Sign in to follow this  

Non-linearity, a second look

This topic is 4714 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I have no idea of whether this is a new concept or not. Given this particular industry, it is probably not. In any case, I am just looking for feedback. Take a typical, "non-linear" story/game. For ease of communication, we will depict the story model as a tree. There is usually the main story (the trunk), and the offshoot stories/plots (branches). Now, in the case of most games, whether or not you can access story branches at any given time, they are often limited, and you must still follow the main story in order to progress the game. I am sure there are more closer-to-the-truth models out there, but this is how many developers define their games to be "non-linear." Reasoning this way, then in my point of view, Diablo II wasn't linear. Which it is. But what if we chose a model where there were several main stories, or "trunks"? Even though these trunks themselves might perhaps be linear (with possibilities of branches, of course, and maybe crisscrossing - interacting with another trunk), on a macro level this would be much closer to something we could at least call emergent non-linearity. Of course, independently of our story-line of choice, there would still be a general direction much the same for them all. It's simply different ways of telling one story, from different perspectives. This would, undoubtedly, mean a lot more of work. But in disregard of that, what is your oppinion on this? Are there already games with this kind of a story model? EDIT What we are looking at here is a system/model to be used in a single player game. I am sorry I forgot to jot that down the first time ^^ [Edited by - mkallman on January 16, 2005 3:18:10 PM]

Share this post


Link to post
Share on other sites
all the truley non-liner game i have played feel like im just doing random errans for people for no reason other than a few gp. I would be a intersting game if you had a decent storyline but could feel like you were actually in control of it
I have played a few story games that give you options but there are usally only a few branch points or all the option have the same end result, this is probably becouse as you already stated it would requre exponentially more work

Share this post


Link to post
Share on other sites
Quote:
But what if we chose a model where there were several main stories, or "trunks"? Even though these trunks themselves might perhaps be linear (with possibilities of branches, of course, and maybe crisscrossing - interacting with another trunk), on a macro level this would be much closer to something we could at least call emergent non-linearity. Of course, independently of our story-line of choice, there would still be a general direction much the same for them all. It's simply different ways of telling one story, from different perspectives.


well, take half-life, for instance. they had the original game, and then other versions were made where you ran through the same story from different perspectives (guard, army guy). in that case, you kinda have to pick who you are at the beginning, whether by menu or a more masked "career day" sort of level. chances are this would be the case in most games that were to try this method, since its a little hard to suddenly switch in the middle and still have it make sense.

something like morrowinds houses might work nicely, only incorporate that as a big part of the main quest. there are three different houses, all with their own benefits and weaknesses, all at odds with each other, and all with their own missions (even though there were really only a few missions in that whole game, just different people and places to do them in). at a certain point in the story, you HAVE to choose a house, and that affects the things you do in the rest of the game.

as far as "running errands" for people, you could do something like a mob type element, where you start out doing things for low level thugs, and gain renown, and end up working with the big boys for far better pay. that would give a little more incentive to actually do side missions instead of just killing the guy and taking his stuff to pawn off later (usually that pays better :-D). of course, if you DO kill the guy and take his stuff, maybe the mob puts a hit out on you. and no one wants that.

Share this post


Link to post
Share on other sites
I think the next step in multiplayer non-linearity would be to make X version(s) of the same quest. If you are Group2 to do it, some of the changes by Group1 would still be visible.

You took the example of DiabloII, so I'll stick with it. Imagine that you need to kill Diablo in Act4. In order to do that, you need to open 6 gates first. Let's say Group1 didn't succeed in its first attempt. Well, maybe in your "instance" your party would see dead corpses of Group1 with still some gates activated.

Sure DiabloII was more real-time than a normal MMORPG, but if you start with this idea, you can mix it with your mega story tree of doom and it would be some sort of "new" step.

Share this post


Link to post
Share on other sites
Nut sure if it's the same as 'tree trunks' or different, but we've been describing _Xenllure_'s plot structure as a tapestry, where there is one strand for each major NPC, the y-axis represents time, and the player can slide throught the tapestry along any direction as long as they are also moving down (forward through time and plot arc). Technically our tapestry should also have all the strands starting together in a big knot at the top to represent the game's single start point.

