Thursday, October 11, 2018

Kick the cover box

A soldier hiding behind a gray box in a futuristic lab, from Deus Ex Human Revolution (2011)
The room pictured above from Deus Ex: Human Revolution is, I argue, bad level design.

The playable area consists of an open flat floor with lab counters, yet all the counters are the same height (they have to be, so the player can recognize them as "those boxes I can hide behind") and each box offers basically the same affordance to the player. (Hide behind it! Look over it! Shoot the NPC that's programmed to pop his head out every 7 seconds!)

Any given object becomes bad design when it is numerous, redundant, and lacks context to the rest of the game. If you automatically repeat any type of shape throughout your game world, as a catch-all solution to fill a space, then that object is basically functioning like the dreaded video game crate. Whether it's a pallet of barrels, or a stack of bricks, or a concrete road barrier, it all boils down to a "cover box"...

Level designers often place these objects in the same faux-haphazard way, like tasteful glossy interior design magazines forgotten on a coffee table. But they're mostly responding to the game design they've been given, especially in a AAA system where combat systems feel like immutable facts. Water is wet, crunch must happen, and shooters need cover boxes. It's going to happen, live with it.

So whose fault is it, really? Well, I blame Steven Spielberg.

(content warning for video above: lots of gunfire, bombs, blood, guts, and dismemberment)

In 1998, Saving Private Ryan (director: Steven Spielberg) ushered in a new wave of American interest in world wars between 1938 and 1946. That film's 20 minute Omaha Beach D-Day assault scene rose to be the holy grail of military shooters, setting the bar for the huge scale, drama, and grittiness that we now associate with depicting war in popular media. Spielberg totally recognized how lucrative this was, going on to help direct the original Medal of Honor game (1999), the TV miniseries Band of Brothers (2001), and Medal of Honor: Allied Assault (2002) which very literally follows Saving Private Ryan's Omaha Beach sequence, from the stark white-on-black introductory text to the themed map objectives.

Several World War II game franchises followed shortly afterward to compete with Medal of Honor. Famous shooters like Battlefield 1942 (2002) and the original Call of Duty (2003) have basically erased Medal of Honor from recent memory. But both Battlefield and Call of Duty clearly aspired toward the scale, drama, and gritty stoic masculinity of Saving Private Ryan / Band of Brothers.

Cover transforms these games into military shooters, enabling slower shooting dynamics that weren't so clear in fast-paced arcade shooters like Doom or Quake. Circle-strafing is not realistic, but crouching behind sandbags is "authentic" and "real"! Are you pinned-down, are you out-flanked, are you getting ready to charge? Both Battlefield and Call of Duty expanded this player vocabulary with a prone posture to crawl flat on the ground, miming realistic infantry tactics.

If Doom and Quake are about shooting, then Battlefield and Call of Duty are about getting shot at.

Kill.Switch (2003) is known as the first modern cover shooter and inspired Gears of War (2006).
With these cover shooters, the first person shooter genre had finally gave the player a body, and enemy fire finally felt dangerous.

But this virtual body also felt sloppy, broken, and unheroic. Players were not Tom Hanks leading the charge on Omaha Beach. Instead, players felt more like the rando Midwestern farmer recruit who has a girl back home, who then promptly dies to a sniper. In first person games, you never really know when you were exposed or safe, which often leads to player frustration. They wanted the fantasy of Omaha Beach's gritty chaos, but without the gritty unheroic uncertainty.

Kill.Switch (2003) was an obscure third-person military shooter with an innovative cover system that seemed to defuse the uncertainty of cover. When holding down a "cover" button, the player could snap to a nearby wall or barricade and also "blind-fire" around their cover-box. Battlefield and Call of Duty are first person games that can't do what Kill.Switch does: in a third-person perspective, you see your character place their back against the wall to indicate they are in "cover mode", and thus invincible from front-facing gunfire.

If Kill.Switch represents the beginning of snap-cover systems, it also represents the invention of the "cover box" necessitated by a snap-cover system. Here I will define the cover box as a frequently duplicated rectangular object that conveniently accommodates the bounding box of a crouched humanoid character.

In the video above, starting at 3:40, notice how Kill.Switch's training level starts with gray concrete blocks everywhere -- then, after you've trained yourself to recognize cover (standing-height, chest-height, boxy dimensions) you can finally begin a mission "in the field" where the cover objects are a bit less readable but still very recognizable. This is a game about learning to see the entire world as a collection of gray boxes.

Gears of War 1 (2006), as documented by history and developers interviews, took the basic cover implementation of Kill.Switch and polished it to a sheen. The feel of running and taking cover in Gears of War is much faster and tighter, epitomized by the "roadie run" which shakes the camera, pulls in the camera's field of view, and feels like an "extreme" sprint, even though you're not actually moving much faster.

