level design

[anubite]

I um, decided to do this just because. I hope the developers are mindful of what I want to show here, but I imagine they are. Right, so. Here.

 

 

What I want to illustrate here is mostly from my memory, so it's probably not 100% accurate, but basically, you follow each of these colored paths to correctly navigate the dungeon and complete it.

 

This is the very first part of BG2, mind you. The very first zone in the game. I think you can notice and appreciate how non-linear it is. There are side-rooms you must investigate and places you must explore. Notice how the red passages make up the center of the level and how doors are locked until you get key items to open them. At the major juncutions of the game, you can't proceed without exploring all facets of the level, but you still have choices to navigate. The level design here is very simple and straight-forward, but on your first play-through, you'll definitely be wandering around a bit. The level, in its "solved" state is pretty linear (there aren't any choices for our player to make here - they need to get certain key items to progress and there is a certain optimal order and way to clear this zone), but for a "tutorial" area, it's pretty ****in' neato.

 

Let me compare this with an extreme, well-known case of linear level design.

 

 

That's a pretty significant portion of Final Fantasy 13, which is now infamous for its nearly 100% linear level design. Every area is a literal corridor that you must walk down. This is a very extreme obvious case of linearity in level design. But I want to use to illustrate a few points.

 

1. In this case, "linearity" is literal. We have nearly straight lines composing levels. But when I talk about "linearity" in this post, I don't mean that at all. A level can be linear and use only curved, parabolic, or non-euclidean shapes. What is meant by linearity is sort of a kind of "graph theory", that is to say, there is a solution to a problem, and that solution is singular. You can create a path of least resistance. There is only one way to play the game. That is what is meant by "linearity" - that a level can be broken down into a superficial solution that requires little to no thought.

 

2. Linearity is bad. I honestly cannot come up with a good reason to have a linear game. Maybe there are good practical applications for it, but I can only see the negatives at the moment. Why is linearity bad? Because everyone experiences the same thing when they play the game. A linear game will play the same way for everyone; there is no emergent behavior. It's like writing a computer program that can only do hard-coded things. It's worthless. What's the point of making a video game if players cannot make choices or deviate from what is expected? Linear games often feel too restricting and are very tiresome in how there's very little room to /play/ with the mechanics of the game. Linear level design is determinism of the worst kind.

 

In the case of Final Fantasy 13, Square was pretty much humiliated and the "splendor" of the Final Fantasy series is a lot grittier ever since. At least, in my eyes. Not that Final Fantasy was ever known for non-linearity, but go play the first Final Fantasy for the NES, there's a lot of freedom of choice and movement, even if the game is rather linear.

 

3. RPGs as a whole, have been suffering from linear level design ever since World of Warcraft got popular. I don't know if it's just a coincidence or what, but I think we can all agree RPGs need to offer some degree of non-linearity.

 

 

Just another example of excessive linearity in RPGs. Here is Dragon Age: Origins, which, while not nearly as bad as Dragon Age 2 in its linearity, still has many levels which are practically straight lines.

 

But allow me to put what I mean in a little more abstract and mathematical manner, so I can better define what linearity is.

 

 

Imagine the red rooms are "encounters" or "rooms" in a dungeon. Imagine the black lines are hallways that connect them (or some other method of navigation, like a teleportation trigger).

 

Example A: This level is linear if and only if there is one entrance and one exit to this level layout. There is one choice to be made at the beginning of the level, where you take the left branch or the right branch. You explore either branch of the level and find the exit. Since there is only one exit, one branch of the level contains it. The level is linear because there can be only one solution to it. Because there is only one choice to be made, the complexity of this level is also very low. You could actually simplifiy this entire 'dungeon' into 1 room connected to two others, one containing the 'solution' or exit to the dungeon.

Example B: This level is linear under the same circumstances. If there is only one way to solve this level, you will find it exploring one of the branches. The point of this is to get across the idea that ADDING SIDE ROOMS DOES NOT MAKE A LEVEL NON-LINEAR. The numbered rooms all have "choices" - two to be exact, but these choices are imaginary. Since there is only one solution to the level, once you replay the game again, you will know the optimal way to beat the game. I'm not saying example B is necessarily a bad way to design a dungeon if you only intend players to play a game once, but I feel video games can and should have a high replayability factor. Furthermore, all players will perform the exact same solution as any other, all that will change among plays of the same RPG is how many wasted steps were taken. Nobody is going to creatively solve the dungeon above, you either find the exit, or you don't.

