Every game, when you think about it, is a kind of puzzle. You’re given a problem, and your goal is to solve it. Traditionally, we think of puzzle solving as a very obvious sort of thing, like an old-school adventure game or something like Portal, but puzzles are everywhere. In a racing game, you’ve got to solve how to get ahead of everyone in the race and maintain your lead. In a shooter, you’ve got to solve the problem of all the enemies trying to prevent you from obtaining what you want.


Of course, puzzles can be good or bad. Battlefield 3’s puzzles aren’t particularly great: climb onto a very exposed area and shoot lots of dudes who are running at you. Easy-peasy. It’s not the only game like this, of course: Max Payne 3 features something like five sections where you sit tight somewhere and shoot at guys who run at you. Games like FEAR, STALKER, or Halo, in contrast, make use of the space you move within, giving you mechanics (such as bullet time), level design (multiple approaches; freedom of movement), or tools (enemy AI that can be manipulated and tricked, weapons, cover, etc), creating much more flexible, interesting puzzles to be solved.

The more tools you have to solve a puzzle, the more fun it can become, though it doesn’t mean that the puzzle needs to become some wholly flexible, open experience: Half-Life’s Blast Pit, a puzzle about killing a giant tentacle monster, features a series of interlocking, linear puzzles that takes players through a space that feels big and natural, leaving the combat puzzles themselves up to the player.


I realize this all seems kinda weird and paradoxical. Most people think of puzzles as only having one solution, but that’s inside-the-box thinking. All that really matters when solving video game puzzles is having to figure out how to obtain a goal in a way that enables progression. Playing by reflex isn’t the same thing; at that point, you’re merely reacting—puzzle solving is a proactive behavior.

The more tools you, the player, have at your disposal, the more ways you can solve a puzzle in interesting ways, and the more satisfying the experience can be. Remember how I talked about how self-motivated play is awesome yesterday? Well, this kind of choice-driven puzzle solving is absolutely rooted in that idea.

So, let’s say you’re a prospective game designer. Let’s say you want to create a game with a big emphasis on this kind of puzzle solving. What do you want to do?


You want to make a first-person stealth game.


A few weeks ago, a writer on some critically acclaimed legal dramas came to give a talk at our school. I was racking my brains over what to write about, because I knew I didn’t want to give my teacher the hundred-and-fiftieth recap of everything she said. Fortunately, I managed to remember a couple interesting things she’d said, almost as asides, and put them together to come to one conclusion: people like formulaic television because the drama of the situation is easily understood.



In a video game context, that means giving the player a very clear goal, and stealth games are great at this, because the player is often something like a thief or hitman. So you can tell the player “get this thing” or “kill this guy,” and the player knows that they have to obtain it. At the same time, the player has enemies of varying types, like security cameras, guards, tripwires, and other stuff.

But stealth games also do one other thing: because the player needs to be in the mindset that they have to be sneaky, the developer is encouraged to make the player feel vulnerable. There’s this big, huge essay I’m going to be writing in the future on the nature of vulnerability and how it increases emotional engagement in a game, but for now, I’ll simply say this: vulnerability is really hard to create in games. The most obvious solutions—making it easy to die or making the player alone—aren’t always enough to create that sense. Halo’s one of the few games that has pulled this off, not just by switching the major enemy type half way through the game, pulling players out of their comfort zone, but by creating enemies who were exactly as powerful as the player, not to mention clever, and occasionally outnumbering the player. Elites are, and forever will be, some of the best enemies in video game history. In a stealth game, “I have to hide” helps create that sense of vulnerability, thus enhancing emotional engagement.

In other words, what all of this means is that the player WANTS something, but SOMETHING ELSE is keeping the player getting that thing, which means that you’ve established an inherently dramatic situation that can lead to an awesome, thrilling outcome.


