Return to “Suggestions”

Post

Emergent multi-part missions, quests and epic stories

#1
I would like to discuss some general concepts for missions. This has as far as I could search, not been discussed a lot yet, except specifically for escorts in the Thread on Escorting Ships and ponctually, for example here and there but without discussion on "how" this is realized.

Now I first would like to understand the concepts, before possibly giving some ideas for the system (missions are sort of a hobby for me. I delevoped quite some for the excellent game Escape Velocity and its successors). This is why it is into the suggestion forum...

In the first thread mentioned, Josh said that missions are not spawn based:
JoshParnell wrote:As you discuss possibilities, try to keep in mind that LT missions are not "spawn-based." This is an important detail and strongly influences how these missions will play out. When you take an escort mission, it will genuinely be because the AI feels that it needs protection for flying through a particular sector of space. Furthermore, if you encounter enemies, it will be because they were already in the region, saw/scanned the convoy, and decided to strike. I will not resort to spawning them just because you took a convoy mission.
Now the game will actually spawn depending on the player action - only the region of space visited will exist, for example, so obviously there is an equivalent of "spawning" systems and NPC, planets, ships as the player moves around and "needs" them.
In my understanding, the difference lies in the fact that when spawning, the nature of the object being spawned is defined by the universe and its root, not by the player and his needs. In other words, the game will not spawn a ship needing escort because you accepted the mission. It will spawn ships that may need escorting and propose the mission.

This is very different, but not fully incompatible. For example, I could - when proposing a mission to escort - "search" for a suitable ship needing escort close to the player and make it the subject of the escort. let's keep this in mind for later posts, it will be important.

Returning to the game logic, this means when entering in a system (or when arriving "close" to the border of the already created universe) a new system is generated, with its ports, corporations, ships and NPC, which have as consequence some generic missions:
  • A corporation/planet needs to move things around: delivery missions.
  • A corporation/planet does not like a guy: bounty hunting mission.
  • A ship feels unsafe: escort missions.
  • A corporation/ship needs to map a region: exploration missions.
All of this with some variations (deliver from here to there, or go there to get cargo and bring back here. Cargo can be mail or illegal drugs or plain metal, and so on).

Of course succeeding in a mission may earn you cookie points with the ship/corporation/... and encourage the AI to entrust you with further similar missions perhaps more difficult and with more rewards, but I don't see however how these kind of standard missions can evolve into epic stories...
Now that may not be the aim of this game, and it will already be awesome with only those standard missions on top of trading, pirating, improving your relation with a corporation and exploring.

Still in my humble opinion, having the possibility to have long winded missions composed of many small steps that build up a story is one of the key element to make the universe interesting - because it makes you really "belong" to the universe.

I am actually thinking of a procedural way to generate such epic stories. The major difficulty will probably be in the packaging (i.e. generating the narrative that binds everything together), and there I don't have a ready made solution yet. The procedural system can generate many different stories, but of course one could imaging a plug-in system that tweak it to add depth.

If there is some sort of confirmation that I understood the concepts correctly, I will expand to a short explanation of the procedure soon to open the discussion to more conrete elements.
Image
Post

Re: Missions, quests and epic stories

#2
I can't believe really nobody's interested in complex missions - probably my formulation was too unclear.
I'll give this thread a second chance and a bump against netiquette before it leaves the first page... :oops:

The core question (lost in a wall of text), was:
CSE wrote:Now the game will actually spawn depending on the player action - only the region of space visited will exist, for example, so obviously there is an equivalent of "spawning" systems and NPC, planets, ships as the player moves around and "needs" them.

In my understanding, the difference lies in the fact that when spawning, the nature of the object being spawned is defined by the universe and its root, not by the player and his needs. In other words, the game will not spawn a ship needing escort because you accepted the mission. It will spawn ships that may need escorting and propose the mission.

This is very different, but not fully incompatible. For example, I could - when proposing a mission to escort - "search" for a suitable ship needing escort close to the player and make it the subject of the escort.
To make complex stories - or a series of missions following each other - you need to make sure certain assets are/were available. If you want the player to chase a ship ("bounty-type"), but then to give him a reason (e.g. via dialogues) to ally with that ship and switch allegience (hence building up a story), you need to make sure the ship is there in the first place.
In normal games, you can just spawn the ship as required. In LT, you cannot do that. But you can search for an appropriate already existing ship and offer the mission focusing on that ship. This allow to obtain the same effects, but instead of spawning afterwards, you must check in advance the existence of the assets before offering the mission.

