Operant Conditioning & RPG System

[Blake Hutchins]

Zak,

You're absolutely right.

Let me add a point of clarification as to why I made my comment.  My experience is that you get better results in game design (and other areas, such as child-rearing) with positive reinforcement.  Diablo II is a simple model that encourages simple behavior.  However, it's bloody addictive, as evinced by the titanic sales figures.  I suggest a starting point for design focus on determining what behavior you're targeting for reinforcement (buttons) and what reward systems you want to offer (treats).  Then I'd move to fill in any gaps with the negative consequences.

I echo the comments about this being an exciting thread.  Good stuff here.

Best,

Blake

[Bankuei]

On a purely gamist viewpoint, its also interesting to note some of the different reward systems and their psychology.

1) D&D's Level System
 Gaining Experience by itself does you no good, but when you get enough, you gain a significant goodie in going up a level.  While you know how many points you need to get there, you have no idea what you'll get this session(random intermittent reinforcement).  You are also gambling that the choices you made in your previous levels as far as choosing skills, feats, spells, will allow you to survive to the next one(negative reinforcement, losing all the work you've put into this character).  As a gamist mechanic, you start thinking about how you'd  like your character to develop over time(positive reinforcement, potential advancement, remember Pavlov's bell?

2) WhiteWolf's xp system
 Experience is immediately useful to you as you gain it.  You may spend it whereever you like(strategizing), but it provides diminishing returns as you raise any stat higher and higher.  Experience tends to come more regular and predictable, and benefits accured are gradual, in many cases, its hard to see how the benefits affect you("yaay.  an extra die. whee.").  On the other hand, being given the freedom to specialize gives the players a lot more strategizing to do, but also tends to make the concept for characters more blurry as they advance(levels provide distinct decision points, buying attributes an open choice at any time).  Experience is not tied to any specific action, and therefore rewards do not enforce any particular action in the game.

3) Chaosium's skill system
 As you use abilities you get the chance to increase them.  Obviously the increases are random, but based on the act of using that skill.  Therefore the reward is reinforces using your skill.  On the other hand, the reward is only a percent or two, so again the growth is gradual, and perhaps not very satisfying.  Since it only matters that  you use the skill once, this system encourages you to use a variety of your skills rather than focus on a few.

It is also interesting to note that all three reward systems "bank" any goodies in the character rather than the player.

I also think that Scattershot's (I think it's in Scattershot) idea of having a mechanical reward that follows the player rather than a particular PC is really brilliant.

I also did this in Persona/Forgotten Fist, primarily because the negative reinforcement of losing everything with character death is avoided.  It's interesting to note that as long as everything invested is reliant on the character, players are encouraged to be overprotective of their characters and also reluctant to change characters.  

In this way, death becomes the biggest negative reinforcement, regardless of its likelyhood in the actual game.

Chris

[Le Joueur]

It is also interesting to note that all three reward systems "bank" any goodies in the character rather than the player.

I also think that Scattershot's (I think it's in Scattershot) idea of having a mechanical reward that follows the player rather than a particular PC is really brilliant.

I also did this in Persona/Forgotten Fist, primarily because the negative reinforcement of losing everything with character death is avoided.  It's interesting to note that as long as everything invested is reliant on the character, players are encouraged to be overprotective of their characters and also reluctant to change characters.  

In this way, death becomes the biggest negative reinforcement, regardless of its likelyhood in the actual game.

I'm sorry to say that TQuid must be thinking of some other game.  That's not how I originally conceived of Scattershot.  (But that won't keep me from stealing such a good idea.  Thanks!)

Fang Langford

[Zak Arntson]

I talked with my wife about this, making sure I understood the theory behind Operant Conditioning. Most importantly, she talked about the merits of each type.

Positive Reinforcement - (Adding a consequence to support a behavior) - Example: Giving a tasty treat for pressing a button. This is shown to be the most effective, short- and long-term. Intermittent (read: sometimes you are rewarded for the behavior, sometimes not) Positive Reinforcement is the most effective. In either case, Interemittent or regular Positive Reinforcement has the strongest effect.

Negative Reinforcement - (Removing a consequence to support a behavior) - Example: Introduce an eir-piercing squeal that only stops if you press a button. This works in the short-term, but has little long-term benefits. This and Positive Punishment are on par with their benefits.