 

 

Example C: Okay, now we're entering non-linear territory. Let's say the 0th room is the entrance to this dungeon. The "E"'s represent a solution to the level, they are the exits. This level is non-linear because the player has two concrete choices to make: does he take the left path, or the right path? Let's say the purpleish rooms represent very dangerous encounters in example C - this means, if he takes the right path, he has to fight a dangerous enemy RIGHT AWAY, this makes the right path a lot scarier to a player doing the level for the first time. He may choose not to take the right branch and instead take the left branch, which is fine. He will continue on and explore the dungeon further. He'll find a consumbale item in the room marked "I", making the dangerous encounter in the room ahead easier. The exit to the dungeon on the left path is just past the difficult encounter.

 

By the very nature of this design, the player had two choices to solve the level, each choice has its own ups and downs - the right path actually has one less room to clear before reaching the exit. But, taking the right path gives you access to an item, which you could always hoard for later.

 

This is non-linear level design. Note how all you really need are two solutions to a level and branching paths (your level isn't very non-linear if the two solutions to the level are connected one identical room, or other silly design decisions). A similar kind of design can be accomplished by turning this level on its head: What if the "E's" represented ENTRANCES and our 0th room is our exit/goal? The level is non-linear then again, because the player had two ways to access this level from the outside, giving some level of player agency to choices made prior to entering the level.

 

VTMB is well-loved for its non-linear level design. One reason why that game is non-linear is because nearly every major area has a sewer system, allowing you two or more entrances into the same combat area. Similarly, VTMB often allowed you to complete its "dungeons" through the use of lock picks, stealth and persuasion, not just combat. Rewards were dolled out for completing tasks and not killing enemies or fully exploring areas, meaning, you didn't feel like you "missed anything" by finding the best solution to an area for your character (though, there are some ups and downs to solving levels in various ways, but I don't want to delve into that).

 

Example D: This is how you'd attempt to implement some kind of computer algorithm. You assign rooms "difficulties of traversal". Notice how each path adds up to 7, meaning, each branch is considered to be "equally difficult". If branches are "equally difficult" in sum, then no branch is an optimal solution. Meaning, the level is truly non-linear, the player must decide what is the best solution for the current situation of his character/party. The numbers here are just whimsical, but they could refer to enemy spawn counts by the computer program, or they difficulty of various tasks in the room (solving a puzzle, having sufficient lockpicking skill, et cetera). So, in theory, non-linear level design is so simple and intuitive, designers could just write a program to do all this stuff. I'm not asking for all that much here. This isn't rocket science here, I'm sure most computer scientists/programmers are aware of what I'm trying to suggest here.

 

So thanks if you bothered to read through this tripe. It might read a little heavy-handed, but I'm honestly sick of being spoonfed my RPGs. Go replay Baldur's Gate and appreciate how circular the level design is there - de'Arnise Keep is a particuarly interesting level to examine under what I've described above - what's the optimal solution to the level? How many ways can you approach the level? How many "solutions" are there? How many entrances are there? Why is this early zone in the game so well-designed, and why haven't other major RPGs implemented what has already been done before?

 

If anyone has any C&C I'd like to hear it - I want this post to be as concise and clear as possible because I really want to pitch the ideas in this thread to whomever game designers/programmers will listen. Is everything here crystal? Am I right about how level design should be gone about? Is this the kind of level design you want Project Eternity to have? Or do you not mind the kind of level design we see in FF13 or "The Dead Trenches", above?

 

I will grant that SOME amount of an RPG has to be linear - we need "choke points", otherwise games get a little too complex. A game like BG2 followed a linear story, I'm not asking for macro-non-linearity (which is something a roguelike could only deliver without spending millions on development), but 'micro-level linearity' - non linearity for dungeons and the areas that connect dungeons to the game/hub world.

Edited June 17, 2013 by anubite

[JR.9613]

     To begin, everything made sense to me and, to my understanding, you got your point across. I agree with most of what you say except for a few things.

