Intermittent Mechanism

Part 5 is up!

Parts 1 & 2 were adaptations of existing material—namely, the first chapter of my dissertation, and this lesson plan. Parts 3 & 4 consisted mostly of newly-generated material. Part 5 returns to being an adaptation of existing writing—this time around, this blog post.

I had to scale back my ambitions for this particular video. Originally, it was going to feature a tour through some .OBJ outputs of the coastline maps, following in the footsteps of Robert Yang’s visualizations. Everything was going smoothly for awhile: I successfully extracted all the necessary textures with GCFScape, successfully opened the maps in Crafty, and got myself an education license for Maya. But try as I might, I just couldn’t quite pull off the trick that Yang did, and get the textures to affix to the .OBJ files. (He kind of glosses over that crucial step the blog post.) All I could produce were textureless grey blobs of level geometry.

So I fell back on a tried-and-true method of compositing a bunch of noclip screenshots in Photoshop. In addition to not having that cool 3D model look, it was also an enormous time sink, though, and slowed me down a lot. C’est la vie, I suppose.

Script below the jump.

Picking up again, for our coastline road trip—and in this video, I’m going to be talking about pacing.

5.1 Expect Delays

“A crooked road, a road in which the foot feels acutely the stones beneath it, a road that turns back on itself—this is the road of art.”

That’s a line that appears early on in Theory of Prose, a book by the Russian Formalist literary theorist Victor Shklovsky, and it’s basically the thesis of the entire book. Again and again, Shklovsky makes the point that if writers just took the most direct route from the beginning of their story to the end, then there would be no story whatsoever. So, instead, they take circuitous paths. Sometimes this just means manipulating language, adding repetition and redundancy at the basic level of word usage. Other times it means literally forcing their protagonist take inefficient circuitous paths.

The end result is what Shklovsky calls “stepped construction.” Here’s a graph of what a theoretical story would look like if the protagonist coasted easily to their goal. There’s not really anything here. So a writer’s main job is to add things that impede progress. Often, this will be some sort of sub-task that the hero needs to complete before they can reach their goal. So the hero moves forward toward their goal, is sidetracked by an obstacle until they complete a task that resolves it, moves forward a bit more, and so on.

These sorts of impediments serve as a delaying tactic. They exist to decelerate the pace of the story, to draw it out, to make it crooked.

In the early history of game studies, there were several scholars who read literary theorists like Shklovsky quite closely. And they did so because they wanted to argue that games and stories had irreducibly different structures. They didn’t think that games shouldn’t tell stories, and they didn’t think that game storytelling wasn’t a valid topic for analysis. This academic debate raged on for a few years, which all seems a bit silly in retrospect.

If anything, game stories are often perfect examples of this sort of step-by-step construction, with side-obstacles nested into the primary drive forward. Take even just a cursory look at a game like Dead Space, and you’ll find that each of its chapters perfectly maps on to a new obstacle—some new mechanical failure that protagonist Isaac Clark needs to go fix before he can get on with his main goal of getting off of the Ishamura alive. Pretty much every task the player undertakes in a game like Dead Space is getting past some new impediment introduced as a delaying tactic.

But if you lean on these sorts of transparent delaying tactics for too long, it begins to get tiresome. Many single-player games struggle when it comes to balancing the pacing of fun player activities with the pacing a satisfying story-based experience. Even if your gunplay is really fun, there’s only so many times you can escalate things, throwing even-bigger waves of dudes to shoot at the player, before you end up facing diminishing returns. And I don’t think I’m alone in thinking that game developers struggle with game pacing. Achievement statistics show that even when it comes to heavily story-focused single player games, less than 50% of players will actually reach a game’s ending. And often, the figure is closer to 20 percent, or under. So I can’t be alone in considering the back half of far too many games to be a slog.

5.2 Introducing stream pools

Half-Life 2 isn’t perfectly paced. I mean, it’s no Ghost Trick: Phantom Detective, which is perhaps the best-paced game I’ve ever played. (Seriously, go play Ghost Trick: Phantom Detective. It’s an utter delight.) But this road trip section is good, and uses some tricks to vary up the pacing beyond the usual impediment-based delaying tactics.

And this makes sense, I guess, when you think about it. If you look at the history of the road movie, forward motion toward the finish line tends to be broken up more by digressions and pit stops than clearly-defined obstacles. The history of that narrative form has given rise to much looser forms of pacing, based more around exploring local color than encountering roadblocks … literally, or figuratively.

The pacing of the Highway 17 sections of Half-Life 2 is largely defined by the environmental features of its level design. The various maps that make up this sequence all embrace what I like to call a “stream pool” design.

I’m borrowing this term “stream pool” from environmental science. (Yep, that’s right: despite its extreme simplicity, it is a bona fide technical term.) It describes exactly what it sounds like it describes: those deep, low-current pools that collect at certain points along the path of a meandering stream or river. They are often positioned at bends, after riffles, or partially separated from the main path of the stream by sand bars.

