Universal Truth Number Three (pt.3)

I wanted the Halo weapons to have depth, so I began thinking about all the guns that were in the matrix. I needed to understand what they were, and how they fit into the design.

The Human weapons were easy to understand. I’m a Human, and I know what we use guns for. But the weapons used by the aliens of the Covenant were another matter. The easiest place to start would be to simply say that the alien guns were simply analogs to the Human weapons on the matrix. The pistols, assault rifles etc. could be basically the same, only with different visual presentation. Easy, yes. But that seemed like a huge missed opportunity to add depth and richness to the game. So I started thinking why would the Covenant choose these particular weapons in the first place?

We (Humans) have guns. And once guns were developed, Humans developed systems to protect people from bullets (bullet proof vests, riot shields etc.) And then in the relentless march of progress, people invented ways to kill other people inside of their body armor (armor piercing bullets etc.) Remember that at the time there wasn't a lot of settled "lore" about the game story. I decided that in my model, Human Spartan armor was created as a desperate response to the Covenant attacks. It had similar functions, like a personal shield, but was based on different technology.

So how about the Covenant?

There were some notes about the bad guys and their guns, but the honest truth was that the aliens shot light-up bolts of energy because they looked a lot more visually impressive coming towards the player on screen. If the bad guys shot nearly invisible bullets and you couldn't see them coming at you, it would be a total drag every time you died. But just knowing that they were colored lights wasn't going to help me balance my combat data.

Clearly they had guns. And they had an equivalent to our body armor – personal energy shields. I could imagine Covenant warriors facing off against enemies across the universe with their plasma weapons blazing. Or more specifically, their Plasma Rifles. As an poor man's analog to the Human pistol, the Plasma Pistol was a pretty dull thing, only useful as a desperation choice for one of the two gun slots you were limited to. I stared at the various data fields in the Halo toolset for quite a while, trying to imagine what to do with the Plasma Pistol to make it cool. And then a question occurred to me: What if the Covenant had to fight an enemy with shields like their own? Or what if they had to fight themselves? They’d need their own armor-piercing capability.

In the Halo tools, every projectile had a “shield damage” value. Most were set so that they would damage shields at a rate that matched the damage that their bullets would do to the player's health bar. None of the projectiles were really aggressively balanced against shields. And you know how I feel about data balance in a matrix! I started to experiment with making Plasma Pistol bullets designed to specifically shred shields. It was a snap to make a projectile that blew them off quickly, but then it seemed overpowered to also make those bullets do good levels of “body” damage as well. Then it occurred to me: maybe the shield-shredding effect could be assigned to a different bullet. The one assigned to the secondary fire-mode for the gun – the overcharge.

This proved to be very fun. In my early playtests, I'd grab the Plasma Pistol and use the overcharge specifically to blow up the shields on enemies that I ran across. But it was frustrating when I missed the overcharged shot (full disclosure: I am a much better designer than I am a player) So to compensate, I gave the shield-busting projectile a terrifying amount of magnetism so that it would track towards whatever I shot it at. I loved it – I could overcharge the Plasma Pistol and let the shot fly, and it would whip around corners and blast targets, stripping off their shields just as I came running in behind and mowed 'em down!

In the short term, I won a lot of playtest games. Unfortunately, once this tactic became known to other players, battles essentially started with “overcharged salvos” of tracking shots whipping across the battlefield. The only thing you could do was hunker down in cover and wait as the first round of supercharged shots came whipping overhead before you started moving. It was interesting to see how these data adjustments changed player behavior during our playtests, but a bunch of auto-tracking missiles wasn’t very true to the spirit of the Halo combat model which rewards player skill, fire and movement. So alas, the “super tracker” overcharged shots had to go. But I did keep some tracking, to help reduce the frustration of a player using the overcharge trick but missing the shot entirely.

So my mental model of bullets/armor/armor piercing was working to create fun combat. But what else could it do for the game?

I thought more about the Covenant weapons. Since they were the “bad guy” guns, I really didn’t want them to overshadow the heroic Human weapons . I wanted the Human guns to make the player feel like a badass space marine, putting the smack down on alien scum. The Covenant weapons needed a different mental model.

