Designing Enemies

Almost every move you can think of has already been created.

Do not be ashamed to use the ideas of other designers, as it is no more offensive to use a pre-constructed library in Java, than it is to take the ideas from another game and reapply them. The ideas don`t make the designer, execution does. Throw off these chains of constantly reinventing the wheel and get to stealing borrowing mining. The mining process begins at the greatest source. At the fount of cool, flashy and powerful moves: Marvel vs Capcom 2.

The year is 1998 and Marvel vs Capcom explodes into arcades. Taking the ideas learned from the previous VS titles, it expands them to both the Marvel and Capcom casts. But two years later its follow up, Marvel Vs Capcom 2 (MvC2), was the pièce de résistance. Do the math and you quickly see why this game is a treasure trove of ideas. 56 characters in total, each with several normal attacks, air attacks, special moves, assist moves, hyper moves… it quickly adds up.

The quantity of moves is not the only reason MvC2 is an inspiration. The design philosophy of this game, of the VS series in general, is that if you `break` everyone, everyone is balanced. Too often, when things seem overpowered, our instincts tell us to swing that nerfbat like the game owes us money, but if you are going to make a move impotent, then why even have the move? For each character the designers strove to give them a “trick” — something fun to give them style. Here emerges the beauty and joy of this game. Around every corner is some new eye-exploding, opponent-smashing, or health-restoring nugget that is yours to discover, and there is a lot to discover. Yes, many moves are character-themed, subtly-twisted derivatives of other moves, but it still leaves us with a large amount of ideas to mine both thematically and functionally.

Having a cast of 55 characters means you have a great collage of themes to pull from: plant guys, giant robots, samurai, cyborgs, ninjas, demons, hulks, egyptian gods, juggernauts… the list goes on. When stuck thematically, this game can be an inspirational source. Even if you have a character that doesn`t fit a mold cast by MvC2, the level of creativity can be helpful to get ideas popping.

As a thematic tool MvC2 is great, but what is more beneficial is its help in defining functionality. When designing enemies your goal is to make sure that every guy gets his own little trick. The trick can be something as simple as the pest enemy you use to swarm the player (emphasizing your big area attacks), or it can be something as special case as the medusa from god of war (emphasizing movement and timing your attacks).

Knowing a good source of ideas is only the first step. Next, you must understand how to deconstruct the game, because knowing how the moves are constructed gives you a fighting game lexicon. Then, you must understand why and how you reconstruct them. Just taking random moves and smashing them — square peg round hole — into your cast of characters defeats the purpose of this exercise. Additionally, if you are making a single player action adventure game, there are certain philosophies that are incongruous with a fighting game. Your cast has roles to fulfill, so you will need to know what moves help you define these roles, how to modify these moves so they are compatible with the game you are making, and how to maintain a balance in your play.

Deconstruction – Action Reaction Result

Now we are going to get into the nitty-gritty fun of breaking down moves by understanding not only what you are seeing, but also why you are stealing it.

Just to be clear: my vocabulary is NOT congruous with the MvC2 community. These are the terms I was taught while working on god of war, which I have adapted and applied to the game of MvC2. I do this, number one, because it is easier for me to think in terms of god of war terminology, and number two, I do it to reinforce the idea that this is not “about” MvC2. The purpose is to talk about how to effectively construct ideas for use in an action adventure game, with MvC2 being the greatest source.

All moves are broken down into three stages: Action, Reaction, and Result. The player performs some action, his opponent reacts, and it results in both of their status changing. Some moves involve all three stages (throwing a fireball), while others might only involve one (teleporting), and yes, sometimes it can be hard to fit moves into any mold. Do not lose heart but instead remember: eventually you are going to reconstruct these ideas into your own moves. A structured approach helps in both construction and balance; additionally, this way of thinking reminds you of the important questions to ask: what does it look like, what does it do, how do I react, why should I use it, and how can I counter it?

Action

Every move your character makes is some kind of action. The first step in deconstructing the actions you see in MvC2, or any game, is by understanding their basic structure. A move is made up of frames of animation, and these frames are broken down into three stages: Startup, Hit, and Recovery.

Startup
The move is starting, but nothing has happened yet. Also referred to as Anticipation. The longer this period last, the more time your opponent has to figure out what move you are doing and react. When someone says a move is ”telegraphed”, they are likely referring to the fact that a move had a long and obvious startup.