In applying this term to videogame level design, I’m using it to describe geographic nooks in otherwise linear games, where players are encouraged to wander from the otherwise-apparent path forward, and explore some tangential features of a game’s geography. In these coastline drive sections of the game, that will basically mean any time we get out of the dune buggy and explore roadside areas on foot. We’ll do that a lot over the course of these road trip levels.

Sometimes, we’re forced to get out of the buggy, due to some actual roadblock, or obstruction. There are moments along the drive where Valve has unavoidable combat encounters planned, where we’ll have to switch from driving to first-person shooting, and then back to driving after we throw a switch somewhere. But there are plenty of other areas that are entirely optional. We might just stop and see what sort of goodies a house has because we’re we’re low on health or suit energy. (Remember back in episode 3, the points I made about health packs as a motivation for player exploration.) It isn’t always immediately clear which areas are required stops and which are optional stops, so players are encouraged throughout these levels to take a slow and exploratory pace.

To give a sense of how this works in practice, I’m going to take a close look at two specific maps, d2_coast_05 and d2_coast_09, and then the bridge section, which takes place over several maps.

5.3 Unfolding some maps

So, first, a map called d2_coast_05. Like every map in Half-Life 2, it’s a linear experience—made especially obvious here by the fact that we’re following a road. But along the various bends and twists in the road, you’ll see these moments where players are encouraged to slow down.

This might be something as simple as some roller mines getting in our way, forcing us to get out of the buggy to remove them. And then we might find some derelict cars that need punted out of the way. Or a landslide that we can easily drive over, but will cause us to slow down a bit.

Up ahead on this same map, we see a more elaborate stream pool. Combine soldiers have set a trap for us, collapsing some debris onto the road. In actuality, this little impromptu road block doesn’t cover the entire road. We could easily speed past it, if we felt so inclined. But people are shooting at us, and we’re losing health, so we are indirectly encouraged to get out of the buggy here and take on these enemies on foot. That way, at least, if there are health items inside these buildings, we can grab them, so this encounter might actually be a net gain for us. And indeed, this pays off: there are goodies in these buildings, once we scour them.

Around the bend, another small encounter—some more derelict cars and another landslide, this time with guys shooting at us, as well.

This graveyard of ruined vehicles marks the beginning of an area that invites several approaches. We can, if we want, punt the cars and bars out of the way, and drive through. But, if we do that, we’ll eventually encounter a forcefield up ahead on the bridge, and we can’t go around it. Even if we get out of the buggy and punt it, there’s no way to slip through crack here.

So the best way to deal with this area is to actually use the makeshift crossbow, which we can find right up on this grassy knoll here. With it, we can snipe some of the soldiers before we get in close. And what we’ll eventually want to do, once we’re over here and have gotten rid of all the hostile soldiers, is remove the wheel jams that are parking this armored vehicle. It is powering the forcefield, and once we allow it to roll backward into the sea, that will open the way forward. And, of course, there’s loot in these buildings, so we can further explore and scavenge.

These little pauses don’t make the coastline of Half-Life 2 any less linear. We’re still barreling down a single road, to our single destination. But they give the player a decent amount of control over the pacing of the experience. We can drive forward, trying to avoid combat as much as possible, if we want to just get things over with. But we can also get out of the buggy at every opportunity, pursue every enemy on foot, take our time to visit each and every one of these little roadside attractions, search every building, to see what they have in store for us.

For another close look at an individual map, here’s d2_coast_09. This one has a pretty subdued start—first there’s just a crashed truck with some boxes piled around. This isn’t much of an obstacle, but it serves its purpose, which is just to slow us down. And then, once we’re already slowed down, we see this farmhouse on the horizon. And there’s several things about it that can catch our attention—for one, there’s smoke rising above it, and secondly, it actually has its own music cue.

This farmhouse is my favorite bit of the Half-Life 2 coastline section, because it is entirely optional. When we looked at d2_coast_05, we talked about ways that one might minimize roadblocks and combat encounters by not getting out of the car. But nothing in d2_coast_05, or any of the other coastline maps, goes as far as this farmhouse. It’s not a roadblock at all—it’s just purely an invitation to explore, nothing more.

From the perspective of Valve’s level designers, this distinction between forced combat roadblocks and optional areas to explore is pertinent and obvious. But it’s important to note that they bleed together more from the perspective of a player. The first time you take your road trip down Highway 17 as a player, you will rarely be entirely sure whether getting out of their vehicle to explore a given cluster of buildings will prove necessary or not. Roadblocks and ambushes are often well-hidden enough that they can take players by surprise, and so it quickly becomes learned behavior to get out of one’s vehicle whenever one sees a building around, and approach trepidatiously, looking for supplies, enemies, and a lever to defuse an as-yet-unseen trap. This subtly encourages a stop-and-smell-the-roses approach to these chapters of the game, modulating player behavior so as to modulate the pace of game events.

