Wednesday, May 27, 2009

Multilinear Level Design For Left 4 Dead

Wednesday, May 27, 2009 17

Two pages from my sketchbook, as I puzzle out nonlinear level layout possibilities for L4D.

Left 4 Dead is a well polished, mostly traditional FPS experience with a notable innovation: the Director.
Valve designers found that no matter how cleverly zombies were placed, players would quickly begin to anticipate them on repeat playthroughs. So they developed an ingenious system to handle the zombies.

The AI "Director" controls most of the zombie mobs placement and modulate their strength in lieu of traditional designer placement. This makes for wildly variable experiences in terms of enemy encounters, density, and where the Survivors make their defensive stands/when they decide to run.

But on the heels of the L4D SDK's release, my thoughts return to how this governing principle might be more completely extended through the experience--namely the level design.

My own experience of the excellent co-operative L4D experience was that I enjoyed it, but felt not much prompting to replay it because I had experienced the whole "ride".
Despite being a superior specimen thereof, it's still of the linear "funhouse" style FPS experience. I creep around the next corner of an expertly crafted environment, but my friends and I have no say as to where we're headed, which seems contrary to the spirit of a "surviving a zombie apocalypse" game.

Just as the Director unmoors zombie placement from direct designer control, what if the level layout were (to a smaller degree) unpredictable? What if alongside semi-random obstacles and dead-ends, the players had to fight but also actively navigate to freedom?

Arguably this is counter or more complex than Valve might deem desireable or saleable, especially for a game so well designed for perfect casual 30-45 minute play sessions. But for players who have invested enough time in the game and want a more challenging but also more varied experience, I think the idea has merit. Certainly worth further inquiry.

The initial plan is for the player start to be in the dead center of the map, with roughly four main paths radiating outward, like a pinwheel. In addition to these four main trunk lines, there would be interconnecting spokes. This gives players four possible paths with out-bound linear paths, but also lateral movement by way of the spokes.

The "pinwheel" design sketch. The paths are in a roughly square configuration to minimize lateral travel time between main trunk lines. In a finished layout, the actual player-paths would be more s-curved and kinked.

Instead of simply choosing one of four paths--in which multiple playthroughs would very likely expose one to be the easiest path and become preferential to players--there is a degree of chaos introduced by semi-random blockages or dead ends that would appear at various points along these paths and spokes.
This plan is far from a "sandbox" approach, but would ideally allow for complexity by way of multilinearity.

Further updates as I refine and implement a test case with the L4D SDK.

UPDATE: Multilinear layouts are not supported by Left 4 Dead.

Screenshot from test level. It shows that despite the test level having two safe rooms, the escape_route--how the Director and AI reckons player progress through the level--can only route to one of them, not both.

The bright spot here--other than providing a nice way for me to familiarize with L4D's AI/navigational architecture--is that I successfully predicted a design issue with a multilinear path.
Above I wrote that the trouble with multiple routes is that players would likely adapt to one preferential route quickly, without something like semi-randomizing dead-ends or blockages.

After writing the above update I downloaded the SDK and starting working with it. Then I remembered that I hadn't played the L4D commentary mode, and so I played that.
In it, one of the developers (Mike Booth maybe?) mentions trying an early non-linear city block level, and that the trouble was in fact that experienced players quickly found a preferred route (and that bickering about which route to take taxed teamwork).

Now, I don't think this would be as much of a problem if different routes actually provided vastly different endpoints--such as heading north means the suburbs, west means the slums, east meant the industrial outskirts and south meant the forest--but at this stage that's pure conjecture.
L4D is a particular sort of game and a very good one for the design elements chosen. I'll have to wait for another opportunity to experiment with the kind of level structure as outlined above; in the meantime I'm still quite piqued at building a L4D map/campaign.

Monday, May 11, 2009

Thoughts on Game Recycling

Monday, May 11, 2009 9
I spent about two years of my life working on an FPS called Darkest of Days.
While the release date is not yet known and the game's popular and critical reception remains unfigured, the experience taught me just what an arduous experience a production cycle can be.
As rewarding as the whole experience was personally, in retrospect something seemed extraordinarily wasteful about the whole enterprise, even at the fairly "indie" part of the spectrum. Is this how it's supposed to be?

There were so many assets created, tools furnished, lines of code completed and yet there are significant portions of the finished game we weren't happy with.

Or at least I wasn't happy with it; it wasn't the game I thought I ought to be making. However close we were to a better game is immaterial--games have to be finished, even if it felt as though we had only just laid in the foundations for a better experience.

It would be far less work to pick up where I left off and re-use assets and design concepts I am already familiar with to make a better game. Yet the way the industry works means I am assured to have to try my luck working on a completely new game, or possibly end up working on an expansion or sequel.

