I realized that despite working on Destiny and Destiny 2 encounters for years, I’ve never really talked in any sort of detail about what goes into that. Let’s change that!

If you haven’t done it already, it’s worth reading my Fundamentals of FPS Gameplay article, which is a 5-10 minute read that I originally wrote for this Destiny article.

The Prototypical Destiny Encounter

Most Destiny encounters are variants on a particular encounter design that the Bungie team brought forward from the Halo days.

The idea is simple: Most encounters are about claiming space as the player pushes forward. The onus is on the player to move combat forward. Changes to the battlefield are triggered by player progress, either through space, kills, or through objectives.

The encounter starts with the player approaching the encounter space from a non-combat area, labelled “player owned space” below. The AI are assigned to different zones on the battlefield, usually split conceptually into front, middle, and back lines. 

The player needs to fight through the front line first. If they kill enough combatants in the Front Line zone, then the AI controller will disable that zone, and combatants will no longer go there. Any remaining combatants will retreat to the middle (and overflow will go to the back). Often this also triggers the spawning of reinforcement waves.

Repeat for the middle zone. 

The back line will show the stiffest resistance, but upon either completing an objective, defeating a boss, or eliminating all enemies, the encounter will end with the player having safely claimed all the space.

Most Destiny encounters are just increasingly complicated variants of this. Sometimes the encounter “resets” in the middle and you need to re-orient yourself as enemies spawn and claim space elsewhere. Sometimes we have multiple possible approaches. The exact shape of the zones, cover layout, and line-of-sight considerations may all change. 

Implementing AI Logic

I won’t go into fine detail here, but a quick overview of some concepts.

Authored vs Systemic

The tools for Destiny encounters are built to give designers authorship over the zones and the squads of enemies (usually 3-6), and therefore the overall flow, but to automatically act intelligently at the individual combatant level without requiring manual authorship. We certainly were not, for instance, scripting every combatant to do a fallback behavior one-by-one.

Firing Areas

Firing areas are volumes we placed in the environment. Squads can be assigned to firing area via Tasks. 

Tasks

Every squad is assigned to one task at a time, and tasks can have logic and stats like “Max squads”, “Max Deaths”, and so on. Max squads is just what it sounds like. Max Deaths means that after that # of combatants have died while assigned to the task, the task is disabled permanently. Tasks are always in a prioritized list, where squads check every task in order until they find one that has room and is valid. The list of tasks is defined in the AI Objective.

AI Objectives

There is typically one AI Objective for an encounter, and all squads within that encounter are assigned to it. It contains a list of all the tasks. “Objective” is a bit of a misnomer, since the combatants don’t really have a goal other than to be where they are supposed to be. 

Example

If we look at the prototypical Destiny encounter, we can map these concepts pretty easily to it. 

There’s one objective, with three tasks, one task for each zone. The Front and Middle tasks have a Max Squad count. The Back task has no maximum because we want an uncapped number of enemies to be able to fall back to it. The Front and Middle tasks also have a Max Death value: the number of enemies that must be killed in that zone before it turns off, causing them all to fallback and redistribute the squads between the Middle and Back tasks. Repeat once more for the Middle. 

And that’s it! Other than spawning the enemies, that’s all the setup required to make the prototypical Destiny encounter within the Destiny engine. You can tell that the tools were designed with this kind of encounter in mind, and that they’re focused on letting designers think about encounters in term of owning space.

From a second-to-second gameplay perspective, what does it mean for combatants to “own space”, anyways?

Owning Space

The John Wick Fantasy

Within Bungie, designers would throw around the idea of “the John Wick fantasy”. John Wick is an action movie protagonist who moves quickly and skillfully through the battlefield, much more skillfully than his opponents, who are slow and can only react to him. He gets to the perfect spot, gets line-of-sight, and kills his enemies with a rapid double-tap. They fall before him and he can quickly claim their space and clear a room.

We had a lot “infantry fodder” combatants in Destiny encounters that exist primarily to make players feel fucking awesome. They die quick, with either with a few body shots or a single headshot. They’re dangerous en masse but not alone. They don’t really project danger to the player very well, because the player can kill them faster than they can kill the player. 

They cannot own space. 

The Anchors

We had some other combatants in Destiny whose function was to be the “anchor” for a squad. They have more health and cannot be killed quickly like the fodder can–at least not without the player putting themselves in danger or using up resources like heavy ammo and abilities with cooldowns.

Being tanky is what prevents them from being cleared trivially by just a few shots. But their other important function is that they can project danger. It is dangerous to be near them. They have shotguns, or AOE attacks, or homing missiles that move slowly enough to be dodgeable at range, but not when up close.

Anchors are always physically large, because we want the player to know where they are. If the players aren’t aware of what space they do or don’t own then a lot of this gameplay breaks down.

With these three traits, anchors can own space.

Something that’s interesting to note: as players grew more skilled and powerful in Destiny 2, the ability for anchors to hold space, assuming equal level scaling, was reduced. This meant that we had to gradually introduce more and more Elites (beefier versions of combatants) and have more and more enemies on the field. This was ultimately a big factor in the Destiny 2 combat experience slowly breaking down over the years. It can be illuminating to go back and compare to the first Destiny, where base, non-elite anchors still function in their roles much better.

Other Combatant types

Not every combatant in Destiny fell into the “infantry fodder” or “anchor” categories. Melee units, for instance, would rush into player-owned space to flush the player out into the open. Snipers filled a role similar to anchors: they could lock down space they had line of sight on, and you had to take them out to be able to safely exist there.

