Ask not what you can do for your enemies, but what your enemies can do for your player. You might think that enemies exist to challenge the player, but their actual job is to make the player do things. All the obstacles in your game are there to make your player actually use those cool abilities that you gave them, and enemies are just a particularly fancy obstacle. You can tell, because sometimes obstacles are dressed up as an enemy, but are basically just environmental hazards with a face. Particularly complex enemies or bosses are a collection of different obstacles, which makes a boss battle not unlike a level design.
The majority of this article will apply to designing either enemies or bosses -because as we know, the difference between the two is often just a matter of timing, fanfare and whether or not a HUD element appears alongside them — but fear not boss battle purists, we’ll have some big bad exclusives later.
Let’s start with some common sense: anything an enemy can do, the player has to be able to somehow anticipate, and subsequently defend against or avoid. This means two things. Firstly that any attack the enemy can perform must come with a warning (we’ll get back to that later) and secondly — the enemy move list is ultimately determined by the move list of the player.
If the player can jump, the enemy can perform a shockwave which the player has to jump over. If the player can aim and shoot easily, the enemy can fire projectiles that must be shot out of the air before they reach their target. If the player can already perform a block or dodge, well, this makes creating the enemy move list considerably easier.. but at the very least the player can usually move around, which means the enemy can perform attacks which affect a particular area of the environment and require the player to run around/away.
By the same token, when an enemy does one of these things, the player must respond with the relevant counter, or they die. Flipping the script, an enemy must have a way to be defeated, and if the player doesn’t do THAT thing — they can’t win and still probably die. Let out your best evil laugh, because you’ve got ’em right where you want ‘em.
So, what do you want the player to do? As a game or level designer, you always have an idea in your head of how you want the player to play. There’s a motto that my friend and I often repeat:
Given the chance, players will optimize the fun out of their own experience.
If you want the player to lower their shield, if you want the player to come out of cover, if you want them to dip deeper into their inventory than the highest DPS weapon, you need to make them.
A lot of this work can be done on the systems side but beyond that enemies are maybe the most powerful method of encouraging the player to do the things you want, or discouraging them from doing things that you don’t want them to be doing.
THINGS you say? Weapons, abilities, mechanics — strategies.
Highway to the Fun Zone
Huge swathes of this article will be borrowing examples and terminology from Hugo Martin’s DOOM Eternal interviews because they demonstrate an incredibly firm and articulate grasp of what their enemies and weapons are for. They talk about pushing the player into the “Fun Zone”. Fun can be a toxic metric without definition, but at id Software they’ve bottled it as a list of key activities (things) they want the player to be doing in their game e.g. moving, swapping weapons, glory kills and prioritizing targets.
Every enemy you add to your game should have a role they are intended to play, a pressure they exert onto the player to change how they behave. Simple examples are that an aggressive melee charging enemy will cause the player to retreat, a sniper emplacement will cause them to take cover, a shield will cause them to flank. Low-threat “fodder” enemies are often multi-purpose; both walking resource crates that tempt the player forward, and distractions that lure the players attacks away from other enemy types long enough for them to get their own attacks out.
In his brilliant NoClip interviews Mr Martin talks about each enemy type as part of a chess set.
The Right Tool for the Job
Martin also mentions a flaw he felt in DOOM (2016) was that the player was not pushed to vary their weapons and strategies outside of the highest modes. Powerful weapons like the Super Shotgun and Rocket Launcher worked in pretty much any scenario, so they saw the lions share of screen-time while other weapons languished in the inventory unless ammo got especially desperate. In DOOM Eternal they want players to see each weapon as a tool that fits specific jobs.
“Toolbox” design philosophy is a strong backbone for any combat system and you don’t need to make it complicated for it to land with players.
The premise is simple: Establish a relationship between an obstacle and a tool (player ability). Give the obstacle an extremely clear and consistent identifier, and let the player read that information and choose the right tool. Pick almost any game and you can probably identify these relationships clearly. Some studios refer to this set-up as “combat puzzles” and it’s a huge part of encounter design, which is when your enemy types start to mix. If it sounds familiar, it’s probably because -as promised- it’s a lot like level design. See for example: the Tomb Raider reboot trilogy.