 

You say that "Adding side rooms does not make a level non-linear" and I agree with that. However, the same can be said about adding an extra path (or three or four)  through the dungeon. I don't think that having multiple paths through a dungeon really matters if the decision of which path to take doesn't matter. If it is a decision between fighting a goblin boss or a ghost boss then I don't think that a dungeon can really be considered non-linear, because, in the end, it doesn't really matter and you probably won't even care in a week. I would much rather have a "linear", but fun and compelling dungeon with memorable fights than play through a non-linear dungeon where the choices you make don't really have consequences.

 

What I am saying is; Quantity ≠ Quality.

Edited June 17, 2013 by JR.9613

[anubite]

     To begin, everything made sense to me and, to my understanding, you got your point across. I agree with most of what you say except for a few things.

 

You say that "Adding side rooms does not make a level non-linear" and I agree with that. However, the same can be said about adding an extra path (or three or four)  through the dungeon. I don't think that having multiple paths through a dungeon really matters if the decision of which path to take doesn't matter. If it is a decision between fighting a goblin boss or a ghost boss then I don't think that a dungeon can really be considered non-linear, because, in the end, it doesn't really matter and you probably won't even care in a week. I would much rather have a "linear", but fun and compelling dungeon with memorable fights than play through a non-linear dungeon where the choices you make don't really have consequences.

 

What I am saying is; Quantity ≠ Quality.

I'm not sure I understand.

 

If you're playing an RPG, shouldn't choices have consequences? There is a difference between a goblin boss and a ghost boss. A ghost might have "dark magic" and a goblin might do physical attacks. Meaning, if your party is more equipped to deal with the ghost boss, then the path you take matters, the fight you choose to fight matters.

 

The only advantage linear levels give is developers can put extra special detail on the linear path before you. But to me, a dungeon has no quality to it, if all I'm expected to do is walk forward and kill the next encounter. Fighting becomes a grind; a chore; a slogfest, when levels are designed like "The Dead Trenches" where there is no choice but to move forward and to always fight that group of enemies the developers have tailor made to fight you. There is no emergent play and there is no quality to the game, if there are no choices.

 

Now, the number of branches that you want does matter. If you have a dungeon with 99 branches, then making custom story content for that level is impossible, it will need to be generated procedural and I understand there are people who want a more "cinematic experience", but I think you can marry non-linear level design and that "cinematic experience" because Baldur's Gate 2, at least in the dungeons I show/mention above, does manage that. I don't know what the popular opinon is of de'Arnise Keep, but I happen to like the level a lot for what it is.

 

There's a big difference between a cRPG and a Roguelike, I'm not asking for RPGs to become roguelikes or anything of the sort, however, I think we can demand a little more weight to our choices than "always going forward".

[mcmanusaur]

I honestly don't really find linearity in the physical layout of dungeons to be too much of an issue (it's all just more "dungeon" to me), as long as there aren't artificial boundaries like unmovable furniture. But if you think about it there will always be structures that lend themselves to a "linear" design. If anything, I'd rather not have to do backtracking to clear an entire dungeon, and that's less of an issue in more linear dungeons; I'm more concerned about avoiding linearity in other respects (such as there being multiple solutions to a quest as you mention, but that doesn't necessarily require separate pathways).

[TrashMan]

Non-linear level?

 

I give you "harbor" level from Crysis.

 

I must have found a dozen ways to infiltrate the harbor and finish that level.

[anubite]

I honestly don't really find linearity in the physical layout of dungeons to be too much of an issue (it's all just more "dungeon" to me), as long as there aren't artificial boundaries like unmovable furniture. But if you think about it there will always be structures that lend themselves to a "linear" design. If anything, I'd rather not have to do backtracking to clear an entire dungeon, and that's less of an issue in more linear dungeons; I'm more concerned about avoiding linearity in other respects (such as there being multiple solutions to a quest as you mention, but that doesn't necessarily require separate pathways).

Right, I can agree back-tracking can feel a little dull, however, I don't see it as a significant issue if it allows the level to have an open-ended approach. Back-tracking SHOULD NOT be the way you make a level non-linear.