Share this post


Link to post
Share on other sites
Well, what do you want to be different? Another slant on "non-linearity" is that you want the same relative story time location (story(t)) to have different meanings based on the player's actions or inactions. The kinds of differences are really up in the air, but so far the "meaningful choices" we've seen in prior games have included dialogue, mission goals, rewards, and level path. The latter is used very infrequently, and the most notable example that I can think of was Marathon: Rubicon, a mod released several years back.

But I digress. Going back to the story(t) bit, as sunandshadow said, the hope for nonlinear games is that you have several possible storylines woven between eachother. In math terms, story(t) is not a function because story(t)={a, b, c, d} and we can be at one of many locations at that point in time. The catch for being here rather than there, however, is in the choices the players make.

So here's what I think would work best for this: write a series of scenes, that only contain or are concluded by a clearly defined choice. You can run and gun down the stairs, or you can jump out the window to escape. You did or didn't bring back that heavy piece of evidence that proves the duke's sister's treachery. Whatever. Write and design a storyline off of both results, label it appropriately, and put it on the stack.

Your content developers will hate you forever, but you have a nonlinear story. I don't think it's done very frequently commercially for one very simple reason: money.

Share this post


Link to post
Share on other sites
Quote:
Original post by Sta7ic
Your content developers will hate you forever


It doesn't have to work that way. The only major difference duplication of content in our desgn is in the script and possibly the scripted scenes/FMVs; all the settings, models, etc. are the same for all paths through the plot.

Share this post


Link to post
Share on other sites
Quote:
Original post by sunandshadow
Quote:
Original post by Sta7ic
Your content developers will hate you forever


It doesn't have to work that way. The only major difference duplication of content in our desgn is in the script and possibly the scripted scenes/FMVs; all the settings, models, etc. are the same for all paths through the plot.


Granted, it doesn't have to work that way: I'll cop to exagerating. But depending on how drasticly one branch veers from another, you may (or may not) end up with extra stages to set and dress.

mkallman: Don't'cha hate how when you hint "role-playing game" or "story-driven", the unwashed masses immediately think "massively multiplayer"?

Share this post


Link to post
Share on other sites
Perhaps that is so. I myself only think of the topic :P In any way, it hardly matters: the only difference would the suggested persistence of, for an example, a Diablo II quest.

Share this post


Link to post
Share on other sites
The problems with non-linearity (Well, closer to 'real' non-linearity at least) are keeping the a high quality level, and avoiding the 'digging a pit' problem.

Think of a game like Half-Life 2. When you're not jumping out of a dam in a makeshift speedboat, you're having... err... 'coversations' with the few remaining rebels and scientists, driving a hobbled-together buggy through dangerous sand dunes and then hiding from a hovering gunship. High quality all the way, each area tweaked for months on end.

When you start approaching it from a non-linear angle, you cant spend the same amount of time making every element operate like clockwork, you've got to make concessions. The more possible paths is inversely proportional to the amount of time you can spend on any one path. Generally unique situational animations are the first to go, followed by a good set of unique events and probably followed by voice-overs. What you're left with is a character that moves the same as a lot of the others and reading plain text to move the story along.

A lot of people dont have an issue with this, normally in an RPG vein because the RPG genre seems to mostly rotate around plot, exploration and battle systems as opposed to the 'wow a minute' that a lot of FPS games try to achieve, but they're seperate goals. If you want a game with a deep, involving, intricately-woven story, but dont care as much about production values, go with an RPG. If you're after something that will constantly keep you on your toes and will introduce cool things one after the other, go with an FPS.

The other problem is the whole 'Digging a hole' syndrome. When you give the player the freedom of choice, chances are (If you're a fatalist like me) that the first thing they're going to try is the one thing you didn't think of. Gameplayers are notorious for trying to break games the second they get their hands on them, so you better make sure your story is air-tight. If you could kill any NPC, for example, what would happen if you killed all of the pivotal NPCs to make the story move forwards? Of course, it IS possible to avoid situations like this, but the more choices you give the player, the more playtesting you have to do to make sure the player cant get into a situation where, basically, you need to get the key to open a door, but the key is on the other side.

