Why “RPG” stopped meaning anything

“RPG” stopped meaning anything because it was never one thing. The label fuses at least five separate properties—a progression mechanic, a fantasy theme, a long single-player format, a particular perspective on a character, and a tabletop idea of playing a role—and for a few decades those properties happened to travel together. Once stats and levelling spread into shooters, sports games and almost everything else, the bundle came apart, and the word was left describing whichever piece the speaker had in mind. The fix is not a better definition of “RPG” but the habit of naming each axis separately.

A word with a tabletop origin

The phrase comes from the table, not the screen. Dungeons & Dragons, published by TSR in 1974 and written by Gary Gygax and Dave Arneson, grew out of miniature wargaming; Jon Peterson’s history Playing at the World traces in detail how a hobby of moving regiments across terrain turned into one of guiding a single adventurer through a dungeon. The shift in scale was the invention. Instead of commanding an army, each player controlled one person, and that person persisted from session to session.

Two things arrived with that shift, and it matters that they are two. The first was a character sheet: ability scores, a class, hit points, experience points, a level. These numbers let a single figure grow more capable over time, and they gave the referee a consistent way to resolve fights and hazards. The second was the role itself. A player at the table could attempt anything they could describe, and a human referee decided what happened. You could bargain with the troll instead of fighting it. You could decide your character was a coward, or a liar, or devout. “Role-playing” named that open-ended act of taking on a part and having a world respond to it.

The numbers were scaffolding for the role. They were not, in the original sense, the point.

What the computer kept

Early computer games in this tradition, such as Richard Garriott’s Ultima (1981) and Sir-Tech’s Wizardry (1981), inherited the tabletop’s dungeons, monsters and character sheets; Matt Barton’s Dungeons and Desktops traces that lineage game by game. What they could not inherit was the referee. A computer cannot listen to an unexpected plan and improvise a fair ruling. What it can do extremely well is arithmetic. So the computer kept the part of tabletop play it could execute, the character sheet, and kept the dungeon, the monsters and the treasure. The open-ended role mostly stayed behind.

This is the first split, and it happened at the very beginning. The computer role-playing game was a translation that preserved the mechanic and lost most of the role. Designers noticed. Garriott’s Ultima IV: Quest of the Avatar (Origin Systems, 1985) is often remembered precisely because it tried to put the role back: the game tracked the player’s conduct against a set of virtues and made becoming a certain kind of person, rather than a sufficiently powerful one, the condition of finishing it. That it was remarkable says something about how far the form had already drifted from the table.

Five things in one word

By the 1990s the label “RPG” carried a whole profile. Pulled apart, that profile has at least five independent axes.

• Mechanic. Statistics, experience points, levels, equipment, and a character who grows stronger through play.
• Theme. Usually fantasy, sometimes science fiction: swords, spells, ruins, a threat to the world.
• Format. A long single-player campaign, often dozens of hours, organized around a story.
• Perspective. Both the camera and the player’s position relative to the character: inhabiting one self-made person, or directing a party of authored ones.
• Role. The tabletop inheritance: deciding who your character is, and having the world take that decision seriously.

Each axis can vary without the others. Final Fantasy VII (Square, 1997) is a canonical RPG on four of them: levels and an equipment system built around Materia, a science-fantasy setting, a long linear campaign, a party seen from a third-person view. On the fifth axis it is close to empty. Cloud Strife’s personality, history and choices are authored; the player can shape his combat build but not who he is. The common shorthand that splits “Japanese” from “Western” RPGs is, underneath the geography, partly a disagreement about this fifth axis.

The Elder Scrolls V: Skyrim (Bethesda Game Studios, 2011) sits differently. Its progression is unusually tied to use—skills improve when you use them, and levels follow from those skill increases—and its first-person perspective puts the player behind one self-made character in an open world. It is very strong on mechanic, theme and perspective. On role, it offers wide freedom of activity, but the world asks less often who your character is than what they are good at.

Disco Elysium (ZA/UM, 2019) inverts the usual profile. It has no conventional combat at all. Its skills are not tools for winning fights but voices in the protagonist’s head that interrupt, argue and mislead, and its checks are resolved with dice in open imitation of the tabletop. Its theme is not fantasy but the politics of a post-revolutionary city. On the progression axis it is modest; on the role axis it is one of the richest games ever made, because nearly every line of dialogue is a decision about what kind of person this detective will be. It has a strong claim to being the most role-playing game in this list, and it shares almost nothing on the surface with the others.