 

A clear example of excellent non-linear design is the first major mission in VTMB - where you get the Astrolite. The area is EXTREMELY tiny, but there are six (or more? I've only found) six ways to complete the task, with two ways to enter the building and several ways to distract or incapacitate the guards.

[Prometheus]

I don't like it, if I have to backtrack a long way in a dungeon, but I like it if there are more than one way to go to the "end" of the dungeon.

Josh Sawyer said that they want to avoid long dead ends. see quote below:

 

The cave level is being developed for technical purposes (developing organic cavern levels) and is small in overall size.  The level above it has two entrances (though in this dungeon they are relatively close) and can be explored non-linearly.  In terms of labyrinthine layouts, we will generally avoid long dead-ends in favor of loops (small dead-end branches are fine).

 

E: In my own campaigns, I always try to design dungeons with multiple entrances and multiple ways to explore them.  I'd like to do the same in PE when the size of the level supports it.

 

[Nonek]

I always like dungeons and other locales with their own little tales to tell, like the vaults in New Vegas. The vault with the lottery, think it was 11, had so much potential and tragedy from such a simple mechanic. That certainly hit you in all the right places, it left me with a burning desire to harm the Enclave in some manner, then of course I realised that judgement had allready come in the denouement of Fallout 2.

 

As for the actual level design, i've always preferred multiple approaches but also multiple areas inside the actual labyrinth. I often designed areas to be quiet, secluded and undiscovered for centuries. Hidden behind secret doors and cunning riddles these areas would feature glimpses of the past, of lost cultures and heroes etcetera. A nice little organic way of fleshing out the gameworld outside of a manual or journal entry.

 

Crumbling broken hallways, where earth falls with your every careful footstep (a fabulous method of splitting up the party,) and any open conflict would result in thousands of tons crushing the protagonists.

 

Endless rows of Sarcophagi, scrawled with undecipherable runework, the angular artwork of a lost civilisation and an unpleasant sensation of heavy watchfulness.

 

Heaving and boiling pits of activity, where monstrous abominations toil and tussle in their own filth, and you are forced to pass by in the shadows, hoping that none of the creatures will glance in your direction.

 

Desperate chases and dirty little melees as you run from a veritable horde, only to find your pursuers stopped and a dread feeling squeezing your very soul with its malign presence. You have stumbled into the lair of some forgotten dread bane, one that your enemies fear.

 

Coming out of the gloom, to be greeted by the kiss of high winds and the heady sensation of standing a thousand feet above the ground, the crumbling narrow bridges barely wide enough for a mans feet. While far above the winged sentinels spread their pinions and behold fresh prey.

 

A timeless place, a strangely dreamlike locale where vast wizardries were manipulated in an effort to tear it free from reality. The enchantment worked too well, the people of this place were trapped between the tick and the tock of the clock, while things that should not be, many angled abominations crept forth from the dark places between to feast and fornicate.

 

A hero of old, preserved by sorcery or alchemy, hanging in a cocoon of spiderlike silken strands that hum with life and warmth. The scars of vicious battle adorn his frame, mortal wounds it seems, and yet they are long healed. 

 

Ornate, decaying and richly furnished apartments where half heard whispers are percieved and disturbing images glimpsed from the corner of ones eyes. Luxuriant couches adorn every room, with braziers beside them still thick with the stench of the black lotus leaf. Here the priests of a slumbering god came to dream, and in so doing leach from the foul essence of their titanic master. Here the mind of a slumbering god is near, and some measure of awareness still exists. What dreams may come if one should rest in such a place?

[PK htiw klaw eriF]

I do agree that non-linear level design is optimal in most cases, although I wouldn't mind a few areas that are fairly linear.

 

Oh and OP, there is a thread in computer and console right now about DA3 actually. I think it could benefit from your perspective, seeing as you made a very long video about the flaws of DA2.

[anubite]

I do agree that non-linear level design is optimal in most cases, although I wouldn't mind a few areas that are fairly linear.

 

Oh and OP, there is a thread in computer and console right now about DA3 actually. I think it could benefit from your perspective, seeing as you made a very long video about the flaws of DA2.