Directing Player Attention

One of the less obvious things an encounter designer has to do on Destiny is think about the player’s attention. Where is the player looking? What are they seeing? What are they aware of? How can we get them to look where we want them to?

In the example encounter layouts above, it’s pretty simple: the player’s attention is always forward. They’re gonna see stuff that’s down range from them, where the enemies are. 

But Destiny encounters get a lot more complicated. The player’s attention may be split between different paths, and between enemies right in front of them vs snipers far off and to the side. There might be special mechanics happening, like an objective they need to reach in a time limit, or an object they need to pick up and take to a receptacle.

The only way to really be sure is to playtest, but there are some ground rules we can follow to direct player attention: 

• Players generally approach combat with the goal of fighting to the end of the encounter space, or the goal of defeating the boss. We can generally assume the player’s attention will always return to these goals. 
• Strategies that level designers use to make environments that are navigable apply here: players will tend to look at brighter things, towards landmarks, and will follow paths suggested by the layout (unless they can just jump over everything, which is a real factor in Destiny!)
• Anchors! Players tend to look at the biggest visual mass of enemies. So if there is a cluster of a bunch of fodder together, or an anchor and some fodder, those are prime candidates for what the player will want to look at. 
• We can try to anticipate which enemies players will defeat first, and extrapolate from that to assume certain paths that they’ll take in the environment, but that’s very unreliable. 
• UI elements. Sometimes we simply need to call out things with quest markers.
• Loud noises can pull player attention, but these are unreliable for a lot of reasons. For one, the audio mix in Destiny has a lot going on. It might play at the same time as the player using their super! Another problem is that players may not be hearing the audio at all: they may be hearing-impaired, or they may be playing with sound off, or they may have voice chat turned way up and the game audio turned way down.
• Enemy movements. The players’ eyes will tend to follow when they see enemies moving to new areas, or they’ll seek out enemies to fight in a location. 

One of our most common failures to direct player attention is when we spawn enemies behind the players. We rarely want to do that, but players move around quickly and combat environments and timing can get chaotic. If we’re gonna spawn reinforcements we either needed to do it behind enemy lines, or we had to be really sure that the player would see it.

Sidebar: The Infinite Forest

I was the primary designer for the Infinite Forest procedural dungeon/encounter system that we shipped with the Destiny 2 DLC, the Curse of Osiris. I was pretty inexperienced at the time so I had to get a real crash course in understanding Destiny encounter design.

For most of the Infinite Forest development, we did not have the ability to define firing areas for the dungeon rooms: there was always a single firing area that took up the entire room, and combatants could be anywhere in the room, relying purely on their own NPC AI logic to pick where to stand and what to attack. This was terrible! The encounters were very boring, had no sense of progression, and tended to feel chaotic.

Later on we got the ability to define fallback firing areas, but we were missing something: we had no way to tell the system what direction the player was coming from. We couldn’t define a progression through a linear space like most Destiny encounters. Instead we had to define a big firing area that covered the whole room, and a smaller one that covered just a central location for the survivors to fall back to after enough of them had been killed. It also meant that every room had to be somewhat circular.

If you ever wondered why Infinite Forest encounters could feel kind of anemic compared to others, this is a major reason why: we couldn’t finely tune the ownership of space and combatant movement flow.

Combat Legibility

This was a drum I was trying to beat a lot towards the later years of my time at Bungie: our encounters were getting illegible! What do I mean by that? Well, legibility is how well we can discern the shapes and meaning of something that we’re looking at. Handwriting is legible when you can read it. Traffic signs are legible when you understand their meaning.

A prototypical Destiny encounter of all red bar combatants, with just 3 types of combatants, in a mostly flat plane with some cover, is very legible! Pretty easy to see it all and understand what everything is.

But there are a lot of things that can decrease legibility:

• More unit types
• More unit “tiers”: base, elite, boss
• More combat fronts
• More verticality in the space
• Special mechanics
• Flashier VFX
• More factions: combatants are designed to have unique silhouettes and abilities as much as possible, but within a single faction they are also evaluated, designed and tested as a set. If you have mixed factions things are less likely to read well
• The same unit types but with different weapons
• The same unit types but with different tiers
• Color and lighting of environment vs combatants in it

And much, much more. I had a document at Bungie where I listed something like 50 of these bullet points. My (perhaps unrealistic) goal was to try to come up with a point-scoring system that allowed us to “measure” legibility for different combat designs. I was laid off before that goal came to fruition, but even without objective measurements, designers had to think about this a lot as they built encounters. 

Ultimately, in almost all cases we want the player to be able to see and understand what’s happening in an encounter, so they can engage in tactical and strategic thinking, make interesting choices, and choose to take risks. Simple chaos is rarely desirable.

Other encounter types

The concepts I’ve discussed in here so far are largely universal across all Destiny encounters, and the prototypical encounter structure I mentioned above is simply a strong foundation that hits all the notes of a good encounter in flexible ways. Encounters can and often were designed in all sorts of other ways. Some other common encounter structures:

• Boss fights in an arena
• Waves of enemies in an arena
• Defend a point
• Simple open world encounters where there’s just one zone and the player can approach from any direction
• Tight quarters (ie combatants hiding around corners in a maze)

Most of these still used anchors and still had an underlying philosophy of designing around owned space. They simply used different progressions through space, or reset ownership in different ways, or changed the conditions that controlled which units could project danger.

Thanks for reading y’all!