On Incentive

Or, In Defense of the Experience Point

In the OSR and OSR-adjacent RPG communities, I’ve encountered a rather baffling trend: the opinionated rejection of XP and XP-based progression systems. Even in new-school D&D, XP have been going out of fashion for a while now as a simple consequence of the shifting priorities of the playerbase—the typical 5e player has no use for XP, so they prefer milestones instead. But to many an OSR gamer, XP is Bad Design, and they have theory to prove it! The chief contention seems to be this:

XP are a system-level mechanic that imposes an extrinsic incentive upon the player, and this is inherently a bad thing.

In a trivial sense, the first point here is correct: by awarding XP for certain things, you do, in fact, encourage players to do those things. But the claim that this is a bad thing makes major assumptions, both about the implementation of the mechanic itself and about the entire rest of the game in which it is being implemented. What is it that’s being incentivized, and in what context? Who decides what’s being incentivized, and how are they making that decision? Without answering those questions, it cannot be said whether XP—or any other “reward mechanic”—is good or bad design.

To prove I’m not tilting at strawmen here, here’s a specific example: “Against Incentive” by Luke Gearing, who is personally notable as the designer of Wolves Upon the Coast, an innovative OSR system with a fascinating alternative to XP for goal-based advancement. The essay is three years old by now, but it’s what people have been linking at me when I ask what’s so bad about XP, so between that and the popularity of Wolves, I’m assuming it’s still relevant.

Despite All My Rage

Throughout the essay, Gearing consistently frames the concept of incentive in behavioristic terms. He never uses the words “Skinner box”, but he does show an illustration of one: a rat in a cage with a food dispenser on the wall to incentivize doing the “right” thing and an electrified grid in the floor to disincentivize doing the “wrong” thing. The GM is the researcher, the food and the electric shock are the incentives, and the player is the poor little rat.

The analogy is far from baseless. Roleplaying gaming has a regrettably long history of GMs treating their players like animals to be trained or children to be disciplined, using in-game rewards and punishments to reinforce desired behavior. It’s is a socially fraught, dubiously effective, and fundamentally misguided approach to GMing, and I don’t like it one bit.

If this point—“Behaviorism bad; don’t treat players like lab rats; talk to them like adults, you dork!”—were the only one that Gearing was trying to make here, I would have no argument with him. Unfortunately, every other point in his essay takes one thing or another for granted, leading him to some pretty myopic conclusions. I’ll limit my critique to a strictly OSR perspective—for example, I’ll take it as a given that it’s more important that players have agency than for them to experience some epic story the GM has planned, even though I know there are entire cultures of play where this is not the default—but even then, Gearing’s sweeping generalizations fall apart under scrutiny. (This means that if you’re not from the OSR, it may sound like I’m talking smack about play practices that are perfectly normal where you come from—but if they work for you, they work for you, and that’s OK! Different playstyles require different approaches.)

The Point of the Game

Although Gearing’s essay is entitled “Against Incentive”, the only concrete example of incentive he gives is XP-based advancement. He argues that, by gating advancement behind a specific objective—gold, for instance—you encourage players to pursue that objective regardless of whether they actually want it. They’re not after the gold; they’re after the XP. Thus, the acquisition of gold arbitrarily becomes “the point” of the game, when the game could instead be about goals that the players choose for themselves.

Though I do see value in letting players set their own goals, it’s not obvious to me that every game has to be like that. Consider the classic, basic campaign format in D&D—“basic” not as in Holmes or B/X or BECMI, but in the sense that the scope of play is contained entirely within the dungeon. No wilderness to explore, no NPCs except your hirelings, nothing at all but you and your party versus the dungeon. Why have you come to the dungeon? To slay monsters and hunt for treasure, of course! This was the original context of D&D’s XP mechanics: they were a scorekeeping device. The game was not about slaying and looting because that was what earned XP; slaying and looting scored XP because that was what the game was about.

