Does anyone have any experience on dynamic story-telling as a mechanic? Things like Fable's story progression based on player behavior or the director used in the latest left for dead title? (check out this article if you haven't heard of it, the link will anchor to the appropriate spot) What might be some things to consider (human behavior, plot branching, ect.) and how would you break-down & communicate such a mechanics from a game design stand-point?

It's a bit old now (2005) and I've moved somewhat away from the topic, but there's the thread I made soon after joining the forums: Automated Storytelling and Interactive Plot in Games. There's a list of resources in the first post.

Thanks, I'll look right into it!

I see Story in games in two different ways:

- Telling
- Creating

In Story telling, the designer of the game has a specific narrative (or set of narratives) that they want the player to experience. This is like a movie or a "Chose-your-own-adventure" book.

the second way to have story in games is through creating. Under this concept, the player is the one who creates the narrative and the designer just provides them with the tools (the game) with which to do it.

Often both techniques are use in games, just that the technique of creation is limited to individual scenes of the game and the scenes are strung together in a "Telling" type narrative. So these are not mutually exclusive as such.

However, what I think you are trying to do here (and correct me if I am wrong), is to give the player more control over the creation of their own narrative in the game. If you approach this goal using the techniques of "Telling", then you will fail. What you need to do is use "Creating" techniques.

In Creating Narrative Techniques you need to present the player with situations and let them choose how to handle that situation, then (and this is the important step) you have to let the choices the player makes influence what the future situations they end up in.

The best example of a "Creating Narrative" game that I have played is "Sword of the Samurai" (it is an old 1989 DOS game: http://en.wikipedia.org/wiki/Sword_of_the_Samurai_(computer_game) ).

In this game, you get presented with various scenarios, and how you handle those scenarios and the choices you make in them are remembered by the game and used again as seeds for future scenarios.

As an example: In one case I was playing SotS and I insulted one of my rival Samurai (who was the favoured to become the next hatamoto of the clan) and defeated him in combat killing him. However, his son took over the household and held a grudge against me, but now that the family was dishonoured, he sent assassins to kill me (which I then subsequently defeated).

Later his lands were under siege by rebelling peasants (due to the fact that no longer seen as a powerful ruler after what I had done to his family :D ) and I was there to help him. I use my larger army to quell the rebellion and saved him from being overwhelmed by them. I then worked to mend the fences between us and performed all the necessary ceremonies. I did this so well that he eventually (after a long period of time) become one of my supporters for becoming the next hatamoto of the clan.

In this there were essentially 4 situations:
1) Insulting the rival Samurai and killing him in combat
2) The son sending assassins
3) Going to the aid of the son at his time of need
4) Making peace with the son

And it makes an interesting story. It might make a good movie or book. But, each time you play the game you will get a different story.

Maybe the son succeeds in killing you and your son takes over, or the insulted Samurai's son might never make peace with you. What scenarios you are presented with is essentially random, but the content of those scenarios are based off what occurs in past scenarios.

I created my own narrative as I played the game and chose what to do in each scenario. What is important is that in each scenario it is difficult to loose the game (only if you die without an heir will you loose), but that there is many different ways you can complete the scenario.

Even if you fail a scenario (or at least fail the primary goal) you can still use that failure as a seed for future scenarios. In SotS you can take on missions (say to defeat an enemy army) and you can loose the battle. You will be dishonoured and loose respect of your peers and hatamoto. You could choose not to take the mission and still loose honour (but less than if you tried and failed) or you might defeat the enemy army and gain honour from doing so (and there are a few other ways it you can finish it as well).

Then depending on the results of this scenario you get presented with another scenario based on this (may be the other Samurai's decide you are unfit to be a leader and start insulting you trying to draw you into a fatal duel, or maybe the try to assassinate you, etc).

Even if subsequent scenarios occur at random, if the details of these scenarios are based off the history then you will get the players constructing their own narratives of what occurs.

It can be done very simply by using "templates". If a random event occurs that says that an attempted assassination of the player character is to occur, then you can just look at the past history and use someone that the player has made a choice in another scenario that hurt them as the originator of the assassination. How that NPC changes their opinion of the player is then subject to the outcomes of the current scenario.

This kind of player created narrative give massive replayability because each time the player plays the game they get a new story. It means that you create a world that the player experiences in a far more personal way than through "Telling" narratives.

Many designers crate narratives with multiple paths through the game and even multiple endings and then state that the player can choose their own story. But this is not really giving them a real choice over the narrative, it means that the designer has created several narratives and the player can choose which of these narratives to follow.

This is no different then letting someone chose which book to read, or giving someone the choice of a movie to watch (and that these options only slightly differ in their content).

Interestingly in a game like SotS there is actually only two endings that can occur: You die without heir or you become Shogun of all Japan. This is the same number of endings in most arcade games, either you finish the game or you run out of lives.

A "Telling Narrative" game with multiple paths through it and multiple endings is a lot of work for the designers and writers, but in a "Creating Narrative" game it is actually less work for the designers and writers to create these stories because the player is the one that creates them (there is still work in that you have to design the various scenarios and how they link together but it is far less work that constructing a multipath/ending 4 or 5 hour long story).