It works because of the obstacles come with a strong and consistent visual language. This doesn’t need to be as overt as color coding or repeating a single specific prop — Lara’s shotgun can smash any light wood structure, flames can burn anything swathed in white cloth. Sekiro’s axe can smash shields of various shapes and sizes, but also the large bamboo rice-hats that those pesky “Rats” wear — the fireworks can scare any animal from boar to horse to crow. Batman needs to use his blade counter whether it’s a broken bottle or a samurai sword. The trick is that once you establish the rule, it must be consistent enough that players can internalize the rule and apply it with total confidence.
Sound the Alarm
Just like how every category in your rule-set needs a consistent language, so must every enemies attack list. We’re back onto cues and telegraphing now. We won’t linger here long, because it’s basic combat design: you need to give the player a way to predict when an attack is about to happen so that they can pick the right reaction.
Visual: This can be as simple as an enemy glowing a particular color, raising a poised claw, or adopting a pounce stance. It is a few moments (or frames) that precede the part of the attack which actually hurts. Sometimes designers cheat and have a universal cue for all enemies, such as a tingly spider sense around the protagonists head, or a button prompt that pops up somewhere on screen.
Audio: Cast your mind back over any boss AI pattern you have ever memorized and you’ll find that those annoying things they kept saying or noises they kept making were actually paired with a specific attack. Essentially the enemy sounds off with a particular noise or phrase immediately before the attack, and the player can learn that this audio cue is attached to a particular action the enemy is about to perform and prepare accordingly. This is essential for any enemies who might come at the player from off-screen, where most visual cues don’t do anyone any good.
Pattern: Some attacks have very little warning of their own, because they always follow one of the bosses other attacks, or always occur in response to a specific player action or position; such as an attack they will always do if the player gets up close, or when the player has been knocked onto their back by another move. The player learns where these attacks will occur from trial and error, so it’s important there’s no randomness involved in a low-cue attack.
In all cases, the key is that the cues of different attacks are consistent enough to learn and distinct enough to avoid any conflation.
The Art of War
Crucially, different attacks require different responses from the player, ensuring they are constantly engaged with the game; assessing the danger and reacting with the correct input, rather than finding comfort (and thus eventually boredom) in routine. All the best enemies and bosses forge a relationship with us as a player because we have to watch them closely and become attuned to their specific body language and needs. You ever date a Cacodemon? Then how come you know the face they make before they spit a fireball?
Back to the DOOM Eternal coverage again for one last nugget of wisdom: Game Director Hugo Martin puts forth the importance of keeping the player thinking. The games designers goal should be keeping the player engaged. The player is engaged when they have to pay attention and think about their next move. Again, the Q+A can be extremely simple because the player is going to jumping from stimulus to stimulus, applying their knowledge bank of rules at speed and under pressure.
Behold the bedrock of my games design philosophy: The Bop-It Theory.
Thus, we’re back to the start! After all, what is a boss battle if not a collection of obstacles, each requiring different responses? If enemies are intended to teach the player how to play the game, bosses are there to test what they’ve learned.
The best boss battles put in the work to teach players their rules before they start testing them. This allows the player to confront their enemy with confidence, and only have themselves to blame for every missed window and slipped prompt.
While a bosses various attacks might call on the players full range of defensive tactics to maximize their engagement — boss battles can also be an opportunity to spotlight a specific player tool or ability as the key to defeating them, a natural extension of toolbox design. Many of the most memorable boss battles are themed around a particular game mechanic. This gives the designer another opportunity to ensure they’re squeezing every good idea out of their systems, adds variety, and forces players to deep-dive into abilities they might be neglecting.
The Legend of Zelda series strictly follows this approach in the majority of their games. If their dungeons are the classroom where players get to grips with new mechanics and concepts, then Zelda boss battles are essentially the “pop quiz” at the end.
Let’s say you just got the boomerang. This will almost surely be shortly followed by a dungeon themed around the boomerang. Soon after, it’s boss fight time, and how do you beat the boss? Well, you just spent the last few hours using the boomerang and learning it’s various advantages and applications in puzzles and level design, so your guess is as good as mine. ..assuming that you guessed boomerang. But just like how math papers always phrase their “problems” as obscure metaphors about train arrival times and dynamic apple exchange, it’s not as simple as just doing what you have done before in an exam environment. The core rules of how the mechanic has worked before remain the same, but a boss battle should force it to be applied to a more complex and abstract situation.