In the context of D&D’s wargaming roots, there is nothing unusual about this design. A wargame takes place in a constructed scenario where players pursue set objectives, often with a system of points to quantify the achievement of those objectives. The objectives are an integral part of the scenario; without objectives, you have no conflict, and without conflict, you have no wargame. Basic D&D (again in the sense of campaign structure, regardless of edition) is a wargame where the scenario is a dungeon expedition and the objective is to prove your worth as an adventurer by overcoming powerful foes and returning alive with the spoils of your victories. And since it’s a refereed, free-maneuver wargame, you have tremendous agency in how to approach that objective: instead of being limited to a set of options enumerated in the rules, you can try any strategy you can think of! Despite its narrow scope, the basic dungeon campaign is a perfectly cromulent game, and—ethical concerns about murdering innocent goblins for their belongings aside—it holds up surprisingly well even today.

The Experience of the Game

As we see, Gearing’s arguments framing XP advancement as always and only a form of incentive get a bit shaky when applied to XP as a scoring device in the basic dungeon campaign. These XP hardly create “optimal choices” that “once discovered, become boring”, they don’t constrain agency in any way beyond what the scope of the game already does, and the advancement you get from them is a “reward” only in the sense of a prize for winning at the game—a very narrow, focused game explicitly about a specific core activity that the players should find rewarding in its own right (or else should be playing some other game).

Yet it cannot be denied that many GMs do use XP in precisely the way Gearing cautions us against, with all of the negative consequences he describes. How did we end up here? How did we get from XP-as-score to XP-as-carrot-on-stick? It’s difficult to say exactly when the change happened. The textual history of the rules reveals only so much, since the change was primarily one of play practice rather than mechanical formalism, but we do see hints of XP being used this way at least as early as AD&D with its penalties for acting against one’s alignment—likely an artifact of Gygax’s temperamental shift from his earlier open-minded permissiveness to his later prescriptive dogmatism about the One True Way to play and enjoy the D&D® game. Meanwhile, as RPGs spread and found an audience beyond the core wargaming scene, the Trad culture of play emerged, quickly developing a diverse array of tricks and and techniques for using games as a vehicle for storytelling and self-expression.

It is in this context that we see XP go from a simple scoring device to a tool for behavioral manipulation. The change is but one part of a radical transformation that some historiographies of the game characterize as the end of Old-School D&D and the beginning of a Middle School of D&D. The basic dungeon crawl gave way to a near universal adoption of the extended campaign format, with a world to explore and interact with outside of the dungeon and many things to do other than slay and loot. At the same time, the role of the GM shifted from that of a neutral wargaming referee to one of an entertainer responsible for delivering a specific experience—an engaging story if you’re Trad, or the glamorized ideal of an official, on-brand ADVANCED DUNGEONS & DRAGONS™ if you’re Gygax. And whatever form that experience took, receiving it became the new point of the game.

I believe that the emergence of XP-as-incentive was the result of tensions between these two trends: the broader the game’s scope, the more freedom players had, yet the more freedom they had, the harder it was for the GM to deliver a specific intended experience. It makes good sense, then, that XP, shorn of its original meaning by the shift away from the highly gamist dungeon hack-and-slash, would evolve into an incentive used to goad players into cooperation.

Risk vs. Reward

Gearing hints at the connection between incentivization and the legacy of Middle-School GMing when he says that “RPGs can be more than Rube Goldberg machines culminating in your intended experience” (emphasis mine). High-agency gaming, he argues, is rewarding in a way that no low-agency experience, no matter how ingeniously designed or expertly conveyed, can imitate. On this point, I am with him. True player agency is uniquely rewarding—not just for the players, but for the GM, too! Yet Gearing is mistaken in his conflation of reward mechanics with incentive. Used well, reward mechanics can actually complement player agency, and in a game that is low-agency already, removing reward mechanics isn’t going to be the +1 magic bullet.

In the OSR in particular, the concept of reward in general is an important part of the game because risk–reward analysis is a major element of player skill. Suppose your GM says this:

Before you stands a massive troll, wielding his club menacingly. Behind him lies a similarly massive hoard of treasure. What do you do?