So does it make sense to discuss/propose a procedural way to generate missions based on this assumption?
I am asking because putting too much work / enthusiasm into something based on wrong assumption is not very useful...
Image
Post

Re: Missions, quests and epic stories

#3
The feeling I got about the availability of missions was not that the AI tried to provide you with missions at a steady rate, but rather the world itself resulted in situations that could generate missions for an interested player. For example, you would get an escort mission when an NPC, going about its normal business in no way related to your existance, realizes that the area of space they are about to travel through is more dangerous than they anticipated and they need more muscle to feel safe. So they put out a call for help, which you can choose to respond to; if you don't take them up on it, either another NPC will, or they will proceed/cancel their journey, depending on how safe/not safe they feel.

So, assuming that idea is correct, I'm not sure how well your concept would mesh with the way the world flows?
I am 42.
Post

Re: Missions, quests and epic stories

#4
The only way I know for creating complex missions is to script them. At least in other games I played complex missions were always scripted.

And while Josh is certainly doing a very impressive job of teaching us what a powerful tool procedural generation is, I still doubt that you can create a meaningful, evolving, complex, story-telling, multi-stage mission without scripting (and therefore predesigning) at least crucial parts of it.

And because LT is going to be procedurally generated throughout, without the predesigned elements that seem to be necessary for complex missions to me, it seems a given to me that there won't be complex missions in LT.

(Of course, Josh can prove me wrong.)
Post

Re: Missions, quests and epic stories

#5
CSE wrote:To make complex stories - or a series of missions following each other - you need to make sure certain assets are/were available [....] But you can search for an appropriate already existing ship and offer the mission focusing on that ship. This allow to obtain the same effects, but instead of spawning afterwards, you must check in advance the existence of the assets before offering the mission.
Yep. Also, you could accept that some multi-arc missions will just die if the dependent assets become unavailable. However, can you expect to know just what happened?

An example: you're in a bar at a station, and pick up a contract to take a Mr. Estragon to a different station at another star system, whereupon you will be paid by a third party - a Mr. Godot. You take the passenger to that station, fulfilling your part. However, Mr. Godot doesn't turn up. In the reality of the dynamic game world, his ship was destroyed before it reached the station. Now, do you get told that he's not coming and the mission is terminated? Or do you just wait for Godot until you get bored and move on?

Edit===
Commander McLane wrote:I still doubt that you can create a meaningful, evolving, complex, story-telling, multi-stage mission without scripting (and therefore predesigning) at least crucial parts of it.
You could still have multi-arc missions generated from a pool of available configurable components, though explicit scripting would usually do a better job.
Post

Re: Missions, quests and epic stories

#6
CSE wrote:I can't believe really nobody's interested in complex missions - probably my formulation was too unclear.
Not that yours are unclear, I think a lot of it has to do with how these missions would be formed procedurally.

JabbleWok wrote:
Commander McLane wrote:I still doubt that you can create a meaningful, evolving, complex, story-telling, multi-stage mission without scripting (and therefore predesigning) at least crucial parts of it.
You could still have multi-arc missions generated from a pool of available configurable components, though explicit scripting would usually do a better job.
I really think generating multiple-link or complex missions are possible, and well within the PCG capabilities. It's just that it will take a bit of coding and fleshing out in order for them to come to life and not feel forced.

A good example I can think of;
  • Stage 1 - Scout an area and report back the findings. You find there are pirates patrolling an area. They spot you, you have a tussle, and you get the heck out of Dodge while their friends are still warping in. They don't pursue you.
  • Stage 2 - Escort a convoy. You report back that there are enemies. Your quest giver is surprised that there are more forces than anticipated and decides they need an escort and extend the invitation to you to help out. If you only have one ship, they hire others for the convoy as well.
  • Stage 3 - Repair the convoy. The convoy took a bit too much damage and had to make an emergency stop and now need help in repairs. You then do what you can (via scrap/salvage) to repair the ship while fending off a few pirate scouts.
  • Stage 4 - Precision strike against pirates. Upon arriving safely, the quest giver decides they want to take action. For your service to them, they hire you and send along a few fighters of their own to strike against one of the larger pirate bases to show they mean business and to send a message.
  • Stage 5 - Defense. The pirates retaliate with their own force. Suddenly you're caught up in the middle of a war between two factions.
  • Stage 6 - Behind enemy lines. To turn the tide of the battle, you're requested as an escort to a wing of bombers to a major sneak attack at the enemy base. You take a round-about way of reaching the base and meet up with some more forces. You destroy the base.
  • Stage 7 - Clean up. The forces you assisted return with you and attack the pirate faction from behind in a pincer move.
After that, you get your cumulative reward.