Positive Punishment - (Adding a consequence to stpo a behavior) - Example: You press a button and you get a mild shock. The effects are similar to Negative Reinforcement.

Negative Punishment - (Removing a consequence to stop a behavior) - Example: You get a bowl of treats; if you press the button, you lose a treat. This is the least effective, and is not very effective, even in the short term.

How to tie the effectiveness into game design? Your Premise will demand certain behaviors. For example, I'm working on a game, Divine Right, where the players are exorcists (kind of the opposite of Sorcerer :). The Premise: "What personal sacrifices would you take help others?" Tying Exorcism into the Premise: "You have the rare ability to help others at great personal sacrifice, how far will you go to use this ability?"

Going with this Premise, I outline behaviors I want to increase:
* Sacrifice Self
* Perform Exorcisms

When want I increase these behaviors, it is Reinforcement. When want I decrease, it's Punishment.

Increase self-sacrifice: Positive - More power to perform the Exorcism. Negative - Loss of safety & well-being.

Since I know that Positive Reinforcement (giving Exorcism power) is more effective than Negative Reinforcement (removing safety), can I assume that even if you gain an equal amount of power as a loss of safety, players will still prefer to gain power? This is what I'm hoping.

Perform Exorcisms: Positive - Emotional release of tension & Player satisfaction. Negative - Expenditure of power.

Positive Reinforcement (emotional release) will far outweight the Negative Reinforcement (expenditure of power). That's a fine assumption, especially when the Player invests an entire playing session gaining power to perform an Exorcism.

Now what other behaviors does the above imply?
Preserving Self has the consequence of denying power. So Preserving Self has Positive Reinforcement (no loss of safety) and Negative Reinforcment (no gain of power). As long as I design with the benefits of Gaining Power over Self Preservation, my Premise should be kept.

Same goes with Not Performing Exorcism. You would earn all this power and not use it? That would be silly, but the design should weigh against this behavior. So to remove the Positive Reinforcement (keeping power), a Player's power diminishes over time. It's now a resource that must be spent or lost.

In a nutshell: Combining Premise with Operant Conditioning, I've located the areas my design should tackle. Namely, Self-Sacrifice comes at the cost of well-being, but gains the Player temporary power, which can be used for Exorcisms within a set amount of time.

After this initial setup comes the Unknown Factor part of design, but I've already talked about it in this thread: http://indie-rpgs.com/forum/viewtopic.php?t=1211">D/F/K and the Unknown Factor.

So everyone, does this make good sense in game design? I'd hate to think that I'm going in a lousy direction. I feel like I'm on the cusp of some grand Zak's Game Design Theory (as in, it works for Zak. Don't know if it works for anyone else).

[Ron Edwards]

Hey Zak,

It sounds like you're looking for some ....

(wait for it, oh God I'm so funny)



...



... POSITIVE REINFORCEMENT?

Boohahahaha!

Well, here it is. Operant conditioning is controversial stuff, as I'm sure your wife has told you. All manner of protests and debate have arisen, regarding the "self" (a dubious quantity) and cognition (a much more interesting notion) and even, the evolutionary biologist said shyly, underlying principles of why we learn at all.

All that aside, however, as long as we're talking about plain old reward systems (not themselves especially controversial), I think you're on a fine design track. This thread has demonstrated any number of diverse responses, all enthusiastic, so quit lookin' for Pavlov to feed your motivational center and get to it.

Best,
Ron

[Zak Arntson]

All that aside, however, as long as we're talking about plain old reward systems (not themselves especially controversial), I think you're on a fine design track. This thread has demonstrated any number of diverse responses, all enthusiastic, so quit lookin' for Pavlov to feed your motivational center and get to it.

Hee ... in hindsight, I was looking for encouragement with that last bit more than anything. Good catch! As if all this great discussion wasn't encouragement enough! (kudos to everyone) And yes, I am talking about reward systems and not any kind of "brainwashing of the Players" activity. I want to nail down the merits of rpg reward/punishment systems.  (you know, in Sorcerer, there's the Positive Reinforcement of power coupled with the Negative Punishment of losing humanity. Since the P.R. outweights the N.P., psychologically, it works!)

The design should be up front, and with every Player knowing the rules. In my opinion, that's a fact of the player contract. (unless the players explicitly agree to the "players don't know all the rules" rule :)

Oh, and painful pun, there. I guess if you write theses on humping mouses, you've gotta have a good sense of humor!