This might work for an MMO but I don't play them so I'm thinking single or multiplayer RPG: How badly do you think it would hurt your suspension of disbelief to be able to see the internals of the game's world state? Take, for example, Fallout 1: In that game the Master is said to be slowly gaining control over the game world. You can't really see it happening at large, but if you don't take certain actions by a certain time the Master's influence will reach your home, the Vault where you started from and which you're tasked to save. This can be seen as a timer (which actually Fallout gives you in the form of the Water Chip quest), but I think this might be done better. What if, for each faction in an RPG, you can see some estimate or approximation of their resources? What if, for each location in a game, you can get resource related information, such as how much food they have left or how dangerous nearby threats have grown? And what if each of these values is tied to some shift in world state which affects you (such as raising enemy strength or lowering prices)? In a more sandbox oriented context this state info might allow you to seek your own goals. If you talk to a stablemaster and he tells you that feed for the animals is low, you could do something about it. If a planet estimates that terrorists are moving ever closer to destroying the capital, you could choose to seek them out. Obviously for the sake of immersion how this information is divulged would greatly affect your opinion of the game. If you go to a screen and look at stat bars that might not be so good (depending). But if it's colored with flavor text, if the information quality can vary (or be tainted by perspective) and if the information is accessible enough so that you can figure out how to affect the game's world state, wouldn't this allow the player more freedom? Thoughts?

If used as a gameplay mechanic properly, this could actually be a selling point of a game. Lets take the planetary terrorist example. There would be a pretty big margin of error to start, and you would only receive limited information: number of attacks, number of sightings (increases with some sort of panic system), and maybe a few others. Quests can be undertaken (infiltrate cells, boost communications in the area, etc.) which would increase the accuracy of existing stats, and add new ones, like rate of growth, a pie chart of recruitment ages, number of cells, and even some knowledge of the higher ups, unlocking assassination and kidnapping quests. Most quests you did would at least marginally boost intelligence in that area. It would be kind of cool, particularly in a splinter-cell-esque game if you could see more and more info accumulate about an organization over time.

Quote:

Original post by doomhascome If used as a gameplay mechanic properly, this could actually be a selling point of a game. Lets take the planetary terrorist example. There would be a pretty big margin of error to start, and you would only receive limited information: number of attacks, number of sightings (increases with some sort of panic system), and maybe a few others. Quests can be undertaken (infiltrate cells, boost communications in the area, etc.) which would increase the accuracy of existing stats, and add new ones, like rate of growth, a pie chart of recruitment ages, number of cells, and even some knowledge of the higher ups, unlocking assassination and kidnapping quests. Most quests you did would at least marginally boost intelligence in that area. It would be kind of cool, particularly in a splinter-cell-esque game if you could see more and more info accumulate about an organization over time.

I like the idea of more information. What's important is that it unlock more in-depth gameplay. Take your starting items of sightings and attack frequency. These could translate in the player's mind to chances for encountering them in certain areas. But to allow the player to be more proactive maybe this information allows them to estimate the ranges for bases of operations.

This information would need to ultimately allow the player to not only estimate locations, identify operatives and block their resources but eventually anticipate and stop the terrorists' next move.

For anyone still reading, some additional thoughts:

I've long been interested in an AI model where entities have needs and points in the game world attract them to satisfy those needs. A merchant starship may have need of fuel, for example, while a space station may broadcast that it offers fuel. A pirate may have need for spoils, on the other hand, and the merchant may broadcast that it "offers" such spoils. As the pirate drew near, the trader would increasingly have a protection need and defense bases would broadcast that they have protection.

I'm trying to tie this idea more closely with more dramatic events, such as the imminent terrorist attack, though-- and I'm having a bit of trouble figuring out what approach would work best:

On the one hand I could try to simulate the low level steps that lead to the event. Let's say it works like this: A stolen ship enters a star system, passes through security sweeps, docks with a space station and explodes.

At each phase I'd need to come up with attraction points and needs and whatever other tests that create this. But this could result in the mission being aborted before the player ever becomes involved. The terrorists could bungle a checkpoint, for instance, or encounter other elements which throw their plans out of whack (what if they get attacked by randomly spawned pirates in route, for instance?)

The more traditional approach is far less dynamic: Spawn the enemies at a point and tell the player that they're on the board, but make the enemy immune to anything that could harm them until the player gets to them.

The first approach has a wider variety of events that can surprise the player but I'm not sure how enjoyable it would be to ramp up for a mission that solves itself.

On the other hand, the traditional approach suffers from "player is center of the world and must save everyone" syndrome.

Which feeling do you want to create for your players; that they're the great hero who has to save the universe, or they're part of a complex environment that churns out around them? Either is fine, but it depends on the experience you wish to create.

My rough feeling is the sandbox approach will make it easier for the player to make more meaningful decisions in the universe, i.e. if there's a particular base the player wants to protect from terrorists, they could make sure it's stocked to the brim with sensors and guard patrols. It might be harder to integrate "player-exclusive" quests in the universe, however Sid Meier's Pirates! managed to have a mix of both. It tended to work in that game as it was more based on the movie version of pirates than a realistic sim, so you were willing to suspend a bit of disbelief as to why no-one was attacking a unique ship.

Quote:

Original post by Wavinator In a more sandbox oriented context this state info might allow you to seek your own goals. If you talk to a stablemaster and he tells you that feed for the animals is low, you could do something about it. If a planet estimates that terrorists are moving ever closer to destroying the capital, you could choose to seek them out.