Now, this may seem complicated, but it is completely within the realms of PCG, and also leaves the methods open ended to the player. Of course, at each stage, the player can choose to end it, and whether or not they choose to be lazy (just get the mission done and return) or heroic (going above and beyond the call of duty) the paths can branch.

The procedural layout would be something like this;
  • Stage 1 - A scout mission seems simple enough. I'm guessing these will be available in the game. The pirates were already there (to abide by Josh's PCG mission generation methods.
    • Outcome 1 - The player destroys all hostiles, reports back. Your quest giver may not need an escort now. Quest line over.
    • Outcome 2 - The player destroys all hostiles, reports back. Quest giver has new intelligence that the pirates were part of a bigger faction. Turns out, after you left, more jumped in and wondered where their mates were and now enforce the patrolling in numbers. Quest giver needs escort.
    • Outcome 3 - The player shows up, sees the pirates, and then leaves. Quest giver requests escort.
  • Stage 2 - Simple enough. Escort is escort.
  • Stage 3 - Completely dependent on how well you protected the convoy. May be skipped if you did really well.
  • Stage 4 - Depending on how much the quest giver's faction's rep hates the pirates, this may be possible, maybe not. It could also be a 'retrieve item from pirates' if they don't hate each other too much.
  • Stage 5 - This could happen if the pirate faction is large enough, or the two hate each other enough.
    • Outcome 1 - Quest giver's faction is large enough to hold the attack, but the pirates are strong too. All out war. Leads to next stage.
    • Outcome 2 - The player has a massive force of their own to contribute. If the two sides are even, this still leads to the next stage.
    • Outcome 3 - The player has a massive force of their own to contribute. The player's side is stronger. They destroy the pirates then and there. Alternate stage for all-out attack on pirate base instead of sneak attack from behind.
    • Outcome 4 - The pirate faction is stronger, and destroy the quest giver's faction.
  • Stage 6 - If there is enough of a stalemate in the war, this could be needed. Most of the pirate forces are pre-occupied with the war, so a strike against their stronghold could be possible. Meeting with other forces or not, is dependent on whether or not you have extra forces, or the ally faction does. Whether or not you destroy the base could lead to more branches. What if the pirate faction was larger than this?
  • Stage 7 - Comes back to Stage 5 to finish them off.
Of course, a lot of these take into consideration a bit of leniency as well. For example, the scout mission and subsequent convoy could purposely move through heavy pirate territory. If you're near a large pirate base, this could be easily done.

Second, if scouting missions are in the game, this could almost always be used a segue into more missions.

With minor tweaking, all of this could be viable, and heck, you got yourself a full-on mini campaign, fully generated! This all depends on how many different types of missions there are, and how in-depth the generator would be. Generating this with the different outcomes could be possible.
Image
Early Spring - 1055: Well, I made it to Boatmurdered, and my initial impressions can be set forth in three words: What. The. F*ck.
Post

Re: Missions, quests and epic stories

#7
Exactly, DWMagnus is describing one way to make complex mission strings.

I would even go one step further, and not script it as "closely" as in your example, but have generic combination rules for making the plots.

