Hmmmm... players dependent on NPCs. There's something REALLY valuable with this.
Part of the problem is that the player
is so independent that he can normally override every concern except those related to combat (or stealth, or puzzles when allowed).
But at the most simple level (still imagining this was overlayed over a traditional RPG for starters), let's say we want to deal with inhibiting the player's ability to attack allies. Rather than handicap them physically (which is a very situationally specific solution), what if we tried to capture how this dynamic would work in a more generally (and thus widely applicable) case?
What happens if we're all at a party and I suddenly go bezerk and attack one of you guys? At the most simple level, someone will likely try to stop me (depending on what I'm armed with, but let's assume all things are equal for the moment).
The degree of force you use against me is going to be dependent on how much damage I've done. If I kill one of you guys, you'll use a different amount of force than if I bloody one of your noses.
So what good is this?
Having the textured / varied levels of responsiveness now ties my hands as a player if I want to respond to a plot event with violence (e.g., I think one of you is a thief). I can't just whip out my cutlass and behead you, because I now risk the wrath of the group. This has been done in games for ages, but rarely at this level of attempted nuance. (For instance, even if I've killed one of you, you may wish to subdue me so I can be brought before an arbiter
who can kill me legally.)
Okay, complicating it a bit:
The degree to which you dislike / like me is going to modify your response if I attack one of you. The degree to which you like / dislike the object of my attack will also modify your opinion.
Your response will then be based on your opinion of my action (do you like / dislike violence?) and an assessment of what you think is a matching response to what I did.
That's going to give some value that will act as a threshold you'll compare to all the things you can do (this is the typical AI weighting stuff).
Now, two more complications: First, your response will be based on the rightness of my action, which can be derived from an identity (e.g., I get an "evildoer identity flag" for attacking an innocent; I get a "righteous defender flag" for attacking the guilty; I, by default, might have a "I'm a soveriegn lord and can do whatever the hell I want to peasants flag" by birth or title, etc.).
Second, your response is going to be determined by our force ratios. I may be weilding a lightsaber and you may only have a rock. So you may want to run and live to fight another day. If we accept more nuance in conflict, you may be able to threaten me, or barter with me. If there is some way of logging a future plan, you may decide to appear to surrender, but plot to stab me in my back or push me off a ledge.
Now, if you have a traditional RPG that's tactically varied (i.e., teammates have distinct roles), you now make interpersonal conflict, an essential pillar of storytelling, a fundamentally integrated part of the game.
To give this the full punch it would need, you could even have results where:
- Two teammates refuse to help each other, heal each other and fight when in proximity
- One teammate can shift like or hatred against another teammate
- A teammate can abandon another teammate or the party anywhere
- A teammate can be coerced, bribed or otherwise persuaded to stay with the party
Final thought: Erase the player. Use a multiplayer network programming paradigm and make no distinction between an AI or human controlled agent. All of this applies to everybody, and everybody can do it. Then, create game design failsafes for certain situations, such as the whole team killing each other (or not, depending on how raw an experience the player is up for... you could have a Treasure of the Sierra Madre experience, where everybody sets out to get the gold... then there's backstabbing left & right).
Obviously, you need much more texture in the interactions between teammates, but I think the idea of
causing the situational dynamics to limit the player's strategic options is the start of figuring out how to make the player and team interdependent enough that the player actually has reason to care.