Gears of War also, very consciously, inherits Kill.Switch's cover box. In an interview or article somewhere (I can't find it, but it's somewhere), the Gears of War artists and designers admitted they had trouble designing enough different cover box objects for each level. Their virtual world was a world full of crushed cars and concrete planters and crumbled wall slabs, all sculpted to the same boxy dimensions.

Soldiers in Gears of War 2, shooting from behind a gray box
Gears of War 1 may have popularized snapping to cover, but any shoot-y game like the rebooted Deus Ex series, The Last of Us, Mass Effect 2 or later, Uncharted series, etc. now has a strong reliance on snapping to cover. The player's animation change helps erase the uncertainty of cover and body awareness, and it also helps you roleplay as an action hero.

I use "snap-cover" to differentiate these games from other shooter games that rely heavily on cover. Such cover-based shooters like Counter-Strike, ARMA, or Plunkbat (what cool people call PUBG) use lots of cover... but provide a more diverse typology of cover, thus letting players easily expose themselves or misunderstand battle lines. These games don't really have "cinematic" animation systems to snap player avatars to walls, and I think they play much better for it. They forego the fluency fantasy for more depth, and that certainly makes sense for multiplayer games.

Let's review why snapping seemed good and interesting at the time.

An overview of a Gears of War 1 multiplayer map "Gridlock", filled with gray boxes in the middle
Snapping enhances readability and it makes cover more discrete / more predictable. If you play Gears of War 1 and you snap into cover, then you're in-cover, period. Until an enemy flanks around, you can be assured that your foot or your hand isn't sticking out for any pesky snipers, or even if you were exposed, then the AI will magically lose its ability to hit you. It shifts the game's focus away from precise movement, and more to tactics and territory and positioning.

From a production standpoint it makes third-person games seem much cooler, because that means you can snap your 3D characters into expressive emotional states. And because they're responding to the local cover conditions, these characters seem smarter / sophisticated / "cinematic" / expensive, while allowing for the action movie feel of being safely bombarded with bullets.

For narrative designers, snapping into cover slows the player down, so the game can deliver some voice-over dialogue and story exposition. Want to make sure the player will be forced to sit and listen to someone talk? Just spawn an NPC on some higher ground, and script the NPC to lay down a hail of golden machine gun fire. Now the player is stuck. It's basically a variant of Half-Life's "lock the player in the room while people talk" trick, but this time the walls are made of guns, and the player has the illusion of choice and skill.

The result is a bunch of big courtyards filled with piles of evenly spaced waist-high stuff. It's big and spaced out so that it'll take a while to traverse, and it helps pad out the encounter duration.

A warehouse scene from The Last Of Us, featuring 8+ waist-high rectangular objects in a yard
"Cover should always be clear and readable to the player."

Sure, that sounds like a very reasonable level design guideline, but it's also where everything starts going downhill. To understand why, we can confer Epic Games' actual design documentation page for Gears Of War level design theory: "If you take nothing else from this document, take this. A player who is paying attention should always have a chance to see an enemy attempting to flank them, and have a chance to react." 

Or, better yet, here's an excerpt from a section called "Fuzzy Cover"...
"Players need to feel safe in cover. They need to be able to recognize useful cover at a glance before moving to it, and cover needs to behave predictably because players don’t want to experiment in the middle of a firefight. When cover doesn’t fill these needs, we call it “Fuzzy Cover”. Examples of fuzzy cover are foliage, chain link fences, railing where bullets could pass through the holes, short cover with sloped sides that result in parts of the player being exposed, pillars you can take cover on that are too narrow to actually protect you from fire, small alcoves the player can’t actually fit into, etc. Cover needs to be viable for protection, or scaled down so as to remove it as a safe option from the player’s mind."
I get it, and I understand why someone would write this paragraph, but it also makes me scream. Vague cover, fuzzy cover, and soft cover... is WHAT MAKES COVER INTERESTING!

Cover nodes as seen in the level editor, from Nick Urko's scripting example for a Gears of War level
The future of cover is not the cover box. We must kick the cover box aside, and move beyond its limitations. I argue these limitations were borne of baked keyframed animation systems, a played-out "cinematic" action hero aesthetic in military realism shooters, and a boring modernist ideal of always-legible discrete cover. What else can cover be?

I think the cover shooter should learn to embrace uncertainty and messiness, and create a world full of vague fuzzy soft liquid gooey melting bubbling cover. Make a cover shooter where the world hugs me and suffocates me, cover me with danger and irony. Hell, go ahead and cover me with leaves, with bees, or maybe in grilled cheese. Just cover me.

Have you seen the world today? We're all under fire, under attack, all the fucking time. Life feels like this smoldering wasteland full of snipers and machine gun nests. It'd be a shame to reduce this feeling to a flat yard full of boxes.