How does enemy MELEE and RANGED combat work?
Enemy Combat Design IV
This is the first essay series for Signals and Light, and it provides an overview of enemy combat design in action games. It is written from the perspective of a AAA game designer who works with enemies in action games and aims to pay it forward with what he’s learned.
The fourth part of this series explores the design and terminology for various types of enemy attacks based on their distance from the player. It covers…
The core player verbs for interacting with incoming melee and ranged attacks.
A definition of melee attacks that is based on the distances they travel.
How games use motion warping to ensure that enemies can approach a player accurately while a melee attack is active.
The basics of various types of ranged attacks and how to think about them in relationship to animation.
Other types of unconventional boss attacks that don’t fall into obvious categories.
And lastly, an analysis of how the distribution of ranged and melee attacks influences the sub-genre of an action game.
As established in Part 1, enemy attack cadence is divided into four pillars. The second of these pillars is PROXIMITY, which describes the distance between the enemy and the target of their attack when they actually activate the attack.
(See Part 2 for a full explanation of the enemy attack breakdown.)
Proximity plays a role in a few parts of the attack, particularly…
How far the enemy can be from the target, as determined during the Assignment of Attacker Role step.
Which actions can be selected from that distance, as determined during the Selection of Action step.
Where the enemy needs to move to start the action, as determined during the Preparation for Action step.
How far the enemy can travel during the attack animation, as determined between the Anticipation Frames all the way through to the Resolution Frames of the attack.
Proximity starts with LITERAL DISTANCES—numerical values that the designers and animators tune in the game based on the specific attack animation and the overall gameplay purpose and outcome of that animation. To properly analyze proximity, though, it is arguably better to look at it through a lens of what TYPES of attacks enemies can use and how those attack types can close the distance between the enemy’s damage and its target.
IV-A - melee attacks
Broadly defined, melee attacks are those that require the use of the character’s model or some object it’s holding to directly touch the character’s target in order to inflict damage. This could be just a weird little dude touching Mario, or it could be the Nameless King smacking the player with a lightning-wreathed spear as large as their character model. When it comes to designing these types of enemy attacks, one of the most important questions to answer is how far away the target can be when the attack actually activates.
In the case of the deterministic Goomba, there is no logic, it is always moving at the same rate, and it is always dealing damage. The reactive Goomba, on the other hand, waddles at Mario when it’s at the right angle and distance from him. But in these two examples, the enemy is just a hitbox. In deeper combat designs, melee attacks break down into a handful of specific rules, and when one considers proximity in attacks, a melee attack does NOT only mean an attack that the enemy activates from a close-range proximity.
The distances that enemies can travel between attack activation and the first hit frames vary across different attack types, and it is quite common for action games to have a spectrum of long-range and short-range melee attacks. While there is no global development terminology for this, an easy-to-understand set of terms for melee attack movement patterns are:
In Place - The enemy doesn’t move much or moves very little.
Lunging - The attack is meant to close a medium or somewhat large distance.
Charging - The enemy crosses a very large distance to reach the target.
How the player can react to melee attacks depends on a game’s genre and the tools made available to them. Common mechanics include dodging, blocking, parrying, or simply moving out of the way. The ideal action to use relative to the distances that an attack travels depends on how fast the enemy crosses that distance (which ties to the previous pillar of attack timing). Generally speaking, the further an attack travels, the more time a player has to react, so often, charging attacks become more threatening through large hitboxes and deal heavy damage. However, some games are defined by having incredibly frequent long-range melee attacks and a great deal of movement by the player and enemies. In these cases, the distances traveled have no relationship with the danger of the attack, and other features of attack animations are leveraged to communicate the level of impending threat, such as timing and anticipation poses.
This essay will set aside the “how the attack hits” aspect of this diagram. It’s not important here to define how hitboxes, hurt boxes, and melee collision can work in different kinds of games. One can find more information on that topic in other places.
What is important when discussing proximity is that in almost all sections of the attack, the enemy may be translating and/or rotating toward their target. In the Preparation for Action step, the enemy has decided to attack, but the attack hasn’t actually activated, and they’re probably running towards the target.
In the animated sections (past the Anticipation step), animators create the movement in the attack animation itself—such as a character leaping through the air to slam their hammer down. Any distance traveled here is part of the attack itself. Usually, however, gameplay systems can programmatically alter this movement within limits if the target gets closer or further away as the attack animation completes. Similarly, the rotation of the attack in these sections is updated within specific windows and constraints as well, also reacting to the movement of the target. Some gamers might call this “tracking,” but it is usually more complicated than it seems.