I like the idea of letting the player choose their own way to "save the world", or do whatever else there is for them to do. The trick, of course is to have a simulation rich enough to let the world-destruction occur "by itself", and letting the player have the information required to make a plan.

I don't think you need to think of it as exposing "internal state". If you want to know how many resources a given faction has, or anything else they might know, have your character ask them. The information is just a part of the world, like it is in our world. You just have to come up with a believable gameplay method of exposing it.

Quote:

Original post by theOcelot You just have to come up with a believable gameplay method of exposing it.

My thoughts exactly

"I've long been interested in an AI model where entities have needs and points in the game world attract them to satisfy those needs"


You can use the model of the 'random encounter' scenarios used by old RPG systems where you have templates for the different interactions which in turn have a set of precanned possible outcomes which precipitate other encounters scenarios.

The outcomes have logic which make them more/less likely (and enable disable options that dont fit). The scanarios themselve would have a similar filtering logic to match against situational circumstances.

The interactions have results which themelves change the situation (and spread them across a map) and precipitate further reactions.


Pirate looking for victim hunts in area.

Nearby merchant ship on regular route gets spotted

They might see the pirate a ways off and determine what they are and call for assistance

Patrol ship might come and rescue/attack pirates
or
Pirates go unhindered and rob and leave a wreck that could be salvaged

Pirate gets caught at checkpoint nead world they fence their stolen goods at
and try to run check point.

Outside the immeduiate area the macro factions (galactic patrol, merchant guild,underworld later hear about the incident(s) and ajust their operations
to compensate.

Its not a tree of scenario stages but a flatter pattern matching system to
create scenario steps and decide consequences and further a local plot

The player may encounter this little continuum at any point and could change the results which precipitates further interactions. The player can decide what they want to do (effectively creating their own quest because when they get involed then there are repercussions for the actions they take).


These interesting incidents would have to occur more frequently than you would expect (so that the player has a high chance of blundering in on it -- a boring well run universe doesnt amke a fun/interesting game). The simulations can be generalized with minimal processing when the player isnt near, but would be ready to be 'realized' into game objects on the map as soon as the player can start seeing them (anywhere in the middle of a multi part action/reaction sequence).


Mini scenarios happen frequently (ship runs out of fuel/has engine trouble...) but might have infrequent interesting twists (false mayday by pirates or ship turns out to be smuggler with easy picking to a opportunistic player) or it might blow up and damage the player turning them into the ones needing a mayday

etc.. etc.. etc...


Interactions in the simulation have consequences which are accumulative which eventually precipitate bigger situational changes (macro plot) which need built in control in the logic to eventually try to recenter balances of factions
-- so that a snowballing effect (escalating feedback loop) doesnt throw the entire game simualtion to an extreme which destroys the game theme.



Different objects in the simulation may each have somewhat different 'preferences' and 'tendencies' whch color their behavior in the decisions for their reactions to the different scenarios. Memberships in higher order groupings can likewise have policies/standing orders that rule their members behaviors (it gets interesting when there are conflicting goals of different multiplegroups the objects might belong too.)

Quote:

Original post by Trapper Zoid Which feeling do you want to create for your players; that they're the great hero who has to save the universe, or they're part of a complex environment that churns out around them? Either is fine, but it depends on the experience you wish to create.

I want some of both, more of the latter (part of a living world) but with a few traits that make you a little special. I'd go all for the latter but I think there'd be too much meaninglessness in the experience.

Quote:

My rough feeling is the sandbox approach will make it easier for the player to make more meaningful decisions in the universe, i.e. if there's a particular base the player wants to protect from terrorists, they could make sure it's stocked to the brim with sensors and guard patrols. It might be harder to integrate "player-exclusive" quests in the universe, however Sid Meier's Pirates! managed to have a mix of both. It tended to work in that game as it was more based on the movie version of pirates than a realistic sim, so you were willing to suspend a bit of disbelief as to why no-one was attacking a unique ship.

Have you played Depths of Peril? It's more fast paced than I like but one trait that's really cool is that you're not the only hero. I'm thinking about making it so that there are several NPC heros as well as "Level 0" folk roaming the land. So if a player-exclusive (more hero-exclusive, really) situation needs to exist than I can tie this idea in. The ship that needs to succeed until you find it then will be controlled by an NPC hero who has powers like you, so it explains why he's surviving where he shouldn't.

Quote:

Original post by theOcelot I like the idea of letting the player choose their own way to "save the world", or do whatever else there is for them to do. The trick, of course is to have a simulation rich enough to let the world-destruction occur "by itself", and letting the player have the information required to make a plan.

Right. Rich simulation means lots of states for objects (beyond "alive" or "dead"), interesting behaviors for those states, and gameplay for learning about them and affecting them.

Take adding morale checks, for instance. Can the terrorists get cold feet if they see that the station they want to attack is heavily guarded? What's more satisfying, that they run away such that you never see them; they're caught because they proceed anyway (possibly looking stupid); or that they attack some softer target; or that they bide their time and wait?

Just like enemies fleeing in combat or hiding so that you have to run them down, there are things that are logical for the AI to do which aren't exactly fun for the player.

Quote:

If you want to know how many resources a given faction has, or anything else they might know, have your character ask them. The information is just a part of the world, like it is in our world. You just have to come up with a believable gameplay method of exposing it.

True. I get the feeling that I'm going to need a combination quest-log and dossier for the important NPCs and factions. It would have to chronicle the info you've gotten from talking to people, but I'd also need some idea of how old or true information is. Something like "Last Seen" is a great example, as it could have a date and frequency attached, which could let players figure out for themselves how reliable something was.