But also… stealth means players have to slow down a bit. Some people still play Halo like it’s a reaction-driven experience, rather than a proactive one, because of the popular misconception that shooting games are about pointing at things and clicking on them until they’re deleted, which is technically true, but not that impressive, kinda like suggesting that a paper airplane is as good as gliders can be. Stealth makes sure everyone understands that there’s a problem and they have to solve it. Plus, to make it good, you’ve got to have more than just a broad variety of great guns, so it becomes immediately apparent to players that there are multiple solutions, thus encouraging them to try different things.

Of course, not all games are like this: Deus Ex: Human Revolution basically says “stealth, no-kills, is the best way to play the game,” so its XP is rewarded accordingly, but there’s no real reason for this, other than some people insisting that stealth was the best way to play the original Deus Ex. Dishonored isn’t quite balanced right, making players feel like they have to stop and start over every time they’re detected—it’s very hard to simply go with the flow and kill people if you’re seen, because it becomes so easy to do that killing becomes the optimal way to do things, and stealth starts to lose its purpose.

So… stealth can be hard, don’t get me wrong, but the potential there, for natural drama, instinctive player understanding of the freedom they have to try things out, and stuff like that? Well, it’s pretty great.



Why first-person, though?

A 2D game provides two dimensions of movement: vertical and horizontal. A third-person 3D game, of course, adds the z-axis to the mix, this change in movement immediately providing another tool, but at the same time, third-person games give players a sense of power, allowing them to do things like looking around corners without revealing themselves, reducing a sense of vulnerability by giving them more information than they need to have.

First-person games have a few benefits, most obviously the way in which they obfuscate information from the player, limiting their vision. It’s not the only advantage, of course; first-person games have a way to put players into the heads of their characters. Not everyone feels this way, but that’s often because they’re playing these games on console, where they’re greatly distanced from the screen, disconnecting them from the experience. For an optimal first-person experience, using a field-of-view-filling monitor or an Oculus Rift, or something similar, the character’s actions become the player’s actions, connecting the player to the experience on a deeper level.


So now we’ve got this game that has the potential for this really deep, dramatic, emotional connection with the experience, which is awesome, but it also provides some really basic tools (six axes of movement) and obfuscation (less omniscient viewpoint).

What else can we do?

Right off the bat, there’s sound. Nobody pays attention to sound, but every game I think is bad (but has great reviews on Metacritic) has some really awesome sound design. Anyone in film will tell you that the final mix is one of the most important aspects of a film; they’ll remind you that films are an audiovisual medium, not a visual-only medium. Sound is amazingly crucial to the video gaming experience, and most games use it in a strictly utilitarian manner.



I’ve noticed, with some worry, that Eidos Montreal talks about Thief as if it were a game about light and dark, and that’s certainly the most obvious element, but if Thief and Thief II cornered the market on a kind of stealth, it was aurally-driven stealth. It’s important to note here that I’m not talking about how crouching makes you quieter so enemies can’t hear you—every stealth game has that, and if you don’t have that, you’re probably off to a bad start with the stealth experience.

Thief played with materials: walking on a carpet doesn’t sound the same as walking on a marble floor, which alters the ways in which the player can be detected. In some situations, the players could walk; in others, they could run and not be heard, allowing them to dash up behind enemies without being detected. Simply through the use of materials and the effect they have on sound, Thief created a significantly less binary, more varied experience.

Still, that’s player-dependent. Thief’s enemies make sound too: some snore, some chat, others whistle. The surfaces they walk on affect whether you can hear them or not—be cautious entering a room with carpet floors, but be thankful for rooms with metal grates. In other words, Thief helps you determine when enemies are nearby, but it doesn’t tell you exactly where they are, or where they are looking.


And the great part about this is that, if enemies can hear you, and they can create sound, then it logically follows that they can hear other enemies, which means that you can manipulate enemies into doing things you want them to do.

Down In The Bonehoard is an amazing level. Thief is an exhausting game to play, and the Bonehoard is no exception. There’s a certain enemy type in the level that isn’t too bright, but makes plenty of noise. Oh, and they’re very hard to kill—you need holy water to do the job. Me, I didn’t have enough arrows to kill them all, not with a good twenty or so of these guys wandering around the level. Then I realized that the enemies alerted other enemies, who would then come running. I formulated a plan.