Thanks Edtharan, nice post. You really hit on some topics (in a summarized way, which is something I needed) that Trapper Zoid and his many sources pointed at. (Also really good stuff by the way) Thanks to both of you for all the great info. If anyone has something to add, I'd love to hear it.

Good post there Edtharan. Sounds like a good way to go about allowing for broader decision making in the game, but the designer would still need to design templates for a huge number of situations and figure out if all of them work with each other. It would still be a complex system to work with, but I think it would still have a better result.

(btw, please don't write "loose" when you mean "lose". Its very distracting to read even a good post when grammar mistakes scream at me> I'm not picky about it but that's a big one and hard to miss.)

I'm reading this, if anyone happened to have questions on any of my older posts. I don't think I really have anything new to add since I haven't recently done any research or work on this topic. I've been working 24/7 producing vector art for my Anthro Breeder game, so I haven't done much of anything writing-related. I continue to favor the Choose Your Own Adventure storytelling approach, with a focus on increasing how fine-grained, reactive, automated, and modular it is.

But here's a list of semi-related threads I was involved in after the one Trapper Zoid linked to:
perma...life? and rejecting quests
A structure for more interesting enemies (more characterization and dialogue for more recurring enemies).
Visual Diagram of a Generic Plot

*looks for but does not find thread about ideomonsters concept* Maybe I never actually posted that???

Quote:

Original post by sunandshadow I continue to favor the Choose Your Own Adventure storytelling approach, with a focus on increasing how fine-grained, reactive, automated, and modular it is.

In some instances this highly structured narrative design is better than a more open one. "Telling" stories is a long tradition and it even has a place in games.

In cases where you have a specific story you want the player to experience (eg: Movie tie in, etc) then this is the best method to use.

however, if you want the player to have a more general experience of a world, then a "Creating" narrative is much better at performing the job.

It is a case of "The right tool for the right job" (to quote Scotty from Star trek :D ).

Quote:

Original post by jmau0438 (btw, please don't write "loose" when you mean "lose". Its very distracting to read even a good post when grammar mistakes scream at me> I'm not picky about it but that's a big one and hard to miss.)

Gah! Doh! *facepalm* etc...

Thanks for that. Spelling checkers don't pick up that kind of spelling mistake.

Quote:

Original post by JasRonq but the designer would still need to design templates for a huge number of situations and figure out if all of them work with each other. It would still be a complex system to work with, but I think it would still have a better result.

Actually you don't need all that many templates. Sword of the Samurai only has around 4 or 5 of them (iirc). What differs is the inputs to the templates.

It is a complex system rather than a complicated system (complex systems have many interactions and complicated systems have many components - according to the mathematical definitions of them). SO yes, there is a high degree of interactions between the components, but it doesn't mean you need a lot of components to make it work.

The way you can do this is to look at a game as see what sort of situations the player can get into. For example, if you want them trying to rescue princesses from dungeons, or what them sneaking into a castle to assassinate someone, then these two scenarios are variations of the same basic template. But if you had these, and also wanted them to lead armies into battle, then the battle scenario might belong to a different template altogether.

In a typical cRPG game, you have wandering around dungeons, conversations, battling monsters, buying/selling gear, and so forth.

Wandering around dungeons and battling are usually the same template (but don't always have to be - just look at the old Gold Box D&D games like Pool of Radiance). Conversations are a different template, and typically handled by a conversation tree.

In these games, these are not usually linked together. Sure, some conversation trees lead to a combat, and sometimes a combat leads to a conversation (with the defeated opponent) but there is usually very little information sent between these template (as these situations are generally pre-scripted to occur). Because the content of one template is not really influenced by the content of the others the interactions between them are not very strong.

If actions you take could influence the content of the next template you enter, then the level of interaction between them can increase.

For example: Lets take the Combat and Conversation templates and see what you could do to increase the interaction between them.

First of all, you need to allow them to switch back and forth as needed and with each switch the game will remember the state of the last one (so you can stop the combat for a chat and you and the enemy will keep your hitpoints and such).

Next we allow actions taken in one to initiate and influence the state of the other. So during the conversation you can have the option to attack the enemy and during the combat you can start a conversation.

SO entering into this, the aim might be to get some information from the enemy (the Grand Vizier), and for the sake of the example this could be the name of an assassin who is trying to kill the King.

The scenario starts off with the players party entering into the Grand Vizier's room. At this point they are in the "Combat/Explore" template with it set for Exploration. They can then either enter the "Combat/Explore" template (yes the templates can change to themselves) but with it set for Combat with the Vizier, they can enter the "Conversation" template and try to convince the Vizier to tell them without fighting, or they can stay in the "Combat/Explore" template as explore and not interact with the Vizier.