Frame 5: Readies the kick

Hit
The move is now “active”. If it was a punch and it was touching someone else, this would be the part where they take damage and react to it. The longer this period lasts, the easier it is to hit someone. At first glance it appears that longer is better, however the longer your hit frames are active the more time your character spends locked into this move. In Street Fighter 2, M. Bison`s crouching heavy kick (c.HK) makes him slide along the floor a decent distance. His hit frames encompass the entire slide, which makes it a risky move. If the other player blocks the move, since you are still locked into the slide, they can easily counter attack.

Frame 8: Extends and connects

Recovery
The move is no longer active, but I must spend time recovering back to my “rest” state before I can perform another move. If I was punching you this would be the time it takes me to bring my hand back. If I really put my all into this punch, it could also be the time I spend trying to catch my breath. The longer the recovery, like having a large number of hit frames, means that you are locked into this move, or”committed”. A particularly powerful move might be balanced by having a long recovery — high risk high reward.

Frame 12: Brings the leg back

Fighting games, action adventure games, beat em ups… they all make it or break it by the strength of their melee moves. Do they feel powerful, do they flow, and do they make sense. The majority of moves in MvC2 — in fighting games in general — fall into the category of melee moves. From quick jabs that let you poke your opponent to death, to arm-stretching overhead smashes that flatten them to the ground; the number of ways you can slap some part of your body against someone else is as varied as our imaginations. This variety comes from the interplay between three attributes: Power, Range, and Speed. These attributes are a sliding scale and not a harsh choice: imagine dials not check boxes.

Understanding how to adjust these dials requires a firm grasp of our three stages of a move. You can have the most powerful move in the game and I the weakest, but if my move has fewer start up frames I will beat you all day every day — speed defeats power. While we could fix this by giving that powerful move greater range, the better solutions present themselves when we have a robust understanding of reactions, and how different kinds of reactions (or a lack of them) can be used to balance moves.

Reaction

Actions are the fun part — the part we fixate on — but they are not the most important part. A powerful move feels like a powerful move only when it is sold by a powerful looking reaction. Understanding how and why characters react to various moves is integral in making sure your moves don`t come off feeling underwhelming, or even worse, comical.

As you study moves, observe the difference between an opponents normal reaction and his reaction when guarding (blocking). When you are hit by a move, much like when you perform an action, you play an animation, and your character is locked into this reaction for a period of time. The reaction animation you play is visually different when you are guarding, but more importantly, the animation you are forced to play is shorter (fewer frames). The variance in length between taking a hit and guarding a hit can mean the difference between landing a combo, or getting brutally punished. A normal reaction means that you are stuck in Hit Recovery (also called Hit Stun). If you are guarding, you are stuck in Guard Recovery (also called Block Stun). Note: I do not like the terms stun or block. The former because you are not actually stunned, and the latter because blocking implies a negation of all damage, while guarding implies a reduction in damage, which is more accurate.

Let`s say I have a punch with 2 hit frames, 4 frames of Recovery, and a kick with 3 frames of Startup. The punch causes 10 frames of Hit Recovery but only 5 frames of Guard Recovery. If I attack someone with this punch (and they do not guard it), I can link my kick and get a free hit (my Hit, Recovery and Startup are less than his Hit Recovery). If they guard my punch, however, he will recover before I do. This is just a simple example, but you can see the importance in understanding the difference.

The attacker recovers first

The defender recovers first

This is just the beginning of what you can do when you begin to understand reactions. Return to the example above, with the ridiculously powerful but slow move and the weak but very quick move. The quick move was going to have the advantage, but what if we can introduce a new trick to the slow move? Some moves allow the player to have armor or ”tank” through attacks. While tanking you take damage, but do not react. This gives you a clear advantage. The quick move will hit you, and you will take damage, but it`s no matter because you plow right on through and smash that punk to bits — huzzah! Tanking is an example of a modifier that we can place on a move, and it`s not the only one you can use. Note: even though modifiers are an attribute of the action, I feel it`s important to talk about them here in reactions, because they deal almost exclusively with how you react (or don`t react) to moves. There are quite a few modifiers, and again, these are terms I learned from working on god of war, but you can see most of them in some form in MvC2:

Tank
You take damage, but do not play reactions. Sometimes a move can only tank a fixed number of hits. For example, you can tank the first hit, but if you are hit a second time you play a reaction normally.
Invulnerable
You take no damage and do not play reactions.
Guard Break
Attempting to block a move with this attribute forces you to play a special kind of reaction. You do not take damage, but you are left vulnerable.
Unblockable
Attempting to block a move with this modifier forces you to play a normal reaction as if you did not block it at all.
Crumple
A special kind of reaction. Your character loses control for a period of time. Being hit while crumpled causes you to play a normal reaction. Note: THIS IS NOT THE SAME AS BEING STUNNED. When you are stunned and receive a hit, you remain stunned until a period of time has passed. The Crumple state, however, ends once you are hit.
Frozen
Turned to ice, turned to stone; theme is irrelevant, the reaction is the same. You are locked down, stunned, and you cannot break free until a timer runs down. Note: this is the first time I have used the word stunned. That is because this is the first time it has been factually correct. Do not use it when not appropriate. It`s confusing, and only serves to weaken communication and meaning.