My perspective was acknowledged in the very-long (and hopefully very dead) DA2 thread we had, where even a BioWare employee acknowledged my posts (though I don't know in what capacity they chose to read into them). I don't think my points would be appreicated there, as I've made them thoroughly in the past, and people think I'm too negative/hard on the series, which may be the case (but I don't believe so).

 

This is neither here nor there, of course, but I don't have any confidence for DA3, with BioWare's latest comments about merging "fast paced" gameplay of DA2 with "tacticalness of DA:O" -- you literally cannot have tactical gameplay that is fast paced. That's an inherent contradiction. All tactical games are slow paced, because all tactical games require thought. You cannot be tactical and make split-second decisions - that's twitch reflex and finesse, which is usually abstracted in an RPG. The fact BioWare cannot even understand why BG2 was a success, or how the RPG genre even works, suggests to me the game will only be saved if it can pull a Skyrim (make a huge ass open world and let modders fix the game for them). To put it simply: DA2 is a cRPG that wishes it was an ARPG, so it does neither of them right. If they try to do that again, DA3 is doomed to repeat DA2's failure.

Edited June 19, 2013 by anubite

[Jarmo]

Ok. It certainly was annoying how every damn dungeon or cave or whatever in Skyrim was just a pipe from A to B.

A bit less of that wouldn't hurt.

 

But I do also appreciate the convenience.

If there's something to be learned there, is how players (like me) really, really don't like backtracking from a damn dungeon.

Just work in something to give me a backdoor away and I'm all happy.

 

The real thing I really dislike though, is fake non-linearity. Just as in the first example of the first post.

It's like... you have 3 doors to take, but only 1 is the correct one and you have no way of knowing.

Pick number 1 and there's an overpowering monster and you die. Pick 2 and hey, you can't pick it because you dont have the key thing.

There's really just one correct path, it just takes backtracking, guesswork and multiple reloads to figure out which it is.

 

It's more a text or point and click adventure game thing than an RPG thing. 

Don't mind that time to time, but I wouldn't like to see such designs too often, the less the better.

 

True non-linearity is something else though and something I like.

It's just it's harder to tell a story if you dont know in which order A, B, C and D are going to come about.

 

ToEE was pretty great in giving a non-linear dungeon.

Too bad the dungeons were all pretty silly and didn't make any kind of logical sense.

[mkreku]

I actually don't mind if the dungeons (or levels) are linear, as long as they're scattered across a varied and truly open world.

 

This is how I've always figured you'd make an open world game with a great story: a huge open world filled with excellently designed dungeons (or levels).

 

The game that was the closest to this that I've played wasn't even an RPG. It was Far Cry 2! It had a massive open world and some really cool places you could visit. But the real stars of the game were the mission levels. They could be anything from airports to shanty towns to old mines. Unfortunately other parts of the game were not quite up to scratch, but they really nailed the open world/fantastic levels mix.

[anubite]

I actually don't mind if the dungeons (or levels) are linear, as long as they're scattered across a varied and truly open world.

 

This is how I've always figured you'd make an open world game with a great story: a huge open world filled with excellently designed dungeons (or levels).

What is a "truly open world"? What does that mean?

 

Open-world games can be highly linear, where you can go anywhere in the world, but you can't do anything in the world until you pass through some quest check point or trigger. Skyrim is full of linear dungeons that are boring; a lot of them have nothing to offer you unless you're on a specific quest, too.

 

A "truly open world" would follow the same design constraints I outline above, on a larger scale.

Edited June 20, 2013 by anubite

[teknoman2]

this is what i call a non linear level. you start at the green room and you have to reach the red exit. to get there, you have to find keys that open the matching color doors. black dots indicate a hidden path that can get you to the destination faster, by passing the deadliest hidden room (i just noticed that i forgot to add dots from the scull room to the last blue room), or will allow you to find a clue that will let you use an alternative exit that will lead to a different place than the normal exit

Edited June 20, 2013 by teknoman2

[JFSOCC]

TL;DR below.

Great levels have multiple points of ingress and egress, they offer different challenges based on which path you take, they reward exploration but don't punish you for missing areas, and can be done at your own pace.

 

How to design such a map? I don't know. I suspect I would build it not as a level designer but as a narrative designer (stick to your strengths ) IE: I'd build the Satrapal palace first, and then worry about how to create a challenge.