Jason Jones had (reluctantly) allowed me to include powerups in the multiplayer maps. I’d always loved the powerups in Quake, and how they factored in to the player awareness of the map and the timing of their respawns. (someday, if folks are curious, I’ll talk about the level designs for the Halo multiplayer maps) The Active Camo powerup in particular seemed interesting, especially once the single player designers started to add them to their campaign maps. I reasoned that Covenant troops might use Active Camouflage technology as they were walking into battle, kind of like the Predator. Covenant troops would rely on stealth to position themselves and then unleash hell on their enemies. So instead of the stand-up butt-kicking mental model of the Space Marine guns, the Covenant weapons could be built around ambush.

I asked Adrian Perez (at the time newly hired to Bungie and assigned to support gameplay code for the multiplayer team) if he could extend the “shield damage“ system and give me some control over how bullets affected the active camouflage powerup. Adrian (codename Cuban) was more than up to the task, and had it running quickly.

My goal was to make different types of bullets react differently to invisibility. I reasoned that because the Covenant put a premium on ambush tactics, their weapons wouldn’t disrupt the invisibility shield as much. I adjusted the data so that Covenant weapons could repeatedly fire from invisible, and the player would barely fade into view just as the weapon was reaching it's overheat threshold. And when they stopped firing, they would quickly disappear again as the weapon cooled.

This worked spectacularly well. The only problem was that initiating an attack from stealth wasn’t as effective as I’d hoped. Because of their lower firepower, Covenant plasma weapons lacked the punch to take out an enemy fast, even if you got the drop on them using camouflage. I avoided adding damage, and instead tried to address this by adding a little bit of “hit stun” to Plasma weapon shots. Hit Stun is a value that freezes the target for a little bit of time with each bullet that hits them. It's a very useful tool in the hands of AI bad guys, because it can make even the weakest bad guy threat something that the player has to pay attention to, if only because of the annoyance of the freeze from their shots. But it must be used sparingly. My initial data adjustments made Plasma Rifle shots freeze the target so much that, if somebody started to shoot you with it, you were basically helpless and could only stand still and take it while they fired their gun to overload. Needless to say, that was extremely frustrating to players on the receiving end of the attack.. I got pushback from the playtesters to remove the mechanic entirely, but I had faith in the mental model that I had developed and tried to find a level of “hit stun” that gave a Plasma weapon enough of an advantage so that using the “Covenant ambush” approach would be valuable but not overpowering.

With those aspects of the model coming together, I started to think more about the Human guns. They already had an edge in damage, and I didn’t want to enhance them with the same benefits from Active Camouflage. In fact in my mental model, they should behave just the opposite. I figured that Covenant Plasma weapons were “tuned” to interface with the camouflage so that the shielded warrior could fade back to invisible more quickly. So then it made sense that Human physical bullets would disrupt the energy field that surrounded a body using Active Camo. To get that effect I adjusted the invisibility "ping rates” for the Human weapons to be very high. The bigger the bullet, the faster it would make the target become visible. After playing with the numbers, I was able to adjust the Assault Rifle so that the shooter would become visible after just a few shots. Adrian’s changes to the code also allowed me to adjust how quickly the shooter would become invisible again when shooting from camouflage, so I was able to add a satisfying delay to the invisibility after the Assault Rifle was fired. Finally, I decided that the numbers for the Rocket Launcher would be the biggest disruptor to the "energy bubble" of active camouflage. Firing a single Rocket from stealth made the user immediately visible, and left the Active Camo field disrupted for several seconds afterwards, leaving the shooter fully visible until it slowly re-formed.