But the trouble with an expansion or sequel is that you are largely constrained by the scope and feel of the original game. If the original game wasn't quite what you had in mind, chances are slim to "complete" it in an expansion.

I reflected on just what a universal phenomenon this seemed to be--glimpse the game industry as journeymen laborers crafting near-exquisite works and then abandoning them. Rare is the chance any developer gets to go back and "make it right" (much of why I am keen to play the updated edition of Butcher Bay included with the new Riddick game Dark Athena); far more common is the endless parade of almost good games, a sea of reasonably good.
At best I play games that feel 90% or even 95% of what I want, yet that last 5% feels unfinished or somehow otherwise mismanaged.

It's as though a beautiful two story house is constructed, with only half the furniture is moved in and no paint on the walls, but the prospective homeowner stops there and elects to simply build another house instead. That's how modern production cycles seem to function.

And so a thought returns, a kernel of an idea.

Keel from Quake 3, done in Painter IX.

The thought occured years ago while I was enraptured by Quake 3, and the nagging thought returns now as I contemplate Unreal Tournament 3, two games removed significantly in time but not in concept or execution.
Quake 3 is a clean machine, a very "pure" game in ways that other id titles have not been, but at the time it was the first id game to eschew a proper single player portion.
(Which was difficult for a lot of FPS devotees at the time you'll remember. Most have returned to providing some token singleplayer game in addition to emphasized multiplayer, but this shift now seems paradigmatic for shooters as a whole.)

The thin scrim of fiction for Quake 3--a line or two in a vestigial manual to needlessly justify a multiplayer-only DM/arena game--was an obvious afterthought, yet it hinted at a far more interesting game made from the same constituent parts.
What if you did make a game about great warriors doomed forever to combat in strangely abstract arenas, a sort of self-commentary on the deathmatch genre?

Or what if you built a game out of an existing game, that was neither a sequel nor a mod? Something that re-used all the interesting world building, assets, characters etc, but went somewhere else with them? Why start all over again?

The thought recurs with Unreal Tournament 3 and its ilk.
My primary complaint with the game is much as it was with UT2k3 and 2k4--namely that game's visual fidelity far outstrip the interest that its traditional multiplayer mechanics give me.

An environment from Unreal Tournament 3.

It took me a good 45 minutes for a cursory glance around the levels that ship with UT3. It's a beautiful game with a surfeit of visual ideas and settings; deeply generous in a way that most games aren't, delivering and endless array of the kind of stunning visuals that Epic has become expert at creating. Yet the gameplay itself provokes no similar reaction.
This is not a complaint that UT3 was made into the game it was, I enjoy it for exactly what it is. (As a friend once pointed out, many gamers' criticisms can be likened to criticising basketball chiefly on the grounds that it is not football. This is not my intent here.)

But the thought persisted--if a game is a nearly-finished house, why couldn't someone move in and, with comparatively minor effort compared to what it took to raise the house in the first place, make it their home? Move some furniture, paint the walls?

Perhaps I'm not making any sense here, but it's a thought largely unarticulated until now so you'll forgive my lack of concision.

But there would seem to me a middle path, something between the land of fierce indie development (relegated to artistic sidescrollers) and the impressive world-building (but conservative linearity) of AAA big budget development.

Some might consider this the purview of mod development but that's not the shape of what I have in mind. There are too many amateur connotations to that word, and not enough fundamental change to the identity of the game in a "mod", despite the look remaining much the same.

Kana, from Unreal Tournament 3.

I'm thinking about what it would be to make a "new" game with mostly old parts. The unexplored connotations of what that great big spacestation level in UT3 could mean, the wide-open rolling plains with already excellent vehicle handling. Imbue all this great art and set dressing with some meaning, rather than completely interchangeable visual glosses. Build a new world from the old one.
And by "meaning" I mean gameplay mechanics, not Story.
(I think most misguided design impulses come from that tourist hag Story, but that's a discussion for another time.)

So picture a single player quasi-free roamer, or even a co-op shooter, re-using as much of all the phenomenal work already there as is applicable, but in a very different play style... the actual design sketch here I'm afraid will have to wait for another update.

Players say that what they really want is deep and engrossing gameplay, not just another empty graphical funhouse ride. But how would they really feel if they had "seen it all before" with the production art, but were really given a brand new game to play?
Could you build something more than just a "mod" that successfully alters the meaning and gameplay content with mostly the same assets?

The above was written and then left in a draft about three weeks ago; some of these comments seem oddly prescient to me in the wake of 3DRealms' demise. However misguided, there was a team dedicated to getting a game finished just so.
A cautionary tale?

Also, my friend Steve Gaynor recently wrote on the subject of what he calls "Single-A" game development, a far more coherent thought on a similar track. Please have a look.