I'm sure it would be possible to create a game with hundreds of branches off the main story tree, but anyone who tries is either a fool or a goddamned genious.

Share this post


Link to post
Share on other sites
I don't see much wrong with "breaking the game". If the key is on the other side of the door it opens, then tough rocks, pal, you don't get to go in there. In Fallout, you COULD kill all the vital NPCs. It made for a less interesting story, but you got some good fights out of it. You had to do this long, difficult quest to get into the Brotherhood of Steel's fortress, but if you had a high enough lockpicking skill, you could just sneak in.

This is a good example of what you describe, though, PlayfulPuppy. If you sneak into the BoS's headquarters, they all talk to you as though you finished the quest and earned entry. There was another time that I had a mission to help a person and one to kill her going on simultaneously, and so she got her group together, followed me to the place where I was going to help her win a fight, and then accused me of being an assassin. Then her graphic glitched out and her team stopped moving. That was embarrassing.

So yeah, I guess it's a lot of work to script all those events so that they work in every possible combination. But to be honest, I didn't break the game until about the fifth time through. You have to know where the triggers are if you are going to bugger it up.

To make it work best, maybe some sort of organic system, where events are less scripted, would be good. Lose the voice acting, put in a dialogue generator (huge hassle all by itself), and let the characters build their own world.

Ridiculous, to be sure, and I'm not going to write the code for it, but it seems tlike the sort of thing that would have a "critical mass". If you put in th enormous amount of effort to build a vibrant virtual world, then you reach a plateau in the process, and so making a vast world is only 20% more work than making a little one.

Sounds like a job for science.

Share this post


Link to post
Share on other sites
Quote:
Original post by Iron Chef Carnage
To make it work best, maybe some sort of organic system, where events are less scripted, would be good. Lose the voice acting, put in a dialogue generator (huge hassle all by itself), and let the characters build their own world.


Perhaps social AI will solve the question.

On a different note, it is true that intensity may be lost if you have an open-ended world/story. Thoug, ultimately, it comes down to how good a storyteller and designer you are, I think. This will merit some serious thought.

And, most of the bugs in software are actually due to the faulty, traditional development approach: code, extend, code, test. On our project in particular, we're trying to at least reduce bugs by utilising a test-driven approach. But this has got nothing to do with game design (unless you incorporate the whole process).

More perspectives, anyone? =)

Share this post


Link to post
Share on other sites
Quote:
Original post by Kaze
all the truley non-liner game i have played feel like im just doing random errans for people for no reason other than a few gp.


I'm not a big fan of procedurally generated high level content for this reason. The elder scrolls games did this a lot. They had a lot of good points and a lot of depth to them, but the NPCS, dialog, quests, settings, and dungeons all came off feeling very bland and generic to me. Everything was lacking in personality. I know there are a lot of people who loved Morrowind, but for me I like the feeling of a rich handcrafted world.

On a completely different note, I think the middle Ultima games (4 - 7) had an interesting approach to a non-linear story. These games had a very rich hand crafted world (6 and 7 especially) with unique npcs, original dialog, and lots of things to discover. However the order you explored the world was pretty much up to you. You could head off in any direction and find interesting things to see and do, and it felt like you had a ton of freedom. Of course the story itself was very linear in the sense that the story really had only one path and one ending, it didn't branch much at all, but the fact that you could explore the story in any order you wanted made it feel much more free.


Share this post


Link to post
Share on other sites
alot of the ultima games seem like that, i particuarly liked Ultima Underworld 1. Sure there was only one ending, but i could go anywhere in the dungeon i want to (whether i'd survive or not was another matter altogether). ;D

Share this post


Link to post
Share on other sites
Quote:
Original post by Iron Chef Carnage
To make it work best, maybe some sort of organic system, where events are less scripted, would be good. Lose the voice acting, put in a dialogue generator (huge hassle all by itself), and let the characters build their own world.