In recent development terminology, this procedural movement and rotation is sometimes called “motion warping.” Without digging into the actual technical development tools, the simplest way to put it is that, in the most refined combat animations, designers are working with animators to define clean sections of the animation where they can modify translation and rotation to bring the attacker to the target and keep the attacker within range for its hitboxes to connect with the target. Their aim is often to keep the fidelity and believability of the animation intact during the procedural movement, especially to avoid what animators call “foot sliding.” Some amount of foot sliding is common across all games, but as fidelity goes up, so does a team’s desire for realism in character movement.
Games experiment with this space much more than might be immediately obvious, and a great deal of the game-feel of melee attacks is a result of what tools and rules the game uses to define these periods of movement. Some of those rules might be globally consistent, perhaps made by a relatively brute-force tracking system, but in modern games, they’re usually more subtly employed to perfectly match the nuances of the animation.
What is interesting is that much of the difficulty in melee action games is a product of painstakingly designed movement possibilities for the enemy relative to the player at every single frame of the attack to result in something that feels satisfyingly challenging. A good example of this is found in Monster Hunter. Much of the challenge of the faster monsters is a result of attacks with faster, more frequent, and more far-reaching rotation and translation opportunities. Monsters earlier in the game have broad-stroke actions and movements that are imprecise and spend less time honing in on a single target after they begin. And then later monsters, by comparison, sometimes reorient to the target to activate a hit multiple times within a single attack animation, requiring more reactive responses.
So, while melee attacks may seem simple on the surface—one thing hits another thing—in reality, there is an enormous creative space for designing them with proximity to the target in mind.
IV-B - ranged attacks
Ranged attacks damage the target without requiring any part of the attacker’s model or equipped items to make contact with the target. They can fire bullets, missiles, or abstract harmful VFX projectiles, they can summon area-of-effect shapes, or they can…Really just do all kinds of other things. The full breadth of ranged attack types that exist would be pointless to define in exhausting depth, but at a high level, there are some broad categories.
First are instant ray-traced or near-instant fast-traveling projectiles. These are most often found in first-person and third-person shooters, and they mirror the projectile methods that the player also uses. When simulated bullets behave somewhat like (or, in a few cases, exactly like) real bullets, many things about the gameplay loop become locked in place. The player can’t dodge individual enemy projectiles and, in fact, probably won’t be interacting with the enemy projectiles in any way other than to be hit by them.
There are three major player defense approaches to hit-scan-focused games:
Cover becomes the primary language of player defense and combat interaction, and these games are often literally called “cover shooters.”
The game focuses on long-term regenerative health and expects the player character to be riddled with bullets and find opportunities to regenerate health.
Or, lastly, PvE firefights are not even the primary focus of the gameplay loop (such as in stealth games), and the player is expected to die quickly if they remain in combat.
The animations used for this firing type are usually very much function over style. The player needs to know who’s shooting or who’s about to shoot from a single glance at the pose, but there aren’t many flashy attack animations.
Next are projectiles that move slowly enough for the player to react to them. This usually means dodging or avoiding them, but it can also mean blocking, parrying, or otherwise manipulating them. Firing animations for these projectiles can range from basic to stylistically telegraphed. Many melee-centric games with limited ranged combat will have enemies post up on a ledge and periodically swap between an idle animation and a simple firing animation. In 2D games, enemies may have simple or even no firing animations at all. In projectile-heavy games, however, sometimes enemies draw clear attention to themselves through animation pose, audio, and VFX to communicate that they are going to fire a particularly dangerous projectile.
Interactive projectiles can do a lot of heavy lifting, with some game genres being entirely defined by their complex enemy and/or player projectile patterns. As a gameplay tool, they have vast utility across genres, and their fundamental value is an obvious one: they are a game object, they can possess any attributes of a moving game object, and thus there is an ocean of depth in how they can behave. With that said, there are arguably limits to how an enemy projectile should behave compared to a player projectile. Predictability and readability are key to good enemy projectile design, and in a deep dive on just that topic, one could likely reveal a relatively small list of common enemy projectile behaviors.