 

So, what does a Satrapal palace need?

Well, there is the impressive main entrance, well guarded of course

Then because the satrap loves his dalliances there is the secret entrance for his lovers, this is hidden, and unguarded but trapped. Interrogating one of his lovers may tip you off on it's location, and perhaps even the traps.

The Satrap loves to hunt, there is a gated entrance near the stables, it allows commoner traffic, but is some distance from where you wish to be. It requires bluffing to get past the gate, or sneaking, and then you'll have to find your own way into the Palace. (maybe you can tail a servant)

You can try and get a contract and be invited in, deliveries to the kitchen have a different entrance, it's not well guarded, but everyone there will immediately recognise that you don't belong there.

You can wait for an invitation for a party by the satrap, which would allow you to pass through the main gate, but poorly armed.

You can scout out a sewer entrance, the sewers may seem like a smart option at first but will lead you past lots of enemies (monsters), lead you out into a high-security area once you leave them, and you'll be smelling awful, cancelling out any bluffs you may have while being found on the premises.

An acrobatic character may find that you can climb to the roof of a neighbouring building, jump to the roof of a small building next to the satrap's romantic garden, climb up to the balcony, and from there into his personal chambers, requiring the bypassing of a single trap.

 

A player can choose to scout out the palace on a day that the satrap is away and security is light, this will allow a player to know the lay-out, but should the player be discovered, will lead to extra security measures on the day of the grand party.

 

Each player can take any entrance, with or without preparation. Prepared players can make educated decisions on what entrance is best, what egress is best, and base it on their playstyle.

Sure you can fight your way through the main door, but it will be more difficult than fighting through the sewers, and the Satrap may flee.

You can sneak through the kitchen, but it will be harder than sneaking through the stables. And you could bluff your way past the guards at the front gate, but most likely they'll ask for your invitation papers anyway.

 

It becomes about skill, but also about your playstyle.

And all players would have the same map, the same level to deal with, yet everyone's experience will be different.

 

So now I know I need

A main entrance (with guards)

A secret entrance (with traps)

A stables area (lightly guarded; connected to the palace but not directly)

A Kitchen/deliveries area (lightly guarded; connected to the palace)

A sewers (populated by monsters, lots of fake exits leading you away from your goal, one or two which open to inconvenient locations)

A balcony and door directly leading into the Satrap's bedchamber.

 

I now have 6 points of ingress/egress, not all of them will be available or useful to everyone, but anyone can find them and attempt to use them. Based on how the player is playing, he or she will be allowed to make educated decisions on their approach. (allowing for a change of plans) I have different challenges set for different sections, which can be avoided (no bonus XP, thank god, so no incentive to do everything) No entrance will be the best option, all paths will create challenge.

-

I think this method creates organic maps rich in potential, moves away from what I like to call "Corridor Hell" maps and helps deliver the setting vs it being "just another dungeon".

 

And you can expand on this too, because once you have a map, you can elaborate on what would be possible there.

Just thinking of this scenario gives us several options for content.

We must cause a scandal discrediting the satrap, what better time than the day he holds a large feast.

We may want to loot his palace

or we must crash the party in order to save the Satrap from Assassination

Perhaps we must assassinate the Satrap, and do it at the day of the party to send a message.

The Satrap is holding someone prisoner, we must rescue the prisoner.

The Satrap wants to test security for a grand feast he is preparing, he's hiring you to test the defences.

A secret organisation you work for wants to get word to the satrap, you must deliver a message without being spotted.

The Satrap is holding a large feast, at the party is a man I wish to contact, I must talk with him about business.

 

And peripherally, in preparation, ahead of time:

We can find out about the satrap's jilted lover, or sabotage his current one. We can manipulate this knowledge and cause a scandal if we're invited into the party.

We can learn about the dungeon under the sewers (just off the top of my head)

We can scout the palace ahead of time, maybe pick up a quest from those found there

The Satrap has the best warhorses in the land, we could steal his finest.

We can forge invitation papers, forge an invitation list and plant it.

We can forge a delivery request.

We can pressure a local nobleman to get us in as his +1

The local crime syndicate can ask us to look for a specific piece of loot thought to be on the premises.

 

-

TL;DR: If a place has a purpose the maps will draw themselves.