Thursday, May 7, 2009

The Sap and the Heater: Q&A

Thursday, May 7, 2009 8
Further development of the "detective game" concept (posted here), prompted by message board discussion. Thanks to SA Forum User Dr. Pwn for the questions.

Q: How would you allow the player to execute any of these dialogue options, gambits, and counter-gambits in actual gameplay?

A: Good question, hard to say. One of those things you'd unravel ultimately through rigorous playtesting, even week one stuff.
I imagine it wouldn't be far from how some games incorporate interactions like human shields/hostage taking, since what choices are available to you are pretty contextual.

You'd need a lot of scaffolding to support unfamiliar gameplay. For instance, very specific audio cues--maybe slightly exaggerated "walking in hall/door handle/door opening" sounds so players have time to decide or react, for starters. It comes back to playtesting.
All the choices in the world dont mean anything if the players don't know they have them available/have chance to use them.

Beyond that, probably something contextual having to do with basic states. Talking to someone with a weapon in your hand would be different than if you're in the "being held up" state; obviously you're trying to reason with them in the latter, while the former case could be a couple of different ones.
Simple one button use/converse key, possibly with a small branching option.

Another important aspect to this sort of system is that it makes a verb out of the sort of thing that action games most traditionally ignore. Characters always react, it's just that in most games they bark and go straight to shooting, and that's it.
It's easy enough to make characters react to you pointing a gun from the hip in their direction, and even moreso when you aim at them.

Say if you've got your gun up and aimed at someone, "getting the drop" on them, he'll voluntarily disarm (that might be a vocal prompt, not unlike SWAT 4).
Then if you don't shoot him right there, further use of "converse" turns to interrogation. Though if you look away or fail to target him, he may re-arm and attempt to shoot you, like the Jim Zubiena clip.
What works so well for Hitman: Blood Money is the use of dropdown menus--not the most action-y choice but a game sketched out like the detective game I describe, with crime scene examination and the like, would generally be more paced like Hitman anyway, instead of an out-and-out shooter.

So I'd say that the actions described are: branching, contextual to time, space, and situation (armed/not armed/drawing gun/gun drawn, having a gun pointed at you/pointing the gun)....

I am somewhat loathe to use it as an example, since the series was for so long the posterboy of out of date control schemes, but how MGS4 allows Snake to crouch, or go prone, and then flip over and lay on his back and shoot all pretty easily.
The key is that Snake can't just go straight from standing to laying on his back (unless knocked back by an explosion); at various postures he has branching access to other postures, generally not all at the same time. Not the best example of what we're talking about here, but something similar.

Q: That's really cool, but how would you reconcile sequences in which the player-character can be incapacitated by a single bullet with real-time gameplay that usually allows the player to absorb ridiculous amounts of damage?
Would every encounter/firefight in the game be primarily a matter of "getting the drop" on someone?

A: I'd honestly go so far as to at least try to center the gameplay around very high lethality gunplay, yeah.
Something with a completely different flavor, to go with the atypical emphasis on having a gun and pointing it at someone being Serious Business...though again, this is only only ballpark stuff with something that would be so central to the gameplay without playtesting first.
[I follow the Valve model in that respect--design iteration is completely enmeshed with playtesting.]

I'm a huge fan of the whole milieu of detective fiction--Chandler, Hammett et al.--and there's a well established trope of the detective getting cold cocked and left somewhere compromising (which we're well familiar with from the bleedover into all sorts of pulp-inspired entertainment).
This is something that would translate well as a gameplay mechanic.
In this setting, gun battles should be dangerous, lethal, and therefore initiated rather infrequently. Usually when the deck is stacked in your favor, and even then it's not very detective-y to cut a double digit body count through the city while protecting the interests of your client.
So while you could get into a proper gunfight, for this exercise I'd like to get away from the all too familiar linear game problem of "you will play the story this particular way, or you will fail completely."

Mount and Blade, whatever you think of the game generally, proves something of a relevatory experience when played in the recommended no-save/loading mode. Falling in battle strips you of a piece of armor, a prized weapon or your warhorse, though you yourself are simply knocked unconscious/left for dead. The game continues in time.

Played this way, "failed" battles that would have otherwise devolved into tedious save/load festivals on a linear storyline in another game (cutscenes and all) went 1.) far faster without any repetition, and 2.) become a part of my player narrative, my story, rather than an impediment to progress through a canned plot.

So to me, getting rid of the traditional save/load mechanics but also ejecting forced failure and repetition of play segments allow you to own the play experience that much more thoroughly.
Leave the flawless runthrough to the savants and the tool-assisted players of the future, embrace your mistakes as part of the experience, not an error to be rewound until you "get it right."
I'm tired of playing along to a fake movie, with none of the impact of a movie and much of the baggage, and a character following someone else's conception of him, not mine.