And here, we’re treated to some light environmental storytelling, and a fun jump scare.

Moving forward just a little bit, we find another combat encounter. We can hear gunfire as we approach this checkpoint, which gives us a heads-up that it is manned, and hostile. We might think we can run right through it, but as we approach this roadblock lifts up, stymying this attempt. Looks like it’s time to get out of the buggy and shoot again.

A fun little detail about this checkpoint is that it is actually possible to run through it. You just have to know what you’re doing. Just another little way that players can modulate the pace of this coastline road trip.

You can’t run through the next checkpoint. There’s a small puzzle here, where you need to collect three car batteries to raise a garage door that’s blocking the way. This takes some peace and quiet to do, so you’ll have to clear out the enemies here, first.

Overall, d2_coast_09 gives us a really good look into how stream pools, be they created through inescapable combat & puzzles, completely optional side-areas, or something in between, can work to break up the otherwise unrelenting flow forward of a linear game like Half-Life 2.

Just to add one final example of Valve’s craft, I wanted to tell a short personal story of my first time playing Half-Life 2, years and years ago. I had parked my buggy in this cluster of buildings, and started scavenging for some supplies. The sound of the headcrab zombie lead me to seek it out and kill it. And as I was chasing down the remainder of the poison headcrabs, just because you always want to make sure those never sneak up on you, I noticed this little door on the side of the cliff, standing out, lonely, with a light above it. And lo and behold, it opens, and there are supplies inside, drawing me in. And then it just … keeps going. Now I was in the inner workings of a bridge, and it just stretched on and on. The occasional headcrab encounter let me know that this is all planned, but otherwise, as I crept along these support struts, it felt as if I was breaking the game. But, like, in a good way. I felt clever, as if I was exploring parts of the map that maybe shouldn’t yet be open to me. And yet, details here and there reveal that this was in fact an intended path. Like the fact that I was shooting a bunch of guys.

And then I pulled this plug over here, anded push a button. And even though I didn’t realize it when I first played this sequence, I had just removed an obstacle from my way. But I wasn’t led over here by a need to remove an obstacle—I was led over here by my own curiosity, and by the exploratory, scrounging mindset that the stream pool design of this coastline section had foisted upon me. As I mentioned back in the “Push vs. Pull” episode, Valve allows players to happen on a solution to their puzzles without even fully realizing the problem. Another game might drown players in updates and new instructions if they break the designer’s intended sequence of events, even accidentally. But Valve just lets us stumble upon things in silence. They cultivate an exploratory mindset by including all of these stops by the side of the road, and they have the confidence that this exploratory mindset will lead us in the right direction when we get stuck along the path. And really, I had no idea what I was doing in this section. I just stumbled onto a boss fight. But it all worked out, in the end, because the level was well-designed enough to allow me to explore my way out of my own predicament.

And, to be clear, the obstacle that I didn’t notice is, in fact, well-communicated. If I had driven the buggy up onto the bridge in the first place, I would have encountered this forcefield, and realized that the way was blocked. And then I could have visually traced these wires, to see that they continue along the bridge. Putting two and two together, based on past experience, I would have known to go unplug them. But I didn’t have to go through that exact sequence of events. The game also supports pure, exploratory curiosity as a player approach. As I’ve said before, it is subtly redundant in the cues it uses to direct us, allowing different players to stumble upon things via different mechanisms.

5.4 Stream Pools in the Wake of Half-Life 2

The “linear path, but with nooks” design strategy of stream pools is a really popular one these days. Valve used it again in Half-Life 2: Episode Two, large portions of which are just an extension of the road-trip-like style of the Highway 17 sections. There, too, you see plenty of geographical nooks that players might initially stop at seemingly of their own volition—because they want to explore, or pick up some supplies, or something … but are then later revealed to be essential parts of the game’s storytelling, required stops on this journey.

I’ve emphasized a couple of times now that this sort of pacing, and exploratory play, really depends on players needing something, such as health items. So it’s no surprise, then, that linear games that have a strong survival aspect make use of this form frequently. Any time a linear game adds something like a hunger mechanic, or some sort of crafting, or anything else that might encourage players into a scavenging mindset, you see the stream pools design. It’s become kind of the official style of post-apocalyptic games, actually, showing up in things like the Metro series, The Last of Us, and then even more experimental games like 35MM, which has very little first-person shooting involved: it’s mostly just a meditation on wandering, survival, and scavenging. All of these games are strictly linear experiences, much like Half-Life 2. But they all cultivate cautious and sharply observant players, and in this way they’re able to slow down their pace, without throwing an endless stream of obstacles at players.

Alright, that’s all of this particular video. And we’re going to have a lot to cover in the next one, so one will be kind of a whirlwind tour. As always, thanks for watching!