[Jarmo]

 

How to design such a map? I don't know. I suspect I would build it not as a level designer but as a narrative designer (stick to your strengths ) IE: I'd build the Satrapal palace first, and then worry about how to create a challenge.

 

 

This led me thinking about the multiple ways to develop a location.

 

Or that's actually one way. Design a location based on who live there and what they need and do.

I like locations like that the best, the ones that make logical sense. Location first, challenge and story after it.

 

One way is to design a challenge. So it's a palace ok, but the design starts from thinking about the encounters

and advancing the story. This is where you start placing barricades and stuff so players don't wander into things

in the wrong order. Some variation yes, but challenge first. Like, we need a big area where the party is ambushed,

hmm... lets make a throne room, should work for that. And then some place where they need to cross a bridge,

because there's this one encounter that'd really need a bridge.

 

And one more way, which I dislike the most (though it's great in other types of games), is to kind of compose the

encounter and location much the way movies are done.  So you start with something like:

dun-dun-DUN-dun-da-duuuunn- ----- ---- du- DUN-DUN-DOOOOMMM so the whole thing has a kind of a rythm,

small encounter - small encounter- big encounter- small - tiny - long encounter-  then catch the breath until the finale.

That's where you take away about all the choice but give a cinematic experience that just a rollercoaster.

 

I think the last one is where ME2, ME3 and DA2 were going or aiming for.

[rjshae]

A big part of the problem with linearity is that it allows you to fine tune your actions following a reload. However, you can turn even a linear dungeon into a non-linear design by creating dungeons that (a) change over time and (b) respond to your actions. If the dungeon population and their activities vary based upon the time you move through it then it throws off your expectations. Likewise, if you leave a dungeon part way through, then enemy deployments should respond in kind. Those parts of the dungeon you cleared should not be guaranteed to be empty upon your return, and the intelligent enemy should respond to knowledge of your previous tactics.

[IndiraLightfoot]

Level design should accommodate complex quests with various approaches and resolutions, but I reckon a few them still can have a linear layout. 

As Nonek's lovely post illustrates - it's what's in the dungeon that counts - the atmosphere, the story, the depth, the variety, the sociocultural context.

 

A level that's cleverly designed should make any backtracking part of the adventure. Although it was nice in Skyrim to always have a backdoor at the end, I soon felt cheap and sordid doing so - I might as well have had a teleportation orb and leave immediately (almost like the map worked). I know, it sounds old school, but I think I like a lot of trekking and backtracking - but under one very important condition - that such movement entails new encounters, the branching off of unexpected adventures and quests, and that those makes sense in the fantasy setting created. So not just a few monster spawned in, but something that is of the same quality as I went out adventuring, as I went in those dungeons.

[Amberion]

this is what i call a non linear level. you start at the green room and you have to reach the red exit. to get there, you have to find keys that open the matching color doors. black dots indicate a hidden path that can get you to the destination faster, by passing the deadliest hidden room (i just noticed that i forgot to add dots from the scull room to the last blue room), or will allow you to find a clue that will let you use an alternative exit that will lead to a different place than the normal exit

No offense, I appreciate that you went to a lot of effort to plot that out, but that's a pretty linear level. The only choices you make in that level are whether to brave the deadly room, whether to look for the secret clue, and which exit to go through. The first level of irenicus' dungeon is less linear than that.

Edited June 26, 2013 by Amberion

[Sacred_Path]

Eh, Irenicus' dungeon was a pretty fu*ing bad opening area for reasons that had not much to do with linearity or lack thereof. However, I'm going to comment on one of your points:

 

2. Linearity is bad. I honestly cannot come up with a good reason to have a linear game. Maybe there are good practical applications for it, but I can only see the negatives at the moment. Why is linearity bad? Because everyone experiences the same thing when they play the game. A linear game will play the same way for everyone; there is no emergent behavior. It's like writing a computer program that can only do hard-coded things. It's worthless. What's the point of making a video game if players cannot make choices or deviate from what is expected? Linear games often feel too restricting and are very tiresome in how there's very little room to /play/ with the mechanics of the game. Linear level design is determinism of the worst kind.