A label works by excluding things. Once every game has experience points, experience points stop telling you anything about a game.

When everything levels up

The decisive break came from outside the tradition. Call of Duty 4: Modern Warfare (Infinity Ward, 2007) attached experience points, ranks and unlockable perks to a competitive online shooter. Players earned XP for kills and objectives, rose through ranks, unlocked weapons and attachments, and chose perks that modified their soldier’s abilities. A Prestige option let them reset their rank to climb again. This is the RPG progression mechanic, transplanted nearly intact into a genre with no story campaign attached to it, no fantasy, no party, and no role.

Nobody called Call of Duty 4 an RPG, and they were right not to. But the move was copied widely, and progression became a default layer rather than a defining feature. Shooters, sports games, racing games and puzzle games acquired levels, skill trees and unlock tracks. Marketing developed a phrase for it—“RPG elements”—which in practice almost always means the mechanic axis and nothing else. It is the same move gamification makes in non-game contexts, taking the most visible layer of a form and attaching it elsewhere; the essay Gamification is not a game follows that thread outside games entirely.

Here is the problem in its simplest form. Classification works by contrast. A genre label is useful because it tells you that this game has something that others lack. When the most distinctive component of the RPG bundle became nearly universal, the label lost its contrast. Saying that a game has RPG elements now tells you about as much as saying it has menus.

Game studies noticed the general version of this problem early. Mark J. P. Wolf argued in 2001 that video game genres should be organized around the kind of interactivity a game involves rather than borrowed wholesale from film’s iconographic categories. Thomas Apperley made a related case in 2006: genre labels inherited from other media describe what a game looks like, while the experience that distinguishes games lies in what the player does. “RPG” is an awkward case for both arguments, because it was never purely iconographic or purely interactive. It was both, plus a format, plus a tradition—which is exactly why it fractured when one of its parts escaped.

Describing games by axes

The alternative is not to abolish the word but to stop asking it to do five jobs. Describe a game along separate axes:

• Genre: the core activity the player performs, the thing a game is mostly made of.
• Mechanic: the specific systems that structure that activity.
• Theme: what the game is about, and the world it represents.
• Perspective: how the player sees, and whom the player is.
• Format: how play is packaged in time and among people.

Run the earlier examples through it and the confusion clears.

• Final Fantasy VII: turn-based party combat and exploration; levels and Materia; science fantasy; third person, an authored protagonist and party; a long linear single-player campaign.
• Skyrim: open-world action adventure; use-based skill progression and perks; Nordic high fantasy; first or third person, one self-made character; very long, open-ended single-player.
• Disco Elysium: dialogue-driven detective adventure; dice-based skill checks and a cabinet of ideas the character can adopt; political crime fiction; isometric, one character shaped by choice; a long single-player narrative.
• Call of Duty 4 (multiplayer): competitive first-person shooter; XP, ranks, perks and unlocks; contemporary military; first person; short matches repeated online.

Each description is longer than three letters. Each is also more useful, because it lets you compare like with like. Skyrim and Call of Duty 4 share a progression mechanic and a first-person view, and that genuinely explains something about why both are absorbing over hundreds of hours. Disco Elysium and Dungeons & Dragons share the role axis, and that explains why the former feels more like the tabletop than games with far more statistics.

The role axis also deserves its own name, because it is the one most easily lost. What Dungeons & Dragons offered was player agency over identity: the ability to decide who your character is and have that decision matter. That is a design property in its own right, and it can appear in games that have no levels at all.

What the word is still good for

None of this means “RPG” should be dropped. Genre labels are not scientific taxonomies; they are names that communities of players, critics and developers use to point at a shared history. The editors of Role-Playing Game Studies begin from the observation that the term has been defined in many incompatible ways across tabletop, live-action and digital traditions, and that the disagreement is itself informative. “RPG” is best read as a genealogy: it tells you that a game descends, by some route, from Dungeons & Dragons and its computer translations.

What it cannot do any longer is describe. To say what a game actually is—what you do in it, what systems shape that activity, what it is about, how you see it, and how it fills your time—you need the axes. The same move applies beyond genre. The companion essay Where the game lives asks a parallel question about games as a whole: not which category a thing belongs to, but which structures are present in it and which are missing. Splitting the bundle is the first step in both cases. A word that means five things means nothing in particular, and a description that names its axes can be argued with.