Lastly are areas of effect (or AoEs), which can come in as many forms as projectiles (or even be spawned by projectiles). The important thing about AoEs is that they often implicitly encourage similar gameplay patterns as level hazards. No matter what an AoE does to the player, the important thing is to NOT be in it when it activates. Some games telegraph this with in-world UI, and some games have the player intuit invisible hitboxes based on gameplay VFX.
IV-C - attacks as hazards
Aside from melee and ranged attacks, there’s a broader and vaguer category of “attacks as hazards.” Many games use these attacks in boss fights, and these attacks are more complicated than just AoEs for the player to avoid. Their “proximity” to the player is generally “in all directions,” as these types of attacks tend to become parts of the combat arena itself.
First are attacks that create dynamic hazards or interactive objects in the arena that feed into how the player navigates the arena and defeats the boss. There are then attacks that completely change the state of the arena, though sometimes these aren’t attacks at all and merely state changes to the level design tied to boss health phases. There are also bosses that are the arena, and the “attacks” that they do are actions that affect the player state to create different opportunities for navigating them. All of these really create a fuzzy line between combat design and other forms of game design, but they are worth accounting for.
IV-D - proximity and level design
The proximity paradigm of a game’s enemy attack design has ENORMOUS interactions with the game’s level design. The distances that enemies travel to perform melee attacks, the distances that projectiles cover, and the sizes and shapes of AoEs are all LITERALLY constrained by the two-dimensional or three-dimensional space in which the gameplay is happening. Less space means shorter distances, more space means longer distances.
Additionally, the number of obstructions within a gameplay space also determines the outcome of an attacker and its target moving throughout the space during the duration of attacks. The utility and ease of cover shooting, leveraging chokepoints, or avoiding attacks through using large obstacles varies significantly across subgenres. All of these factors are some of the main vectors through which level designers and combat designers have to collaborate, and the best outcomes for both disciplines are a result of clear alignment between arena and enemy design.
IV-E: discussion - approach to proximity ≈ sub-genre
THIS ESSAY PROPOSES…that one could map differences in enemy attack proximities within games to a spectrum of simplicity to complexity in combat design, and the choices developers make here help to inform a game’s subgenre within its genre.
On the SIMPLE end of the spectrum, the simplest and most refined combat designs generally have a small set of possibilities for what a player can expect from melee and ranged enemy attacks. Either the game does not expect the player to engage with many attacks at DIFFERENT proximities at the same time (more on this in the next pillar, enemy attack CONCURRENCY), or developers design enemies around strict proximity rules and/or encounter placements such that the player can parse the enemy behavior easily and quickly without getting confused.
On the COMPLEX end of the spectrum, AAA action games open up a Pandora’s Box of complexity in terms of the types of incoming attacks that the player must be aware of from varying proximities. In more complex games, multiple attacks trigger simultaneously or within brief time windows of each other regularly, and the proximity of those attacks can be across the entire arena, whether melee, ranged, or AoE. This requires the player to TRY to be aware of the whole combat arena at once in order to properly anticipate incoming threats, but whether or not they can do this depends on the game’s camera (and, again, more on that in the next part covering attack concurrency).
Take the above challenge arena from God of War: Ragnarok, for instance. Enemies can spawn at a couple of locations, and there are both ranged and melee enemies in the spawn sequence. Any of them can attack the player at any given moment. Something the level and encounter designers do in this case is always spawn some ranged enemies on some specific ledges so that the player immediately understands where to expect ranged attacks to come from. The game also generally restricts ranged attackers from attacking from off-screen as often as melee attackers can (more on this in Part 6), making it easier to parse the incoming ranged attacks as they occur. These considerations clarify the game’s approach to proximity for the player, making chaotic enemy encounters slightly easier to understand.
So, this essay makes the claim that enemy attack proximity helps to determine subgenre. Of course, no team is sitting down and saying, “We want really long-range enemy attacks” when they describe their desire to make a first-person shooter, but even within the shooter genre, there is a clear set of rules and expectations given what developers expect for the longest and shortest ranges of player and enemy attacks. A game where the player only has a short-range projectile-firing weapon requires different enemy attacks than a game where the player has a hit-scan sniper rifle. As another example, when developers decide to give the player ranged options in a melee-focused game, they’re also obligated to make enemies fun to interact with at longer ranges and reckon with what it means for the player to parse ranged versus melee enemy attacks. The team also plants a subgenre stake in the sand when they determine not just how often projectile-firing enemies show up in a melee-focused game but also how the player is meant to interact with incoming projectiles in terms of player actions and combat arena design.