I play 5e every week, and I absolutely love it. So it may seem strange that I simultaneously believe that the 3rd edition (specifically 3.5) of D&D remains the definitive incarnation of the game.
You might think nostalgia's to blame, but in fact I have nostalgia for 2nd edition DragonLance and Planescape. The 3rd edition isn't something I have a longtime emotional connection with, it's just a really good game system.
In fact, as discussions of the "next" D&D have officially begun in late 2021 and early 2022, it occurs to me that D&D 5.5 (or whatever it's to be called) already sounds like it's being pulled back to 3.5. In a way, 3.5 encompasses the natural way of a tabletop roleplaying game. You can start out as a simple game, but eventually situations arise during play that make it apparent that some standard rules are necessary, so you develop more. Next thing you know, you have three rulebooks, a smattering of extra classes and class fixes, adventure-specific feats, and so on.
Here are five reasons I think 3.5 is the correct technical reference point for D&D completeness.
The progression from the earliest editions to AD&D 2nd Edition (what I'll call 2e for brevity) is a bust, because the earliest editions were pretty messy. There was a good reason for that. D&D literally invented the game genre.
Picking up, then, from AD&D, the progression from 2e to 3e was interesting because you could see the abstraction happening. Game elements got tagged, classified, and structured. Flavourful descriptions were adapted to fit rules, rather than the other way round. It was iterative.
From 3e to 4e, the changes were again iterative, even though they were famously drastic. Game elements got refocused and restructured, but the elements themselves were essentially the same. Similarly, 4e to 5e was surprisingly iterative, even though 5e's math is totally different.
I believe that a certain set of rules emerged from 2e to 3e, and these are the core rules of D&D. They're the rules you need for D&D to be D&D. Things like stats affecting skill, a way to resolve complex combat maneuvers, consistency in how magic works and doesn't work, the way armour affects AC and how equipment affects attacks, the way saving throws work, and so on. When an edition omits a rule initially, that rule inevitably makes it way back in.
I default to the Open Game License, and 3.5 is the OGL edition that had it all.
A game is a game because it has rules. As Mark Rosewater observed in a podcast, a game without rules is essentially a toy. There's nothing wrong with a toy, but for me D&D is a game. I understand that it's not a game for all players, and that's an important feature. But for me, the game part of the experience is important, and rules, whether they're arbitrary or carefully extrapolated from flavour, help make the game a fun challenge.
The problem, for me, is when a rule is missing. It's never a problem in the moment. Temporary fixes are a dime a dozen. It's repeat rulings that become problematic. As a DM, I struggle to remember whatever rule I quickly homebrewed to make a ruling, or else I realise that whatever solution I came up with in the moment doesn't work when some new element is introduced into the equation. As a player, I get, possibly unreasonably, annoyed when rulings are inconsistent. I know that in the real world, inconsistency and unfairness is a reasonable expectation. But in my games, I prefer my skill or lack of skill, not confusion over how the simulated world functions, to be the deciding factor in the win or lose condition.
3.5 has rules for everything. You don't have to look that outlier rule up right now. But after the game session, you can do your research, learn the rule, and add it to your arsenal. It's just good world building.
Iterative rule development suggests that all D&D evolves into 3.5 eventually, but 3.5 itself never really ended. Famously, when the D&D brand abandoned the OGL and shifted to 4e, a new brand arose in the form of Pathfinder. Not only did Pathfinder out-sell D&D 4e, but it continued to evolve the 3.5 rule system. Little changes were made to smooth out and improve gameplay. Additional rulebooks got released, most notably:
And those are just the ones I can see on my bookshelf from where I'm typing. There were so many great books that added possibility to the game.
Of course, 3e was released in 2000 so this is arguably a natural evolution for something that's been around for 20 years. Give 5e 20 years and eventually they'd produce more rulebooks to sell (they've already got Xanathar and Tasha, and it's not even 10 years old yet. However, I think there's more to it than that. Paizo is a company that loves to publish. They're the company responsible for the Dragon, Dungeon, and Polyhedron magazines, both of which put out a staggering amount of truly great gaming content. There was a mech system in a Polyhedron issue. Paizo understands that a game is malleable, and that it doesn't detract from your game to publish one or one-hundred books full of optional rules.