A linear game can still very effectively make use of what for me is the essence of a CRPG; make good use of your characters' abilities to overcome obstacles. Players' behavior can differ drastically if they have capable archers in the party (that can draw enemies and disrupt casters), sneaky offensive casters (that can charm foes or bomb a room), assassins (that can take out bosses), or a melee heavy party that can take up positions in the center of a room. Better yet, the linear layout makes balancing all these aspects for the devs much easier, limiting the number of potential loopholes. With just a little room to maneuever, you can even sneak - and thereby pick and choose - with what denizens of the dungeon you want to make contact. Saying that all players have the same experience makes me wonder if you have played any linear RPGs at all.

 

BTW, I'm guessing your definition of a "linear" dungeon still allows choosing between going left or right... because I can't think of any CRPG right now where there are no optional rooms/ hallways in dungeons.

 

IMO the best kind of design is having dungeons with several levels which are each slightly non-linear, but must be traversed in a linear succession.

[Sensuki]

Irenicus Dungeon was good...

 

The first time you played the game

 

Because it gave you that sense of OMFG YES I AM OUT

 

but yeah on replays ... I think a lot of people would just give up before they got outside.

[anubite]

Hm, I think people are fixated too much on my first example. I do concede in the original post that it's a fairly linear dungeon - which is probably why I dread replaying BG2 as well. The point is, that it's a tutorial level and yet still exhibits a good kind of 'circularness'. I probably should have used a different dungeon for my first picture, as it's misleading to call the Irenicus Lair dungeon 'non-linear' - it has a few non-linear elements, but there's definitely one optimal solution.

 

That said, at this point, I haven't played an RPG with level design even as good as that particular dungeon... probably since BG2. I'm struggling to think of an experience that is as good as it after I played it, first time or not. It's kind of sad that engaging level design is such a lost art. I'm still trying to figure out what makes it tick, though I think I've nailed the gist of it in my first post? Whether you agree with the particular examples or not, aside.

Edited June 28, 2013 by anubite

[Amberion]

No, you're absolutely right. It's been a consistent thing in recent years, particularly with AAA games where the plot is extremely linear, particularly in the later stages of the game. Part of this is due to cost considerations and a desire to tell a compelling story within a fully cinematic experience. The recent Tomb Raider reboot is an excellent example; it's a linear game, but has non-plot important nonlinear elements(mostly revolving around gathering items and upgrading weapons). Tomb Raider is, IMO this year's best storytelling experience in a AAA game, but it is definitely linear.

 

I remember a mission you were given in BG2, to raid a slaver compound and rescue some slaves for a barbarian. There were two entrances; the front door, and a secret sewer exit. The sewer exit has an ultimate exit inside the copper coronet tavern, which is incredibly easy to spot from the outside. When you get the quest the first time, the natural thing to do is to fight your way in through the front door; eventually you find the back entrance after killing all the guards and the captain. On subsequent playthroughs, it's more likely you will take the sewer entrance and carve through the slaver compound in reverse. I really liked this design, for while the level itself is linear, the hazards and encounters can be different each way.

[Jarmo]

Good non-linear design in many of Fallout 3 and NV locations.

As much so you hardly think of them as levels.

 

There's this prison compound, you have gate entrance,

but after that its several buildings you can enter in any order and the buildings themselves have many routes to take.

Some of the old vaults are pretty linear, but many are not.

 

There's a bunch of city-based modules for Neverwinter Nights, those are mostly non-linear, you just have some place

you need to go but many ways to get there. 

 

It's pretty much immediately when you move from a city or castle or vault or such to a more "dungeonish" location,

that developer goes from I'm developing a location mindset to I'm creating a level mindset, which is a bit of a shame.

[motorizer]

What they need to avoid doing is making all the l;evels follow the same basic structure

 

There should be some linear and some non linear, you shouldn't go in knowing what to expect because you've already done some dungeons.

 

Pretty much every dungeon in skyrim comes round full circle to give you an easy exit as soon as you've found what you went in for, to the extent that as soon as you have it you're looking for the quick way out that you know is there, after a while it felt lazy

 

I agree that you shouldn't need to backtrack the way you came (unless it's changed considerably by some event) but there should be a bit of variety in how you get out, where the objective is, and how linear it is.