I found a room—my killbox. A few enemies slept peacefully below me, until I shot one with a holy water arrow, and he exploded. As it turns out, the arrows feature a degree of splash damage, which meant that I also damaged the other two, who immediately moaned, and began searching for me. Within a few minutes, the room had filled with enemies, alerted by the cries of their companions. I went to work, dispatching them with my arrows. The splash damage meant I was able to kill all but one of them, who followed me forlornly around the level as best he could, bless him (well, that’d probably kill him, but whatever), trying pitifully to murder me.



I didn’t have enough resources to obtain what I needed, so I used the game’s mechanics, specifically sound and AI manipulation, to obtain the result I wanted.

It’s not the only tool, of course. In his talk Three States and a Plan: The AI of FEAR, former Monlith AI guru Jeff Orkin talks about light switches, and how it took a lot of work to figure out how to get enemies to turn on the lights in dark rooms, and how Goal-Oriented Action Planning, the AI system they developed, could have made this a very easy problem to solve. There’s a whole big thing about AI, how it works, how complex AI makes for better games, and a bunch of other stuff I could talk about here, but the truth is, I’m nowhere near a pro on this subject, and can only tell you what I know from playing games with great video game AI: every game should have great AI. Like sound, it’s literally one of the most important elements of a video game, and also one of the most overlooked. It’s one of the reasons games like Halo and Half-Life still hold up.

Instead, I want to talk about light switches.

When I mentioned my Bonehoard experience, I talked about getting the AI to do things. That’s another tool in your arsenal. Imagine switching off a light in a room, which causes an enemy to deviate from his path, entering the room, allowing you either to bypass the guy or lay in wait to ambush him. Gunpoint, for instance, can let you rewire switches so that, when you turn off a light, and a guy goes to turn it back on, you cause him to pass through a sensor that automatically opens a door you needed opened. Payday 2’s Framing Frame level lets you activate blow-driers, which means you can trick guards into entering the bathroom, where you can get them out of the way.


Because AI’s so overlooked (in part because it’s a very CPU-intensive task, and there are other things, like graphics, that developers seem more keen to devote the limited power of consoles to), this area of game design is often overlooked. Most people seem pretty content to have enemies who run at you, occasionally take cover, and fire. Most enemies vary primarily in terms of how much armor they have, or feature fairly limited enemy types. Developers such as Monolith have explored with giving enemies lots of behaviors and allowing them to interact with world objects, creating cover and channeling the player’s flow through a level. Other developers, like Bungie and Epic, allow the player to fool players, or use weapons to manipulate them—toss a grenade here and the enemy’s going to jump out of cover there. Fire at an enemy from here, and then work your way around to flank while they fire at where you had been. And there is so much more that could be done.

But for now, let’s talk about controlling a level.


At its heart, a first-person shooter is about area control. Most people think it’s about shooting, but that’s a bit like saying a novel’s about words; sure, you can’t really have the former without the latter, but that doesn’t mean it’s what the former is really about. A novel’s about a story. A first-person shooter is about controlling a 3D space. A first-person stealth game is a bit like that.

AI manipulation, of course, is a great way to do this. Unique tools—in a stealth game, this doesn’t have to be weapons, though it can—can control the way you treat the game space. In Halo, which I reference a lot because it’s more or less the textbook-perfect first-person shooter, enemies armed with various types of weapons are going to affect the way you interact within the space, as will the weapon load-out you bring to bear. Level design can change things up too. All this ultimately has the power to result in an incredibly deep, varied experience. In a lesser shooter, of course, it’s often just about “wait here, point a gun at those guys; shoot them until they die.”


Of course, that’s not the only way to do things.

Orcs Must Die 2 is an awesome game. Everyone should play it. Robot Entertainment has created a super fun, hilarious, and a great tower-defense/action game hybrid. Tower defense games are about area control in a different way: create barriers and towers/traps/weapons/whatever, use the level design to your advantage, and control the flow of the enemies the way you want through the space. Of course there can be optimal ways to do things, but it’s always really fun to just kinda determine what space you want to control, how you want to control it, and get ready to rock.