And not just a dialogue generator, but level generator, mission generator etc. I'm with Iron Chef Carnage on this one. Don't model the outcome, but rather the system; just as with particle systems you get more diversity and awe on graphical effects than if you just had the same explosion animation every time something blows up. Rely on emergent behaviour, don't have a strictly pre-written story. Note that you can still control the story to a degree by adjusting the probability distributions of the generated levels, missions etc. You could still have scripted sequences, but generate the scripts randomly.

As for the game becoming "broken" if the player does something unexpected, I'd say it's rather an inherent flaw in pre-written stories than non-linearities and freedom. You should not assume anything. I'm not saying you have to consider every possible player action (as you really can't), but rather to build the system that the story of the game creates itself while the player is playing. Assume there is an NPC that the player needs to talk to in order to accomplish something. If the quest/whatever related to that NPC is critical, then you could have a backup plan. Create a new NPC in the world with a similar personality and quests, maybe a relative of the late questmonger. Or turn the NPC's information into tavern gossip, so the player can find it out there. Or whatever. Don't make rules of the sort "NPC X must not die", unless it is specifically the mission (eg. a quest to protect someone). If the quest/whatever of the NPC is not critical, then big deal. The player won't miss a side quest as new ones get generated all the time.

The mission generator should do much more than to turn ordinary NPCs into questmongers with quests like "fetch me a Sword of Mayhem +2". You could first create a problem that needs to be solved, eg. a humongous green multi-legged creature terrorizes the countryside. Create stochastically effects, such as NPCs start to disappear, villages are deserted, put a randomly generated lair of the monster somewhere on the map where the player hasn't explored, or whatever. Create information about the quest, such as tavern gossip, people who have seen the monster, piles of corpses near the monsters lair, the king promises a reward to anyone who slays the monster etc. Then, if the player doesn't act, maybe a village or two die out, and maybe then some other hero slays the monster. More gossip follows etc. If the player actually kills the monster, his reputation increases, local villges will regard him as a hero etc. Make the player feel that the world is dynamic and there is more happening than just the player's actions, and give depth to random quests. Random quests should also have prerequisites, such as the questmonger having a certain personality etc., in order to prevent obviously silly quests.

Quote:
Original post by Kaze
all the truley non-liner game i have played feel like im just doing random errans for people for no reason other than a few gp.


Not to mention that usually these random errands are of the order of "I seem to have lost a cow. Some broken twigs were found near the forest. Quickly, there is no time! Accept/Reject?", which is pretty much the complexity of the random missions in Sacred. Heck, if you accept the mission the game actually gives you the exact location of the objective.

But this just means the system was done badly. I wouldn't say it was an inherent flaw of emergent behaviour per se, but rather that in order to achieve an interesting world you need much more than just a questmonger and a single objective. These missions should have more effect on the game world than just the fact the player gained a lousy amount of cash.

Share this post


Link to post
Share on other sites
Quote:
Original post by Grim
And not just a dialogue generator, but level generator, mission generator etc. I'm with Iron Chef Carnage on this one. Don't model the outcome, but rather the system


There is only so much that can be done with today's technology, let you put together a team of twenty scientists. What we do have is a society and personality/emotion/behaviour synthesis system. Perhaps "mission generation" could be included into that, at the least. Very interesting post, though! ^^

Share this post


Link to post
Share on other sites
Quote:
Original post by mkallman
Quote:
Original post by Grim
And not just a dialogue generator, but level generator, mission generator etc. I'm with Iron Chef Carnage on this one. Don't model the outcome, but rather the system


There is only so much that can be done with today's technology, let you put together a team of twenty scientists. What we do have is a society and personality/emotion/behaviour synthesis system. Perhaps "mission generation" could be included into that, at the least. Very interesting post, though! ^^

exactlly, the original point of this topic was a non-linear game with a good storyliny, the primary problem with them is they do use ai/level/dialog generators, but game AI cant write storylines or even carry continuity(unless its pre scripted)

Share this post


Link to post
Share on other sites

This topic is 4714 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this