Example:

Example Mission String #1
  1. player just completed a "transport cargo" mission (one of the standard ones that happens frequently).
    IF player got attacked during mission, 10% chance goto 2, 10% chance goto 3 (and 80% chance do nothing, mission string is finished).
  2. Recipient fraction (F#1) ask you to investigate why attacker's faction (F#2) was unhappy by contacting an informer (check if NPC of attacker's faction F#2 exists and select this guy (NPC#1) as informer. else do not propose mission, string finished). Jump to "find a guy string" and return after completion to 4.
  3. Recipient fraction ask you to investigate whether there is more of attacker's faction F#2 guys in a neighboring system where it needs to transport goods (if no such need exists, do not propose mission, string finished). Jump to "explore string" and return after completion to 5.
  4. If you could get the info, goto 6, otherwise, goto 7 with 50%, finish mission string otherwise.
  5. ...
Find a guy String
  1. 20% chance: mission giver (Faction or NPC) give you indication about the system where the guy is now (he may move before you arrive); 20% chance: F#1 has indication of where the guy is going to (if not moving, where he is now); ...; 10% chance: no indication but the name/ship name of the guy (so scan around to find him)
  2. When you find him and contact per radio: 50% chance he asks you to land on nearest planet to meet in pub (goto 3), 10% chance he ignores you... etc...
  3. You discuss with him. 40% chance he gives you the info, goto back to Main mission string. 10% chance he tells you he does not know but another guy could inform you (check if another NPC of attacker's faction F#2 exists and select this guy as informer. restart "Find a guy string" else he gives info, goto back to Main Mission String), 30% chances he request a service from you (goto "deliver cargo mission") and will tell you afterwards, 20% he refuses to tell.
  4. ...
So the mission is actually getting "built" as you follow it, based on some probabilities, on the availability of assets, and on the impact of your actions. The key is that e.g. the "find a guy" step 3 can possibly start another "find a guy", which itself may start a "deliver cargo", which itself may request to find a guy because the customer was no at the expected target planet and so on (probability is decreasing, however, so not all missions will end up in very long strings), and so on. So you get possibly infinite stories - or very short one, you don't know when you start. And all based on a limited amount of building blocks (like "find a guy", "explore a system", "deliver a cargo", and so on) that can be re-used many times.
And as DWMagnus mentioned, the reward accumulates, for instance the "goodwill" of the client faction - if it accumulates enough, it may get you adopted :)

For the logical consistency of the mission string, you will need to store "pointers" to assets like
Client faction F#1
Cargo Transported C#1 ("potatoes")
Ennemy faction F#2
Informer NPC#1
Cargo Transported C#2 ("information about the mood of <F#2>")
...
so that you can built mission introduction text, find correct assets (e.g. return to a fleet of your client faction F#1) and other necessary texts with correct reference. For example when contacting the informant (replace elements in brackets with appropriate name):
<NPC#1> stand up when you approach. 'Hi you, <Player_Name>, you were the guy attacked by <F#2> while transporting harmless <C#1> for <F#1>, weren't you? Got your finger in a complex situation now.' You nick and tell him: 'You know why the tried to stop me?' ...
The whole art will be to make building blocks and text blocks smartly enough to make the player feel it is a real, consistent, mission, while being generic enough to allow the building blocks to be used in many situations. For instance the sentence above is not so good, because it does not allow me to reuse the building block "find a guy" in a context where finding the informer was not linked with a cargo transport mission that got attacked... So a new formulation may be better. Or a different "find a guy" building block made for more generic uses.
Such texts probably will need to have some randomness as well in the generation in order not to feel to canned. Possibly one could "crowd-source" a dictionary of sentences where several formulation can be picked (in categories, like "formal", "colloquial", "vulgar", "strange", etc...) to add more color.

Now there are many refinements that would help a lot, like also having "ranks" (arbitrary). E.g. you could make a normal cargo run. If the client faction is a faction that is mostly active in trade (somehow the engine will be able to characterize factions), than if you complete the mission an operative may tell you the faction may use your service again and give rank of "provider of service level 1 for faction F#1". This may then be a condition when generating branching missions further on.
Image
Post

Re: Missions, quests and epic stories

#8
Ah, some excellent ideas coming out here!

There are probably at least two useful methodologies for constructing multi-arc plots.

a) Simple paramaterised story:
There is a fixed sequence of arcs, but each can be fed with parameters for localisation, characterisations and and sub-constructs.