With all of the various reactions (normal hits, guarded hits, air hits, crumples, etc.) and with the knowledge that to truly sell an attack it has to have an appropriate reaction, you can see how the workload begins to balloon when left unchecked.

Note: If you don`t have time to make heavy reactions, all that time and effort making cool heavy looking attacks is now worthless.

What separates the good from the great is having a truly competent grasp of how important reactions are to the system; not just how to make them look good, but how to make them look good within a budget. Your workload dictates the possible reactions, the reactions dictate the feel of your attacks, and the feel dictates the final piece, your results.

Result

Reactions seem to be results, but there is a worthwhile distinction here. First, a move can have a result without a reaction (I drink a health potion). Second, and more important, reactions are animations, while results are changes in state.

State, in this context, refers to more than just the health of your character. Other changes in state range from altering your current armor rating, to altering whether you are considered on the ground or in the air. For clarity I will break them down into categories, and then provide some examples within each category to illustrate its uses.

Health

  • Damage: A fixed reduction in health. The amount of damage done is scaled by two factors: the opponents armor class, and the current combo count. A character like Sentinel takes less damage from hits than someone like Akuma because they have different armor class ratings, additionally, the more you hit someone, the more your damage is scaled down (eventually reducing to 1 damage per hit).
  • Poison: A fixed reduction in health over time.
  • Heal: A fixed increase in health
  • Regen: A fixed increase in health over time
  • Drain: Reduces health from one person, healing someone else for some fraction of damage done.

Defense

  • Armor: Increases or decreases the amount of damage scaled by your armor.

Time

  • Increase: Making one character move faster than another is a simple to understand, if complex to implement, example of this concept, but a better illustration is to consider what happens if we simply reduce the number of startup frames a character plays.
  • Decrease: Much the same, but in reverse. I could increase the length of my opponents recovery, giving me a major advantage.

Position

  • Impulse: Not all moves shove you across the screen, some suck you in. Impulse is the generic attribute, applied programmatically, of how much (both positive or negative) something shoves you.
  • Launch: Knocking an enemy into the air places him into a completely new state. The game must understand not only that he acts differently, but also that he reacts differently. There are a whole new set of reactions to play when you are in the air; remember what I said about your workload ballooning when you don`t pay attention?

Results are easy to grasp and fun to think about, but do not get carried away. Their simplicity in concept belies their complexity of implementation. For every system you wish to modulate, the programmers must expend time and effort making that system variable instead of fixed. Pick your battles carefully.

Action, Reaction, and Result. These stages help us answer three questions: what does it look like, how do I react to it, and what does it do? We have yet to learn why we should use it and how to counter it, but these questions are answered by taking the deconstructed parts and ideas and reconstructing them into a coherent package.

Reconstructing – Roles, Classes, and Tricks

Our goal until now has been the break down and analysis of MvC2`s combat system, but now we are going to switch gears. Applying these ideas to an action adventure game is not like designing characters for a fighting game, just as designing cars for Gran Turismo is not like designing cars for Burnout; different styles of games have different design goals.

One of the major goals in a fighting game is to minimize your barriers to entry and maximize the potential for mind games. This is achieved through minimizing depth and maximizing breadth; by this I mean not that a character takes no skill to master, but that (in successful games) the mastery is applicable to numerous characters in the cast — I learn to Dragon Punch, I also learn to Tiger Knee. The challenge comes in judging and anticipating the moves of your opponent, not in executing your own.

Action adventure games, however, have a different goal: Accomplishment. This requires a depth of player and a breadth of cast. The player must grow and have a sense of depth, which applies to more than just his attacks. Every system the player learns (magic, climbing walls, world interactions) is a new system to master, so do not go crazy with the player`s attacks. Avoid superfluous attacks. This “back of the box” design has no place in compelling games. The player doesn`t need breadth, as he has no one to play mind games with. Who is he going to fool? He requires only the moves that get the job done, in the coolest way possible. Again, this is why MvC2 is so great, because everyone in that game is doing wacky, powerful shit all over the place. One can imagine the designers on that game standing on a box shouting, “Fuck subtlety!”, and I say Amen.