I made one other change under the hood of the Human weapons, which many people don't even realize exists at all. Jason Jones had designed the Human pistol to be the weapon of choice for players at medium/long range. The accuracy, high damage and the limited sniper zoom on the pistol made it a powerful choice for dropping enemies right at the edge of their "AI awareness" bubble, enabling players to pick off one or two targets as the enemies startled into their alert state and then came charging into battle. But it was strong. Damn strong. Frankly, it was too strong for multiplayer. I toyed with damage settings that made the multiplayer pistol weaker than it's single player counterpart. But to be honest, once it was "nerfed" it became a pale shadow of it's single player cousin and using the pistol became a lot less fun.  Still, I felt that turning the full power of the pistol loose on the Halo multiplayer "sandbox" unaltered would be opening the door to endless criticism, so I decided to made a subtle change. The single player version of the pistol is "autofire" - meaning that if you hold the  trigger down the weapon will repeatedly fire at the precise point you're aiming at. But... that's not true with the multiplayer version of the pistol. I wanted to at least challenge the skill level of players a little more. So the multiplayer version of the Pistol has shot spread. What that means is that, if you simply hold the trigger down and let the pistol automatically fire over and over, each bullet will deviate from the point that you're aiming at. And the amount of deviation will increase with every bullet. I wanted to make it so that players could still use the badass pistol, and it could retain the fun feeling that it had in single the single player game, but only if the player could master the technique of actually pulling the trigger with each individual shot. I still believe that this was a "righteous fix" - meaning that it was justified and the solution was (in my humble opinion) elegant within the restrictions of the established game play. Unfortunately, I lost my nerve a little bit. After all, this was a huge change from the behavior of the single player version of the Pistol. I was worried that players might have to re-train themselves to use the multiplayer version of the gun, which again might lead to huge volumes of outrage from players. So I didn't make the pistol deviate enough while auto-firing. Oh, the shots will spread  if you hold the trigger down, but not so much so that you might not still get the head shot that you were aiming at. To this day, not adjusting the spread rate of auto-fire on the multiplayer pistol is one of my regrets. I wasn't aggressive enough!

But hey, people still seemed to like the game.

One of the things that I’m proudest of is how my mental model for Human and Covenant technologies had profound impacts on the single-player game. For example, the high camouflage ping rate of the Human weapons meant that, even late in the campaign, Human guns were ideal for exposing Covenant bad guys that were cloaked in Active Camouflage shields. A second impact was on the AI development of the game. When the mighty Chris Butcher (AI programmer for Oni and Halo) saw the changes to the Plasma Pistol, it gave him the idea to have the Jackals use the Plasma Pistol in it’s overcharged mode, along with their shields, to greatly differentiate them from the Grunts wielding Plasma Pistols and grenades.

I’d like to take a moment here to talk about why I keep using the term “mental model”. You might ask “Shouldn’t the design document cover all of this?” And my answer would be that my design documents have never captured all the details of the game. I find documents valuable in helping me codify my own thinking, and they can occasionally be good tools for communicating a design to the people that are responsible for implementing it. But I've never encountered a game development team that religiously read every document produced by the game designers. And when you're actually knee-deep in making the game, you rarely have the time to fiddle around with keeping all your design documents up-to-date. So my own process has evolved to be very fluid and organic. I start with some clearly stated intentions as to what I want to accomplish with a design, and then start to build it. But along the way, I watch the design evolve and continually evaluate that process. As things happen I’m constantly deciding, “How is thing coming together? Are we going in the right direction, or should we be going another way?” So my paper specs get me started, but beyond that my mental model is constantly evolving. I once read a quote from Tim Schaffer, that I'm going to have to paraphrase heavily because I can't seem to find the original quote. He described the process of making a video game as building a puzzle out of pieces falling in slow motion. But the pieces fall at different speeds and the shape of the puzzle changes, depending on which pieces you get, and which fit. That is a very poetic and accurate description of what my process looks like: I like to toss the pieces up, and every day take a look to see what’s coming together, what’s falling behind and what shape the final form is going to take. (I apologize, but I can't find the quote out there on the web. If you find it please add it to the comments section and I'll edit this post!)

So that brings us full circle, back to the one-sentence blurb question that I got via Twitter: was quick camo intentional? Yes; entirely intentional. All of the camouflage behaviors are a product of my mental model for Human and Covenant weapons, and my desire to add depth to the gameplay model for players to discover and exploit.

Did it work? As I said before: often players will never know all the details included to add depth to a game. The fact that a person on Twitter was asking about that feature proves that, although my mental model was thorough and effective, it wasn’t so intuitive that players completely understood it, even after a decade of playing the game.

But here’s the thing: even if an audience doesn’t understand all of the influences that shape their experience with a work of art, those influences still resonate in their mind at some level. That’s called subtext. When I watch a performance of Cirque du Soleil, I don’t know exactly what’s happening in the overall story of the performance. But I know there is a story.

And my experience as an audience member is all the richer for it.

Please join the conversation at www.hardylebel.com