Option 1) Fight the Vizier: This will enter combat and the Vizier will attempt to flee (this would be an input form the status of the Vizier entity) and take the state to option 3. If he can't flee then he will fight as best he can (also another setting of the Vizier entity - and pretty much a basic default behaviour of that kind of entity). The Vizier will also try to talk his way out of the situation (what he can offer as bribes - including the information the party wants - are part of the entity's settings) if he can't beat the party (move the template to the Option 2 template). IF he beats the party, then they are typically dead, or might be allowed to be captured (depends on how lethal the game is designed to be) and this could be considered as Option 4.

This means that at first the Vizier will try to avoid combat, but then if he can't he will try to defend himself, but if he can't do that he will try and weasel his way out of it.

Option 2) Conversation: With this the layer will need to use various skills to convince the vizier to tell them what they know. this could include threat, bribes, violence, etc. Threats would include non physical threats like a threat to expose him to the king, or some other political threat. Where as violence might lead to the combat template (option 1). The Vizier will have information that he can use against the party and against their attempts at coercion (from their past actions - so there is interaction from the history).

If the player's attack, then the state changes to option 1. If they can't convince the Vizier to tell them the information they need then the system will revert to Option 3.

If they threaten the Vizier with violence, the effectiveness of this tactic will lie on the past history of the party. If they have show themselves to be powerful combatants (eg: they just beat the Vizier up in option 1), then this tactic will be more effective in convincing him to talk.

Option 3) Wander: The vizier will go about his own tasks, which would include talking to the assassin. However, if the Vizier sees the party, then he will avoid taking to the assassin. This can be handled easily by having a flag with the activity the vizier is attempting to do. These tags would be "Legal", "illegal". If the tag "illegal" (ie: talking to the assassin) is part of the action, then the vizier will avoid doing that if he can see anyone who is not an ally. So the player can use stealth to discover the identity of the assassin.

This actually gives an alternate tactic for the player in option 2, by convincing the Vizier that they are an ally (truthfully or not) and then discovering the identity of the assassin this way.

Option 4) The player is defeated: They might be captured by the Vizier and put into a dungeon. The Vizier might then be sent into a monologue (a variation of Option 2) and the identiy of the assassin discovered this way. If they can then escape form the dungeon they can then inform the king.

Actually, each of these options might involve several templates (bribery might be a type of conversation template, and threats might be another conversation template with different inputs - or a completely different conversation template altogether - it depends on how you decide to handle these things) and switching between them as needed.

All this sounds like it is very complicated, but it is not as complicated as it sounds.

Each entity has a some variables that it can bring to the various templates and which options within those templates are available to it to use (because it will need to make decisions in the encounters too). Also, the player will have variables that allow the game (and the templates) to know what they have experienced in the past (what they know and have done - and games track these already, just see Morrowind and Oblivion as simple examples). Then the old template just has to inform the new template of the purpose of the switch which is just a set of tags that describe the end conditions (a set of new states to enter into and the conditions that need to be met to do so).

Actually, if you think about it a bit, all I am really describing a what is called a "Finite State Machine", but with the addition that the "States" are templates and each state can communicate information between them. IF you ahve ever used (or heard of) a "Switch" or "Case" statement, then you know how to implement a "Finite State Machine" (it is just a Case statement coupled with a loop and in each Case branch it has a way to change the value the case statement is dependent on).

Most menu systems use such Finite State Machines, this is just a bit more complex and complicated than a simple menu system (but it is essentially the same idea).

I actually recognized right near the end there that it is a state machine. I guess when I was originally thinking of it I wasn't generalizing the templates enough and thinking of more specific situations. Having them very general would reduce the number of templates needed but they would need to be very well designed to allow for a wide range of possibilities.

Quote:

Original post by Edtharan Quote: Original post by sunandshadow I continue to favor the Choose Your Own Adventure storytelling approach, with a focus on increasing how fine-grained, reactive, automated, and modular it is. In some instances this highly structured narrative design is better than a more open one. "Telling" stories is a long tradition and it even has a place in games. In cases where you have a specific story you want the player to experience (eg: Movie tie in, etc) then this is the best method to use. however, if you want the player to have a more general experience of a world, then a "Creating" narrative is much better at performing the job. It is a case of "The right tool for the right job" (to quote Scotty from Star trek :D ).

But, to wax philosophical, isn't a general experience of the world made out of lots of specific experiences of the world? Particularly in a science fiction or fantasy world where the general experience of the world is supposed to have a unique flavor and present a consistent picture of what life in that world is like, different from life in the real world.

I think the border between telling and creating is fuzzy anyway. An Eliza program, or something similar like a soupy-bot, which can be used to generate NPC dialogue in a game, has to have both specific (telling) things to say (otherwise it's boring/useless) and general (creating-enabling) patterns of how to respond to input (otherwise it's uselessly inflexible and fails to seem lifelike). Similarly, pure telling is okay for a purely linear game, but as soon as the player can do anything to affect either the overall plot or the specific events/dialogue at any point in the game, there's at least a little creating going on. What about a Colossal Caves style game where you can experimentally try to combine verbs and nouns, and the game has premade satirical responses to some combinations? And for the Samurai example, the concepts of becoming hatamoto, being dishonored/wanting revenge, and sending assassins vs. challenging to combat, are all little units of telling which are necessary to give the player something to do their creating with.