Part of the by-product of being released under the Open Game License is that third parties are able to publish content. Astonishing though it may seem, even Paizo wasn't able to have all the ideas during Pathfinder 1e, and so of course third party publishers were able to account for what was missing. At the time, this was a huge advancement for the game industry. It literally did for gaming what open source (its documented inspiration) did for software. It allowed anyone to develop for the rule system, which invited all the ideas.
Some people roll their eyes at this and complain that there's "too much" content as a result. And because of the volume of content, it's pretty easy to dredge up a bunch of examples of trash content. In the worst of cases, there's uninspired fluff created solely to make a few bucks. In some other cases there's content made by amateurs who were still getting the hang of how to make resources for RPGs, but as I'm arguably a member of that very group, I've never been offended by that, and in fact I appreciate fresh ideas.
And there's a lot of really good stuff, too. There's material released by third parties that ends up being better in many ways to official material. To this day, Kobold Press's Southlands source book is one of my favourite RPG worldbooks. It's almost good enough in itself to justify Pathfinder's existence, if Pathfinder needed justification. Southlands is a captivating setting, as fun to read about as it is to play in. And to think, I only bought it because I couldn't find Al-Qadim source books.
Whatever it is, maybe you appreciate the story that's being told, or maybe you find an idea that inspires you to create something of your own, or a rulebook that demonstrates how different aspects of life can be translated into a game world. There's plenty of material for 3.5, making it less a rule system than a rule ecosystem.
Given a complex but complete system, you can extract a simple subset. I don't have enough data on this to claim this as a universal law, but it pops up in a lot of areas of my life. For instance, I'm a KDE desktop user. KDE is famous for having lots of configuration options available. If you see something on your desktop, you can probably right-click it and modify it. Because there are so many knobs and dials, I can essentially implement the much simpler GNOME desktop.
It doesn't work the other way, though. You can't mimic the complex KDE from the simpler GNOME. There just aren't enough options to customize. After all, that's why GNOME is considered to be simple. There's only one way to do any given task, and the user doesn't have much of a say about that.
Similarly, 3.5 has all the rules. Whatever the scenario, you can find a consistent ruling on it. But you're just as free to ignore the rules you consider excess to requirements, instead rolling with advantage or disadvantage to settle disputes, or discussing a ruling as a group, or whatever method of resolution your gaming friends prefer.
There's a danger here, though. A complex system discourages newcomers, because a complex system most likely has a learning curve. I write a fair bit of documentation at work, and it's my firm belief that complexity can be explained simply. The trouble is, very few RPG systems assume that the reader is a novice. Books tend to be verbose, and they do little to reduce the complexity of getting started, with a few significant exceptions: Paizo's Beginner Boxes and D&D's Starter Sets present simplified versions of the basic rules to help newcomers learn how to play.
It's hard to build a 3.5 rule set from a simple 5e base. It's easy to extract a 5e-like system from a 3.5 rule set.
Despite me pitting 3.5 against 5e, I'm not proposing that 3.5 is "better" than 5e. Different isn't a synonym for "better", and I'll continue to play 5e for as long as my players have 5e PHBs. Should I find a group with Pathfinder Core Rulebooks, I'll happily play that. It's all D&D, and I believe the game is unquestionably flexible enough to solve both too much simplicity and too much complexity alike. Conveniently, 3.5 rules often apply, with appropriate modification, to 5e, so you can have your cake and eat it too.
You make a good point at the end, that it is easier to extract a simplified game from a complex one. That is what the OSR was about, and Chaosium's Big Yellow Book that has all BRP rules in one volume. The problem is this is great for those that know how to play and make it difficult for newbies to figure out where to even start. I think a good system needs both. Call one Basic and the other Advanced for nostalgic sake. One to bring folks in, and another providing the options.
On solar day 03.2022 by ruprecht
Except it didn't.
Advanced Dungeons & Dragons - aka 1e - was published in December 1977 (MM), 1978 (PH), and 1979 (DMG), after not only OD&D (1974), but also the first iteration of D&D Basic Set (July 1977).
On solar day 03.2022 by Erik
Ruprecht: Great point. I may amend this post in the interest of clarity, but for now I'll just say here in comments that by "From complexity emerges simplicity" I [also] mean that a complex system can be made simple for beginners. I think the Beginner Boxes from Paizo do this really well, as does the Starter Set from Wizards. I've purchased them both for new players, and the experience has been unilaterally excellent.
Erik: Point taken!
On solar day 03.2022 by Seth