An example: You are tasked to travel to [parm1] to meet [parm2], who asks you to take him to [parm3] to collect (or action parm) a [parm4] before [parm5], and then to take them both to [parm6]. You may be challenged by [parm7] (random to determine if it happens) and you must choose to 1) fight and defeat [parm7] or 2) hand over [parm4 & parm5] to them. If 1) and you win, the reward from [parm2] is greater, else you are dead! If 2) you are rewarded by [parm7] but you have a bounty on your head by [parm2]'s faction.

b) Randomised (weighted, filtered) arc story:
One [Stage1] arc is randomly chosen, parameterised appropriately. This has a one-to-many child relationship with several [Stage2] arcs, of which one is randomly chosen (maybe weighted due to current status). Similarly [Stage3] etc... until the current stage has no further child references available, i.e. the end. This could include conditionals where your choice of action determines the choosable options for the next stage i.e. the weighting becomes a filter.

The stages would have to be carefully designed so that appropriate next stages make sense. It may even be possible to be clever, such as merging two sequences; e.g. in Quest1, A asks you to assassinate B, for a reward. Before completion you pick up Quest2, and a subsequent stage of that tasks you with assassinating A and be rewarded by B. You can't do both and get rewarded, so you're faced with choices.


The other thing is that this may allow third parties to provide further plots or plot constructs i.e. quest mods become possible.
Post

Alternative mission system idea

#9
Firstly, I know this already has a topic, but I feel my idea is too different to previous suggestions to be a reply (which is what it initially was).
{alternate suggestion to a common problem so... *shove* - Gazz}

I have another suggestion as to how the epic story arch style of mission could be procedurally generated; and it is related more to the way you think about what a "mission" actually is.

To begin with, lets think about what a mission is (or at least what they are in my view). A mission is a task set be someone for you to achieve (or not) to achieve a certain end with a reward upon completion of said task. When considering the procedural side of this, often the ideas are based around how certain "chunks" of a story can be fit together to show a long seemingly continuous flow of story or action.

But wait a second! Why are we trying to fit a story together with predefined chunks when we can make an entire universe with a single number?

My idea for the "mission generation" system is to not have one in the normal view of the term. Instead, what about a system that takes a powerful diplomats self interest, a military commanders orders to destroy a foe, a corporations need for expansion, a pirate desire for booty, or a knowledge companies need for the unknown, then splits up this goal into definable chunks at a high level then converts them to tasks for any of that entities associates to achieve (and be compensated for do so).

Let me give you an example:

We start with a seed: "Evil Corp wants to expand into the Omega system."
For this to happen many, many this need to occur. To list a few:
- Trade lanes need to be established
- Officials need to be bribed to allow Evil Corp to continue its less than legitimate business dealings in both the Omega and its currently controlled systems
- Pirates need to be cleared from the sight of the planned headquarters
- Large amount of resources must be sourced to begin construction of the required infrastructure.
- The CEO need to be shuttled comfortably and safely to the sight to overview construction
- Any existing competition need to be dealt with in the manner for Evil Corp!

Now the magic happens. The player isn't the only person in the game who works for Evil Corp. In a living universe that Limit theory offers, we should expect the employees of Evil Corp to be allocated with most of the work for this expansion; however, maybe one of the workers is feeling sick, or has a hot date, maybe Evil Corp is a man down after a raid. Whatever the reason may be, the players capabilities and trustworthyness is assessed by someone and can be assigned a role in this expansion.

Of course, the player may not even know what is going on. They could just be approached by a man in a bar, not wanting to fly through space on a routine spaceiron ore delivery and lets the player know there is 500 spacemoney for someone who delieves 500 spacetons of spaceiron to the Evil Corp warehouse. Then maybe on arrival, a manager notices that the player is kitted out as more of a merchant/fighter/explorer/builder/whatever and knows that Evil Corp has a task that needs doing. Then another, then another.

This level of detail can now be improved procedurally. Evil Corp, as part of its expansion wants to setup a new private trade lane. By now our player has had a fair few dealings with Evil Corp and has quite a reputation of being a excellent explorer, but can also handle him(her)self in combat; or maybe Evil Corp pays them a salary and simply expects the job done (allowing the player to palm it of to others as well setting up "missions" for other NPCs!). The player is then called up and asked to scout the proposed trade lane. Tee player does so, but it turns out that the proposed route travels right along a pirate nest (note, maybe the player doesnt report it, maybe they dont even check, the concequences are endless) the players "boss" then decides how to go about dealing with this issue. Could the pirates be reasoned with? could they be bought off? could they make vauable employees? The boss then decides that the pirates must be eradicated, completely, lest they present competition to Evil Corp! "I want those pirate scum DEAD! I dont care what it takes, get rid of them!" Now the player could attack them himself, or hire some mercs, or comtact other employees, or anything really!