Zero subtlety

The tight, focused and subtlety-free design of our player also applies to our enemies. We want breadth, but we want it to come from the cast as a whole and not the individual enemies. The monsters are not there to play mind games with the player, they are there to make me feel like a bad ass. I want to smash faces. This does not mean that your enemies cannot be tricky — they can and should have a trick. That is trick, in the singular. The player learns the monster`s trick, and once overcome, gives her a sense of accomplishment. She is getting stronger and better than her adversaries.

Accomplishment is a major driving force behind the best Action Adventure games, so if you want players feeling accomplished and not asking what idiot ruined their night, there are three things to keep in mind. First, you must always keep in the mind the roles your cast must fill. Second, you must type your cast in a balanced manner. Third, you must find your trick.

Roles

When discussing roles it is wasteful to speak in specifics. The mechanics of your game, which are unique to each game, define the roles for your enemies. In God of War, for example, the Gorgon`s beam attack is countered by using Kratos`s dodge, so her Role is to emphasize the dodging mechanic; a Role that makes her compelling in tight spaces or against heavily armored enemies that require commitment. With no dodge mechanic, the Gorgon serves little purpose, which leaves her boring and frustrating. We must instead view our cast from a higher level. Doing so forms a picture, independent of the mechanics, of four major reoccurring Roles: Emphasizers, Enforcers, Smashers and Challengers.

Emphasizers

Emphasizers should, when striving for mass market appeal, make up the majority of your cast. As stated, your player has several tools in her toolbox for dispatching the creatures she faces. An emphasizer is a creature that rewards, not requires, the use of one specific mechanic. The key is to have positive reinforcement. You are not penalized for using the wrong method, simply rewarded for doing it the ”correct” way. God of War has several examples of emphasizers, but I particularly like the Gorgon.

The Gorgon is considered a higher tier creature, in that she poses not only a greater challenge to the player but also requires more work to implement and create. She is fast, avoids your attacks, and has her deadly gaze, which if the player stays in too long, will turn Kratos to stone. Behind the scenes, when the player is ”inside” the cone of the Gorgon`s gaze the game starts a timer, and if the timer ever reaches a certain point Kratos is turned to stone; however, if the player ever uses his dodge (roll) the timer is reset, regardless of whether Kratos is still in the gaze. Dodging (rolling) gives the player an advantage against this creature, but Kratos is not required to do this. For example, if the gorgon is hitting me with her gaze and I attack the gorgon with a significantly heavy enough attack she will stop her gaze and react to my attack. They key here is ”significantly heavy enough” attack. Not just any attack will be effective. Only Kratos`s heaviest ”ender” attacks work, but it still gives the player a choice and that choice is what makes her a compelling monster.

Let`s say I have started to attack a creature and at the same exact time the Gorgon has started up her Gaze attack. I must now make a choice, do I continue with my combo in the hope that I will have time to hit the Gorgon with my heavy attack (risky), or do I forgo the damage I could do and escape (safe). This meaningful choice comes from her emphasizing mechanics and not enforcing mechanics. Very rarely, if ever, should enemies be constructed with the purpose of requiring a specific action from the player. Emphasis rewards experimentation, while requirements brutally teach. Over time the player discovers the best course of action. Enforcement, on the other hand, brings the game to a screeching halt, while the player must deign the correct solution. There are times when you should enforce a mechanic, and those times are when you want to teach your player something core to the game.

Enforcers

The majority of your cast should be emphasizers, but they do have a downside. By allowing for multiple killing solutions they lack the ability to forcibly teach the player. Sometimes you have a mechanic that is so core to your design that a player must learn it. Enter the enforcers. These creatures require a mechanic from a player, and without the player using it, they will not progress.

The Automaton forces Kratos to use his Fire Blades

The shield bearing enemies from God of War are the perfect example of Enforcers. You cannot grab them, you cannot hit them with light attacks, and the only way to get passed their shields is to hit them with a “significantly heavy enough” attack. These moves are known as his “crush” moves, and are made up of all the heavy ending moves of Kratos`s combos. (Square, Square, Triangle) — (Square, Square, Square, Square, Square, Triangle). These moves are so core to God of War combat that the game wants to make sure the player understands their power. Enforcers have their place, but their place should be rare. Use this only when something MUST be taught to the player, and do so with as much obvious flair as possible; this is not a time to be subtle. Subtly and enforcers do not mix well, and if handled poorly can quickly become frustrating.

