Designs & Decisions: Why Old School Score Generation Works

A subreddit I frequent has hosted a spate of memes lately about the old 3d6 down-the-line method of generating ability scores, alternatively looking fondly on a relic of a distant past and skewering both the idea and the conversations generated around it. Under these posts I tend to join the small but extant group defending the practice. My reasoning tends to focus on seeing the system not just as "something we used to do" (I didn't play D&D until it was already a pop culture phenomenon in the late 2010s) but as a valid design decision in its own right. 

Mechanics contribute the most to a type of game and the style of play that embraces the features implied by those mechanics. This is something often missed among faithful 5e fans, as modern D&D's mechanics do not contribute well to any specific type of game and does not foster any specific style of play. It is, however, the core argument behind the mantra that "system matters," a mantra repeated most often by those with experience in multiple rule sets or with an interest in game design. 3d6 DTL was used for so long, and continues to generate enthusiastic conversations, because it did just that: contributed to making the game a survival-horror dungeon crawl simulator while fostering a type of play that focused on exploration, player skill, and treasure hunting. 

The biggest complaint about the method is that a player who wanted to play a wizard may be stuck with a 3 in intelligence, or some similar combination of desire and dice that cannot be realized. This is a valid complaint in a game about character building, which is what 5e tries more than anything else to be. Starting with 3e and it's myriad character skills, feats, prestige classes, and utterly broken combinations, that character building was conceived as a kind of mini game with a focus on designing a mechanical marvel capable of committing 14 attacks of opportunity (I have read this was a real possibility with a fighter build in 3.5, but I will confess to not having checked). With the success of Critical Role, that character building is now conceived as a theatrical progression of personality, sometimes within a linear plot, sometimes within open-ended emergent play, but always as the object of narrative development. The idea of narrative development through theatrical progression of personality has become so widespread that the definition of "role-playing game" is actively changing to mean a game where one plays a theatrical role. And yes, in a game designed for this kind of experience, 3d6 DTL is a bad mechanic, as it violates the (supposed, in the case of 5e) design goals and the expectations of everyone at the table. 

For the first three decades of Dungeons and Dragons, the idea of "character building" was not the design goal or expectation of the game. Instead, it was intended to be a simulation in which a random representative sample of the world is selected to secure treasure and experience adventure by surviving a harrowing series of death traps and murderous monstrosities in an enclosed underground environment. It was about dungeon crawling. 3d6 DTL randomly assigns a character a place on the bell curve in each of the 6 areas of human/oid prowess deemed relevant to the game - it simulates the local population. The random generation of hit points within a low band (an even lower one if you rolled low on Constitution) was a realistic translation of an untested peasant's or dottering wizard's fragility. The high rate of character death resulting from these mechanics is intentional - a feature, not a bug. If most level 1 commoners could walk into a dungeon and expect to return then it's not a survival game. If the monsters do not inspire actual fear of losing the first persona at the top of your character sheet stack, it is not a horror game. If the local village can produce 3 or 4 Paladins to venture into the caves by virtue of the players rearranging scores as needed or ignoring ability score requirements, it is not a simulation.

All this comes with some caveats. The idea of 3d6 DTL as the primary method did not last long even in the days of TSR; in the first edition Dungeon Master's Guide (or as I like to call it, The Bible) Gygax himself already recommends against it. In the Basic rules, which recommended it through the 90s, it was common to set bounds for how bad a character could be before a new one was rolled. Personally, I do not even prefer it as the score generation method for all my games. My current main campaign used 4d6 DTL because, while Arden Vul is challenging, I do not intend for it to be survival horror the way my last one-shot was. It is more exploration focused, with some survival elements. When the physical books come out for Dolmenwood I will be using an extremely generous system because I intend that to be a more character-focused campaign. But this is my core thesis - not that 3d6 DTL is the best method but that it enhances the explicitly stated genre of the editions that use it. It should be seen as a valuable and relevant way to create a certain type of experience at the table, instead of a mere archeological artifact or a mistake we evolved away from.