In some playstyles, that prompt may be interpreted to mean, “How will you interact with the troll?”—the assumption being that this is a setpiece encounter that the group is intended to experience. But in the OSR, the question is broader: “Will you interact with the troll at all, and if so, how?” The choice is yours to make, its consequences yours to face; the referee will not force you to go after the treasure, but neither will they fudge dice for you or make any guarantee as to whether the encounter is “balanced” with respect to the party’s abilities. The situation generates a dynamic that stimulates the players’ agency by demanding a choice—a choice whose consequences will likely change the course of the entire adventure!

The risk and the reward are both essential to this dynamic. If the players can count on the encounter being balanced as a fair fight, or if they can count on the GM to fudge the dice to protect them, going after the treasure becomes an obvious, “Correct” choice of exactly the sort that Gearing complains about at the beginning of his essay. Similarly, if the players have nothing to gain in provoking the troll, they have no reason to risk lives and resources by doing so. What is not essential to the dynamic is whether or not there is XP attached to the treasure. It doesn’t matter whether the players are “after the gold” or “after the XP”; the choice is meaningful if the risk is real and meaningless if it isn’t. The reward doesn’t even have to be gold—it could be a magic weapon, a stone tablet of ancient knowledge, or a kidnapped princess. As long there is a risk and a reward, the fundamental nature of the question remains the same: Is the reward worth the risk?

Reward vs. Incentive

The deeper implication here is that any form of reward can be used as an incentive. To make sure your players experience that super-cool setpiece battle with a troll that you’ve planned out, just find out whatever it is they want and put that behind the troll! This means that the distinction Gearing draws between mechanical advancement (“Number Go Up”) and diegetic advancement (“ranks and titles … equipment … notable treasures (and techniques)”) is a red herring; if you’re the sort of GM who’s inclined to lead players on with the one, you can do so just as easily with the other. There is a hint of arithmophobia in Gearing’s framing: yes, XP is a number, yes, it goes up, and yes, players may even be excited to see it go up, but how is that excitement any different from the excitement of finding a magic sword or learning a new spell? Even video games know how to manipulate players with non-numerical rewards; there is an entire, highly predatory genre based around exactly that.

At this point, my understanding of what Gearing is on about really starts to break down. He clearly recognizes some distinction between incentives and ordinary rewards, but I’m not sure what the criteria for that distinction are. The best guess I have is that he dislikes XP because, to him, it’s a dissociated mechanic. When he contrasts actions made in pursuit of XP with choices made “with” or “as” one’s character, he seems to be asserting that, because the reward represented by XP has no direct counterpart within the fiction, the player’s choice to pursue the XP can’t be mapped onto a choice by the character. So it could be argued that if the player is after the XP, roleplaying has given way to metagaming, and the character is after nothing. If this is indeed what Gearing has in mind, then perhaps he defines incentive as rewarding the player, and believes that the only legitimate reward is one that goes to the character.

I actually think that’s a pretty good test: if the reward is presented for something the player does—doing funny voices, bringing snacks to the game, or whatever—that’s definitely incentive. But I don’t see much point in splitting hairs about the PCs’ in-character motives for the daring deeds they do; it’s still the character stealing the troll’s treasure. The real motives to interrogate are your own: Why did you choose to put the reward—be it gold, magic weapon, stone tablet, or princess—behind the troll in the first place? If it’s because you want the party to interact with the troll, that’s incentive—you’re rewarding your players for going along with your plan. Otherwise, it’s just a part of the scenario—an opportunity for your players to exercise agency by deciding whether the reward is worth the risk.

Closing Thoughts

Gearing and I seem to have fundamentally different understandings of what player agency even is. I look at it as something fairly abstract—the player’s ability to act upon the fiction through the medium of their character, making decisions that are allowed to carry real consequence. To Gearing, it seems to be something more: the ability to set one’s own goals, to choose for oneself what the game is even going to be about, all from within character.

I’d love to compare and contrast his design choices in Wolves Upon the Coast with my own actual-play experiences, specifically with respect to topics of agency and reward, but this post is long enough as is, and I’ve been working on it for over a week now. It still feels a bit incomplete, but if I don’t publish it now, I’m not sure I ever will. I might do a follow-up to it; let me now in the comments if there’s an angle I’ve missed.