Nobody’s really done this in a stealth game, that I can think of, but there’s no idea why the basic concept couldn’t work: board up a window here, shut off a light there, get the AI to engage in the stealth patterns you want… pretty soon, you end up controlling the space, giving you the ability to move around as you please.


Actually, I can think of one game that’s kinda done this: Payday 2. The guards won’t come out doors you haven’t opened. Once the doors are opened, they will. Now, granted, being able to close doors and control where the guards move would be a tremendous boon to the experience, but for now, I’m simply happy to look at what doors are open, keep an eye on them, grab the guards I can, and then open doors to ‘attract’ guards where I need them to go so I can take care of them.

It’s really clever, and it’d be cool if this idea could be explored more deeply in other stealth games.


Another game that toys with this is Hitman: Sniper Challenge. You can shoot an elevator panel, and the sound of the bullet strike—or the open elevator itself—will cause a guard to check things out. When he gets close, shoot him, he’ll go falling down the elevator shaft. You’ve opened an area of the level, used it to affect a guard’s behavior, and used that to take him out of play.

Then you’ve got social stealth, and I don’t mean in a kind of Facebook context. Deus Ex: Human Revolution toys with the idea of conversation battles, the idea that every conversation is a sort of battle or puzzle that the player has to ‘beat.’ Instead of simply being dry, boring “please exposit information at me,” conversations involve active player participation—to get the desired outcome, you have to pay attention to what’s going on. It’s been a while since I’ve played, but I believe the player can use conversation to trick their way into going to places they’re not allowed to go. In other words, social engagement with game characters allows you to do whatever it is you want to do without being noticed.


Hitman (not Sniper Challenge, which totally needs to become its own game, but the proper games in the series) lets you wear disguises, another kind of social stealth. In Absolution, it’s a bit too mechanical and binary, and would really benefit from a more simulation focus, but the basic idea across the series seems to work: wear a disguise, such as a janitor’s outfit, and most people won’t question your presence. It’s a great way to interact with the game’s world. Payday: The Heist features a similar idea in the PC-exclusive Mercy Hospital mission: at one point, you get to wear doctor disguises, allowing you unrestricted access to a ward you’d have to drill your way into otherwise.

There are a ton of ways you can mess around with a stealth level, but at the end, it’s all a sort of dramatic puzzle, trying to figure out how to get what you want, using lots of tools, from the enemy AI, to sound, to social exploitation, to get what you want. Seriously. Payday 2 features ECM jammers, which allow players to bypass cameras, for instance. Imagine players taking on Payday: The Heist’s Diamond Heist map, but instead of alerting all the other guards when they get noticed, the players could convince the guards to cuff themselves—Payday 2 only allows masterminds to cuff guards, and even then, it’s just one guard (regardless of how many masterminds are on the crew, which seems weird to me). Players could use hacking tools to gain access to camera feeds, scouting maps. The Payday games allow players to highlight enemy guards so their crewmates can see them (I mention the Payday games a lot, don’t I?)

So, yeah. How would you change stealth? What kind of mechanics would you implement in a stealth game that might not have been tried before? Would someone please talk about how Amnesia handles stealth? How could you take stuff like, say, AI manipulation, and really go all-out with it?


That’s all for now. Seeya next time!

You can find me, as always, on my Tumblr and Twitter. As usual, if there’s any subject you’d like me to cover, let me know! I’m always itching to write essays during my downtime at work, and, due to the nature of the job, there’s a lot of downtime (part of the reason I’m looking for a better job). Seeya next time!


NEXT WEEK, ON DOC WRITES ABOUT STUFF: “Oh no, Skyrim is so bad,” “I disagree! New Vegas is the worst!” Can these two opposing viewpoints make any sense? What’s really going on? Is it possible that both of them are great, or are they both horrible? Must things always be so binary? Where did these games even come from? FIND OUT, NEXT TIME!