So! I’m thinking about Reinforcement Systems, and what they can actually do in order to effectively (and not ham-handedly) direct player interactions. The arguments against the term “Reward System” which snort at the proposition that players will only do what they’re rewarded to do are, while not concept-killers, at least puissant enough for me to reshape the term as Reinforcement rather than Reward. The fact that these systems utilize disincentives as often as incentives also suggests to me that the term and concept needs an upgrade. Hence Reinforcement, which is not only a little less denigrating but is also a little more accurate.

The way I see it, reinforcement systems mostly work in four ways. The first two manipulate the Fiction to reinforce desired behavior, and the second two see their effects in the social space of the players.

Authority and Credibility

On the Fiction-and-System interface side of the equation, we have the first and straightforward method, and the second and convoluted method. The convoluted method is actually more common, and is what we usually think of first when we talk about Reward Systems.

The first method is to privilege the actions of players who play towards desired content — Exalted stunting is a good example of this, awarding extra dice to over-the-top narration. Simply put, if you’re cool and flashy, you are more likely to succeed. In raw terminology, these reinforcement systems assign desired narration additional credibility.

The second more convoluted method is to reward players for desired behavior by giving them experience points (or the local equivalent) with which they can buy or upgrade character stats. These stats, in turn, are used as authority tools which end up giving the players’ future statements more credibility. GURPS hands out character points for completing objectives and “good roleplay”; Dogs in the Vineyard assigns Fallout and its cousin Experience for engaging in the conflict resolution system. These reinforcement systems give the player authority tools to reward desired action.

Sort of a “2b” reinforcement system is refreshing stat pools for desired behavior, such as Changeling’s willpower, which replenishes when you act according to your nature (or sleep, a caveat which kills this particular rule). Superficially different, such a system nevertheless bestows authority tools just the same.

Full Light, Full Steam has two reinforcement systems along these lines. Thematic Batteries give players’ narration more credibility when they play to their characters; Spoils give players authority tools when they play to others’ characters.

Social Reinforcement

On the System-and-Goal side of the equation, there are two reinforcement systems that I see. However, both of these are potentially problematic for the game designer. The above two are relatively simple to implement, while the following two are difficult if not impossible.

First, of course, is the elusive “being cool” reward system, gaining the social esteem of your fellow players, getting them to say “that’s cool” or nodding and grinning. This is perhaps the most powerful reinforcement system of all. This system is composed entirely of social rules and cues, but is unconnected to any game text. However, as a designer, the only thing I have control over is the text, so I’m not sure this system is of much use to me. I can’t design for what some group of players I’ve never met think is cool. The best I can do is design for what I think is cool, and what people close to me think is cool. (This goes for fictional content, such as superheroes-and-guns as well as situations, like Vincent’s predilection for escalation, both of which you can design for.)

Another alternative is the sense of self-satisfaction in displaying skill and/or creativity in the roleplaying medium. Some players aren’t really playing for the other players, they’re playing for themselves. This typified vast tracts of my MUSHing roleplay, where I would see if I could “pull off” a certain thing (whether it be a tactical objective, a quality story, or a vibrant characterization) in a social setting where the majority of the other players were unimportant to me. However, this system suffers even more from the problem above, in that the standards of success are even more idiosyncratic and inaccessible to me as a designer. The best I can so here is to design for players defining what’s cool for themselves and working towards it, which Thematic Batteries do to some extent.

Designing for Social Reinforcement

It occurs to me that both the “being cool” and self-satisfaction reinforcement systems are, rather fundamentally, extratextual systems over which designers have nothing but the most feeble control. One can extend the argument to say that any playgroup can scrap the authority and credibility reinforcement systems, but whereas I can dictate how those two work (and those dictates can be ignored) I can’t functionally dictate group dynamics or individual preferences. I can only give guidelines (as the first chunk of FLFS’s First Session demonstrates) that can hope to influence those systems. It’s an interesting thought to consider that, as a game designer, I’m not the only one bringing reinforcement systems to the table.

Or perhaps I’m thinking about this under the wrong terms. Perhaps, just as I do not dictate the actual content of a given die roll, I don’t need to dictate the actual content of what’s cool and what’s individually satisfying. What I as a game designer do is provide a system that facilitates player interactions. Is there a great functional difference between “Decide which Thematic Battery to discharge and by how much, then promote your ranks that much for the roll” and “Pick three things which are important for your conception of the character in the context of the game”? The first is the credibility reinforcement system; the second is choosing Thematic Batteries, which leads to self-satisfaction. In neither case do I dictate what the Thematic Battery is, just provide the procedure by which they are created and used. I must admit, I like this option better, but I’m biased as a game designer.

Reinforcement Systems At Odds

Let’s go back to that “2b” example where we refresh stat pools for desired behavior. If one considers Health to be a similar pool-like authority tool, we can easily cast losing in combat to be a behavior that is discouraged, not as an unfortunate part of a story but as undesired roleplay. This makes me strongly suspect that a Health or Hit Points stat is incompatible with pool-refresh reinforcement systems if the players will ever want or need to portray a character losing or damaging themselves. That is, if you usually refresh pools by doing what you’re supposed to do, draining those pools will naturally be seen as doing what you’re not supposed to do.

Similarily, if you usually gain authority tools by performing desired behavior, losing those authority tools will feel like the results of undesired behavior. To some extent, this is a complex way of saying that character death is a reduction in player power when player power is tied to character well-being. In a game where player power is the usual reward for desired behavior, however, character death becomes seen as a categorical failure.

However, two reinforcement systems working at cross-purposes is profoundly common when we take the social reinforcement systems into account. Player death is an easy example — in the height of the climax of an adventure, I sacrifice my character to save yours. Smiles and tears all around the table. Awesome moment. However, I lose all of my authority tools that are associated with that character. If I’ve spent a lot of time performing desired behaviors in order to gain those authority tools, it’s a strange scenario when behavior desired (or appreciated) by the social situation deprives me of those tools.

Certainly such a scenario can be dodged. Dogs in the Vineyard liquidates those authority tools into a currency that is then turned into a new player character’s authority tools (with a little bonus for your trouble). When this is compounded by the fact that character death only happens as the result of player choices, we can frame this as a reinforcement system of its own. Sacrificing authority tools allows the player to get a new and slightly larger set than before.

However, reinforcement systems in conflict may not be as much of an issue as I’m making them out to be. There may be something to be said for deliberately setting two reinforcement systems against each other in order to provoke players into choosing between one reinforcement system and the other. The player would be choosing, not only on the basis of reward, but also on the basis of which system’s desired behavior to exercise. This might lead to some powerful statements by players, but it might also complicate each situation well beyond the designer’s goals and lead to behaviors destructive to the end goal of the game. If it would be in-character for my character to do A which garners me X resource that I don’t need, and it’s very out of character for her to do B which garners me the much-needed Y resource, this strains the distinction between character and player, resulting in (a) an unhappy player, (b) a flat pawn-stance character, of (c) both. I suspect that this is simply dangerous territory — not forbidden, automatically-dysfunctional territory, but easily-dysfunctional be-damn-careful territory.

But How About More?

I’ve outlined four methods for reinforcement systems, but I’m sure there are others out there. What other methods are there to reinforce desired behavior outside of credibility, authority, group esteem, and self-satisfaction? Some of the shorter one-shot games have scoring mechanisms to determine who “won” rather than disburse XP; do these constitute a fifth means of reinforcement?

We just finished our first playtest of Full Light, Full Steam, and nothing blew up! Well, nothing as long as you don’t count the things that the PCs ran into.

In any case, an AP report of the playtest can be found at the kallistipress.com forum.

Wheeee!