Challengers

These are the bosses, the difficult enemies, the ones you use all of the tools in your toolbox to defeat. Challengers are separated from your emphasizers and enforcers due to their complexity of design and cost to implement. Where most enemies in your cast will have their one or at most two attacks, a challenger might have three or four.

These enemies control the flow of your game by inserting spikes in the pacing, and a well paced game follows a sinusoidal curve of difficulty. It starts the player off with simple tasks, increases the challenge gradually up to a climax, and then quickly eases back to let you cool down. Where the challengers control the mountains of our pacing, the final group makes up the valleys.

Smashers

Providing the player with challenges is important, but sometimes you just want to destroy something — a lot of somethings. Smashers are the smaller, weaker and easily dispatched enemies that you throw at the player, sometimes alone, but most times accompanying other enemies, and they let you get your smash on. Remember our number one goal in these games is to give the player a sense of accomplishment, and letting your mass murderer out of the bag every once in a while feels really good. Where Challengers strive for difficulty, Smashers strive for simplicity.

 

Note: this simplicity should be applied not only to their difficulty, but also to their implementation. The Smashers serve a secondary function. By making them lightweight in terms of AI, poly count, texture size, and basic memory footprint, you can sprinkle these enemies liberally across the game.

 

Two groups (Epmhasizers and Enforcers) are defined through the player`s mechanics. Emphasizers, the most common, reward the player for using a specific mechanic without requiring it, while enforcers teach a specific mechanic to the player. The remaining groups (Smashers and Challengers) are defined by difficulty and implementation, which again is independent of the specific mechanics. Smashers are weak and easily dispatched, while Challengers pose a great challenge to the player. These four role archetypes are important, but they are not enough if we want to have a balanced cast that is diverse in play style. For that we need to group our creatures by additional criteria.

Classes

A balanced cast is one diverse in style, complexity, and mechanic. In the old days, it was as simple and clean as a “skinny”, “medium”, and “fat” hockey player. In God of War they break enemies down into several classes, which ensures clarity in discussion, ensures designers (and others) visualize their workload, and ensures there is a properly distributed variation in play. Classes range from “Pests” which contains all of the lightweight annoying enemies, all the way to “Bosses” which, obviously, tracks the bosses.

Clear and distinct classes are necessary for meaningful discussion, and sticking to it lets other departments quickly understand not only what you want, but also how much work it creates (or saves if it`s cut). Visualizing your workload, while important in the beginning, is just as critical near the end when things are being cut. Your cuts need to save time and they need to leave the game without a gap in its play.

Classes can also help you to keep track of your play variation. In MvC2 you can break all characters down into three classes: Huge (Sentinel), Quick (Spiderman), and Ranged (Cable). Huge counters quick, which counters ranged, which counters huge — rock, paper, scissors. Keeping these three classes even maintains a feeling of balance, but say I need to cut three characters out of the game. Without class structure, I could unknowingly choose three characters out of the Huge group and, with fewer counters to the quick characters, leave the game feeling broken and unbalanced.

A full breakdown of the cast of Chains of Olympus

Roles and Classes, while appearing redundant, serve two distinct goals for the game. Roles are about the the player, while Classes are about your cast. The mechanics of your player demands enemies that work well against and with him, but only viewing your cast through that particular lens leads you to end up with lots of functionally distinct but visually similar enemies. The ultimate goal is a diverse cast, and having Classes helps to ensure visual diversity.

Tricks

Everything must serve a purpose. If you cannot explain to me why a creature must exist in less than two sentences then it`s a safe bet the player is not going to enjoy killing it. A great designer I know always asks, “What`s this guy`s trick?” The question has stuck with me, and it`s a great question to ask of your enemies. A trick implies that you can fool me, but once I know your trick it loses its effectiveness. Being fooled keeps you on your toes, while providing you with the opportunity to learn and grow as a player, which leads to our ultimate goal of accomplishment.

And so I return, one last time, to MvC2. If ever there was a game that could be called a Bag of Tricks, it is this game, but there are even more tricks out there. Go, mine, study what they have done, and if you choose to take away only one thing, let it be this: don’t be afraid to be inspired by others.

Every move you can think of has already been created by someone. Mining their ideas doesn`t make you a bad designer, because, believe it or not, taking that idea and executing it to the same (if not higher) quality is hard. If you are smart, you will choose to devote all your limited time and dwindling energy on executing to the fullest of your extent.