So that's a whole lot of words, but I'd love to make a game with these mechanics.

Even say it's actually possible to actually die, straight up. If you're mortally wounded, you start the game over.
Or if you're shot, you need to limp to the phone or otherwise get help and end up in the hospital.
(In which either your whole case goes a whole lot colder and its harder to interview people or you just end up with a "convalescing in the hospital" ending--maybe you can keep this character and his stats for a new game+ type run through?)
For the most part though, you won't get into truly lethal gunfights often if you play smart, like the smart detective you ought to be.
Then when the goons "get the drop" on you and you are unsuccessful in weaseling or shooting your way out, you just get cold-cocked and a similar small penalty for being crappy:

-Bad: you wake to find they have snatched up some small clue/polluted crime scene (end game total for clues collected or how thoroughly you make your case or whatever, fill in the blanks)

your gun is taken (humiliating for your character and you--or are you a detective that prefers not to be armed?), or the thugs sap you and bring you in for a short meet and greet with the nemesis (could be triggerable to happen later in the game)

you wake to find either at the crime scene or somewhere else entirely, a freshly fired gun in your hand and a dead blonde on the floor and the police are soon to knock on the door. Oh shiiiiii--

Monday, May 4, 2009

Level Design Primer: The Sawtooth

Monday, May 4, 2009 4
Level Design Primer is an introduction to the high level concepts of first person shooter level layout and design.
As to credentials: for two years I built the majority of levels for the forthcoming single player FPS title Darkest of Days. Anything written here is backed by my own experience and observation from building and playtesting.

The sawtooth is such a universal staple of level design that once sensitized to, you will notice them everywhere. It is simple, effective, and (generally) foolproof.

The sawtooth stops the player from going backward. Simple as that. But as with other central design tenets of level design, simple can have far reaching connotations.
It takes innumerable shapes and forms: a broken staircase, a drop-down, a chute, a sheared cliff. Some are obvious, some are seamlessly integrated into the environment in a natural way, some are completely blind--like a ventilation duct that drops out from under you (groan).
Careful modulation of both the obviousness of sawteeth and their frequency are critical to the player's immersion. And the more "realistic" a game purports to be, the more carefully masked and generally less frequent they'll need to be to not be disruptive.

But why impede backward travel in a level? For single player games, there are a number of reasons: it makes life easier for the designer and the for the player in several ways.

The designer no longer has to worry about whatever the player is doing--getting lost, wondering where the party is, getting bored, and eventually quitting (which is the worst thing that can happen for a designer).
They can provide natural break points in the physical layout of the level to complement save points--which save in time, but not necessarily in space. Loading from a save point doesn't mean the player isn't still lost and might wander back to the start of the beginning of the level; a sawtooth means the designer knows for sure.
It's useful in tightly scripted "cinematic" action sequences where if the player otherwise may not know where to go or could veer off course too much and lose the immediacy of the engagement at hand.
Or, if the designer wishes to ambush or really rattle the player, certain kinds of sawtooth elements are useful to prevent the player from simply backing away from the encounter and cherry-picking the enemies from a safe distance.
The designer can also deactivate scripting or delete performance heavy entities or effects that they know for certain the player can no longer see.
This may just sound like its covering for lazy design, and while that does happen it also helps designers concentrate on crafting "the good stuff" rather than an increasing load of invisible-to-the-player housekeeping.

For the player it means a discrete point where, so long as he sees the sawtooth in advance, he knows he must be ready to progress.
For compulsive item and ammunition collecting players, approaching a clear sawtooth is a kindness; certain players hate not knowing if they'll be able to back and scrap up every last round of ammunition before moving forward.
And as previously mentioned, it ratchets physical progression through the level itself, effectively gating the play space into manageable chunks. This is helpful for designers, but aids navigation as well. In traditional "funhouse" like linear layouts, nobody wants to spend time wandering around the dead, empty portions of what they've passed through.

They're not just for single player layouts, either. For multiplayer levels whose flow tends to be circular and concentric, like a heart pumping its warring players into collision, sawtooth elements can still be useful in order to restrict two-way flow into more interesting one-way channels, say around a power-up.

To find examples of a sawtooth element, look no further than the best known FPS franchises with a linear single player or multiplayer component. A short list of recent titles would include Crysis, Call of Duty 4, Bioshock, Killzone 2, Left 4 Dead... you can find examples everywhere, once you look.

The critical view to sawteeth elements and their universal employment would be that they're a crutch that continues to prop up sloppy design and overly linear world design. So long as the kinds of things that make sawteeth such an obvious choice to use are present in level design, we'll continue to get fairly rote, linear "funhouse" type experiences.
But something as monumental as challenging concepts of linearity is outside the purview of this update--more on this sort of thing later.
gausswerks: design reboot. Design by Pocket