In Legend of Zelda: Ocarina of Time the boomerang themed dungeon is Jabu-Jabu’s belly, and the pop-quiz is hosted by “Bio-electric Anemone” and “Electric Rotating Intestinal Worm”, Barinade. Let’s call him Barry for short. The boomerang is thus far known for it’s ability to hit things that are out of reach, and stun enemies. The battle firstly calls on the boomerang to sever tentacles which are holding Barry to the roof of the room, a static out-of-reach target as previously experienced, albeit minus the usual visual cues and in a new topical environment for added satisfaction. That was just to get you in the boomerang-ing mood. The player must then carefully time their boomerang throws between spinning electrical jellyfish things rather than simply throwing it willy-nilly-y. Timed throws, check. Barry then takes off around the room and must be stunned with the boomerang to reveal a weak-spot. -Moving target, check- which the boomerang can hit.. assuming the player can strafe Link around the fuzzy electricity texture planes Barry is throwing at you. Timed throws AND a moving target, at the same time!
This battle is both the culmination of a classic Nintendo 4-step-level design and the final test of the boomerang dungeon, and once it’s completed the player has proven themselves as a fully certified and card-carrying boomerang expert. Well, boomerang-carrying.
Although Nintendo have the most overt examples of this boss battle style, it can be identified in many other games.
In many of these case studies, the game forces the player to do what it wants by subtracting the other options. This is acceptable assuming the subtraction doesn’t feel contrived and the remaining options are enjoyable. Granting immunity to some or all of the players attacks is also a popular answer to a question all boss designers must ask: What if the player does EVERYTHING?
When the player strikes an enemy, it is tradition for the enemy to recoil in pain. This is called a “hit stun” and usually prevents the enemy from doing their own attacks. Without a way to escape hit stun, enemies can be “stun locked” by a player simply bombarding them with maximum aggression. This is acceptable for fodder and lower-level bads who can find time to do their thing while the player is focusing elsewhere in the scrum, but it’s too embarrassing for elites and big kahuna’s who face the player without distractions.
If you are chasing engagement as a goal then thoughtless mashing is top of the list of things you don’t want the player to be doing and closing that door is the first step to pushing the player back towards your intended Fun Zone. These enemies either need a way to escape the players onslaught or to control the players opportunities to deal damage.
The easiest (and thus rarely the best) solutions are to keep the boss out of reach or in possession of a standard immunity to the players attacks, either for certain preset sections of the battle, or until the player succeeds in completing a predetermined task or puzzle to lower that immunity/get in reach.
Another classic defense mechanism for bosses is the “weakspot”. If this spot is only available at controlled intervals or after puzzles, it is basically the same as the out of reach/immunity approach, but some bosses wear their weakspots loud and proud the whole battle. The fact it is slightly trickier to hit the bulls eye than the dartboard is sometimes just enough increased challenge to stop the player spamming out damage and draining the boss bar too quickly, particularly if that spot is hidden on the bosses back or lesser spotted areas. This system often links well with another design tactic in boss battles, that of the boss opening itself up to damage through some of its own attacks. Just as each boss attack must be designed so it can be avoided, certain moves may leave the bosses defenses down or “unintentionally” reveal that tell-tale weakspot. Ideally if this is the method you choose, more than one of the bosses moves should be open to reprisals, or the player will simply find themselves waiting for that move to come up, and “waiting” is rarely one of the things you want the player to be doing.
The most sophisticated bosses are open to damage throughout the whole battle but have techniques to escape stun lock and punish the player for being thoughtlessly aggressive, usually a “get away” move such as a teleport or jump, or a “keep away” move like a short-range AOE. These are triggered when the player overstays their welcome in the bosses personal space, with the latter normally prefaced by either an auto-block or temporary “super armor” state. Super Armor is when an enemy is still taking damage but will no longer react, which means their attacks cannot be interrupted. Some bosses walk around wearing it all the time.
If you ever wondered why so many bosses play by the same rules, it’s because they’re trying to solve the same problems. The designers most challenging boss battle is balancing convenience with contrivance.
Too Big Too Bad
Enemy and Boss design is a big juicy topic that shares so much real-estate with other big juicy topics like combat systems and level design that it’s impossible to wrangle into a single thesis. This article runs the gamut from macro to micro in an attempt to show enemy design shouldn’t happen in a vacuum, enemies and bosses should be conceived and designed to serve a purpose. Their attacks, their weaknesses and their defenses are all ways to control and define the way your game is played.
What do you want the player to do?