The example continues, until the pirates are gone, the trade lane warp gates are complete, the area mapped, the permits aquired, so on and so on. Depending on how "well" the player performed in the eyes of his boss, he is allocated better, worse, different jobs and could gain or lose his trust, and just like that, a huge impact on the universe is achieved when Evil Corp now controls all trade though a certian sector of space, due in part to the players efforts! maybe that player is now also responsible for minions that do jobs for him, maybe he now runs part of Evil Corp (given high level jobs by the CEO), maybe he has gained enough trust AND ACTUALLY WORKS FOR THE SPACE IRS!

Hopefully from my example you can see how this would be adapted to other situations, like a military commanders orders to attack another civilisation, etc.


The crux of what I am saying is that the "mission" system in a procedural game can be much more abstract than predefined chunks. Missions become the outsourcing of portion of the will of a person to the player which he is seen to be capable and trustworth enough to complete. This would also allow for amazing engagement when the player is able to set missions for other NPCs to do, allowing the, to "prove" themselves to the player (or maybe they are SPIES!).


Lemme know what you think!
"Every vision is a joke until the first man accomplishes it; once realized, it becomes commonplace." ~ Dr. Robert H. Goddard
Post

Re: Alternative mission system idea

#11
I really like this idea - far beyond any simple linear quest, it's a top-down hierarchical task generator based on a set of initial premises.

I feel this approach could really come into its own if it could be combined with human creativity i.e. a quest engine that can work from human-designed goals. While I'm aware that LT won't have mods per se, perhaps a script-fed goal-based quest engine might be possible? I'm aware any such mechanism would need a lot of thought to get it to work in a balanced way. Ideally it would spit out sub-quests that the player can follow, as well as generate a fair amount of NPC activity.
Post

Re: Alternative mission system idea

#12
@Rhino_aUS
I see what you're saying and understand what you mean.

However, the reason I broke missions into 'chunks' is because there is a very finite amount of mission types that can be achieved. More of a 'greatest common denominator' type combination (since you really don't need to go into detail of the what to do during a certain type of mission).

1) Scout/Go somewhere
2) Fetch quest/Deliver
3) Destroy
4) Escort (subset of #1)
5) Process task (repair ship/mine minerals)

Even though you can break these down into their smaller parts using what you have, I'm tackling this from a high level, and more from a 'What' perspective rather than a 'Why'. Josh already stated that missions aren't spawned based, so I'm guessing the PCG engine would have to do pretty much the majority (if not all) of what you proposed in order for the missions not to spawn the enemies.

I guess I took it for granted that the low-level aspects (the ones you defined) would have to be in place in order for some of the high level things to become available.
Image
Early Spring - 1055: Well, I made it to Boatmurdered, and my initial impressions can be set forth in three words: What. The. F*ck.
Post

Re: Alternative mission system idea

#14
Rewrite of (my understanding of) the original idea into a more general concept:

1) NPC using the HLAI decides on a long term goal (e.g. expand into area X)
2) this long term goal is split into smaller tasks
3) these tasks are handed out to individuals to perform - and these are the missions

Tasks may be handed to other faction members (who may themselves subdivide or pass on the task), if a player arrives then they may be offered the task.

Handing out of the tasks would be based on some sort of cost / benefit analysis.

If the NPC in step 1 is a faction leader then its more likely that a lot of the tasks will go to faction members (as the cost should be lower to them), if they are just an individual then more likely to use lots of hired help (as they don't have the resources to do themselves).
Post

Re: Alternative mission system idea

#15
Welcome, Rhino! And congrats, you've just reinvented the LT mission system :D

My concept for mission generation is almost exactly what you've written here...so there's really nothing "alternate" about this idea! ;)
Bryce wrote:1) NPC using the HLAI decides on a long term goal (e.g. expand into area X)
2) this long term goal is split into smaller tasks
3) these tasks are handed out to individuals to perform - and these are the missions
Literally, this is exactly how it will work (assuming implementation goes well). I'm pretty sure I detailed this in an interview somewhere. But the system you've described is exactly what I am working towards!

So...that's that I suppose :)
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford

Online Now

Users browsing this forum: No registered users and 5 guests

cron