it weight and meaning.
Something as simple as a leather duster that acts like armor, or a hat of disguise, can be a beloved and precious thing, to the point that losing it marks a ruthlessly determined quest to get it back, whatever the cost. Something incredibly powerful but nondescript and flatly presented can be a nothingburger largely ignored.
Unusual items like the feather duster of protection certainly carry a cachet with them and are often popular, but I rarely if ever hear anyone complain on finding an incredibly powerful "nothingburger".
If one went nuts, sure. But the game didn't break if your party had no* or nearly-no magic, nor did it break if your party had gone through a few classic modules and scooped all the rather-abundant loot from those. There was a wide "what/ how many magic items does the party have?" window within which the game worked fine; significantly wider than 3e I think.
Magic items were also much more easy come easy go in 1e, given that some bad luck with AoE saves could strip you clean in a hurry. 3e largely did away with this (I think if you rolled a nat. 1 vs AoE then one of your items had to save, or something like that), and 4e and 5e have no item destruction that I can recall.
* - unless the DM was cruel and threw in opponents that required magic weapons to hit.
Again, though, it depended so much on what those magic items were.
Magic armor and shield? Guess what, you're never failing another save vs AOE again. After all, those bonuses apply to your saving throw. The odds of losing a magic item when you've got a +2 suit of armor and a +3 shield meant that you virtually never lost items. Never minding that now your AC was in the mid negatives, meaning virtually nothing could hit you because they needed an 18 or better.
So on and so forth. I find the whole "Well, you're going to lose all these items" thing to be very overblown. Number one, very, very few monsters in 1e actually HAD an AOE attack that would trigger saving throws. Number two, those saving throws were generally pretty easy to make anyway. I'm sure it happened from time to time, but, most of the time? Nope. You just kept accumulating more and more and more magic items. Negative 7 or 8 AC. Girdle of Giant Strength meaning that you virtually never missed an attack and obliterated most things you hit in a round or two. On and on and on.
