The Three Challenges
The real-time aspect of an action game increases the frequency at which the player has to evaluate some parameters of a situation, and then decide which ability to use.
Evaluate the distance
- Evaluate the distance to an enemy in order to choose which ability to perform
- Evaluate where your character will end up once the ability is performed
Because everything moves in real time
Evaluate the time
- Evaluate the time it takes to perform an ability
- Anticipate how much time your ability will last
Because everything evolves in time
Cleverness and anticipation
- Anticipate a combination of actions to perform in different situations
- Know which ability to use to counter the attack of an enemy
Because each ability is a tactical tool for the player
In parallel, there are many advantages to designing enemies which match the player's abilities. Indeed, one of the main purposes of adding enemies to a game is to create an interesting challenge to teach the player, little by little, the mechanics of the game.
An Enemy is a Challenge For the Player
When we have to design an AI behavior, a common tendency is to try to make the AI seem as clever as possible and feel human. For example, to create an immersive experience, we might want to have multiple enemies behave as a squad, or perform actions the player would interpret as clever ones.
However, we tend to forget that no matter how clever an AI enemy behaves, the challenge created by an enemy when the player encounters him is what will impact the game experience the most. This means that as designers we need to go back to what is fundamental in designing a challenge, and understand precisely what are the most important properties of an AI to offer the right challenge we are aiming for.
Define a precise challenge for each enemy. Because the main function of an enemy is to attack the player, and because most of the time the player can destroy the enemies, there are usually two important questions I ask myself when designing the behavior of an enemy:
- How can the player attack and destroy an enemy?
- How can the player defend himself against an enemy?
Indeed, when we design an entire panel of enemies, we usually try to create very diverse ways to beat each of the enemies.Here are two examples from Spider-Man: Shattered Dimensions, to illustrate this point:


As shown by those schematics, the ninja and the kamikaze bomber each offer a different challenge to the player when he has to defend himself against them:
The best ability of the player to defend himself against the Ninja is the dodge ability.
The challenge of the player to defend himself against the Ninja is a timing challenge.
The best ability of the player to defend himself against the Kamikaze bomber is the web shot.
The challenge of the player to defend himself against the Kamikaze bomber is a distance evaluation challenge.
Challenge the Player's Abilities
As a complement to designing a different challenge for each enemy, it's also interesting to design the enemies with different values of weakness for each ability/weapon. In other words, it is interesting to design each weapon to be more or less efficient against each enemy.
- This pushes the player to use all the abilities he has?- This helps the player to learn the specifics of each weapon?- This also pushes the player to be tactical and to use the right ability at the right time.
To illustrate this point, here's an example analysis of weapon efficiency on the enemies in Halo:

Because of these properties, during a combat against multiple types of enemies the player might switch from one ability to another in order to be more efficient in battle.
Common enemy archetypes in action games
Since everyone is attempting to have enemies that match with the player's abilities, we often see the same kinds of enemies in many different action games. Indeed, there are a lot of advantages to use these common archetypes:
- they offer an easily understandable challenge
- they have a recognizable and dissimilar shape
- the player may already have an idea of how to deal with them without the need of a tutorial.
Here's a list of the the most common archetypes in action games, with the functionality they usually have in the game.
- Shield enemy: challenge the precision of the basic attack



- Bomber: close combat attack to counter

|
Write this book! =)
It mostly seems to focus on position, esp. distance, as the determining factor. I think that needs to be extended to other elements. For example, many games feature some sort of "elemental alignment" or "color", ie. fire weapons work better against ice enemies, AP ammunition works better against armored enemies, etc. This can also be a major determining factor in which weapon to use - even the primary factor, in many games that include such elements.
Thanks for your interesting comment.
I agree with you, I didn't talk at all about element alignement as you said. Maybe I should have insisted on it but it's true that I've focused myself on real time combat system in this article, and what are the major parameter to play wiht in a real time combat system. Elemental or colour don't evolve in real time, distance do. And timing evaluation is inherent to real time combat.
In short, if we frame elemental-style design again in the context of some of Sebastien's original core ideas - risk vs reward, offering different varieties of problems/challenges, effectiveness of weapons vs types of enemies - you can use a system to magnify those elements in your game design if they need more emphasis.
I was playing Halo 1 recently, and I think it makes a good example here. The Shotgun is excellent against the Flood, not so much against other enemies. Most other weapons are weak against the Flood. Covenant weapons are better against enemies with shields (the tougher Covenant ones). But these weapons have other tradeoffs too, which usually dominate decisions about which one to carry and which one to use. E.g. sniper rifle is used for distant head shots, rocket launcher doesn't have much ammo, Covenant weapons overheat and can't be reloaded, etc. Even though some enemies seem to have a specific "elemental" bias (e.g. weak against plasma weapons, or weak against shotgun) its not a simple binary decision. Later levels contain a mix of Covenant and the Flood, so choice about which weapon to use in each section will depend on things like, what kind of cover is available, whether they are open spaces or enclosed hallways, etc.
I'll just add that, in most AAA titles, most players don't even experience all that depth. I played through Arkham Asylum without really understanding the melee combat system. Bit of a shame...but I guess for the players that really get into it, like my friend who really got into the Challenge Rooms, they appreciate it.
One thing I notice, though, is something that seems pretty common to combat AI in most games. Namely, that "tactics" is being defined here as really applying only to personal combat with internal resources. But tactics also refers to the smart use of external environmental elements to obtain and exploit a force advantage over an adversary.
This is mentioned indirectly by describing a "tendency" (presumably of novice combat AI designers) to want to allow multiple mobs to function as a squad, apparently just to make a game feel more "immersive" or to make enemies "feel human." It's a fair point that a small set of interesting choices is better than a big vocabulary of indistinguishable actions. But I think dismissing environmental tactical choices as mere "immersion" would miss some opportunities for offering more enjoyable combat play.
In addition to tactics enabled by internal resources -- e.g., HP/mana, cooldown timers, recovery delays, shields, movement speed -- environmental features can be resources as well. Line of sight may be the most common environmental phenomenon that characters can be programmed to respect, creating opportunities for movement and stealth tactics. But there are plenty of other phenomena that also allow tactically interesting character actions and choices: day/night cycles, rain/snow/fog, smoke, entrenchment (digging into the terrain), smell (what if you want to sneak past guard animals?), sound (the Muffle spell in Skyrim), camouflage, heat, pressure, radiation, mass, and so on.
Think about the Thief games. Both light and sound were functional environmental elements of combat AI design. By allowing characters (player and NPCs) to be aware of and make choices about these environmental phenomena, the range of interesting player choices possible -- and the coolness of what NPCs might do because they could detect those phenomena and make decisions regarding them -- allowed a truly satisfying game to emerge. Tactics weren't just about my internals versus your internals; the world itself mattered.
Not every game needs exactly those features. The point is that external environmental resources can also be valuable in tactical combat AI design because they enable more interesting choices. They're not just for immersion.
Thanks!
And may I add, that my instructor in gamedesign thaught me...that the information regarding distance, position of your gear, bombs legs and arms...are game narrative.
Though he used tetris as an example.
Its an valuable pillar, because the player needs input/exposition on how he should act to overcome challenges.
no
NO
NO!
Change "Deep gameplay " to "thrilling", "emotional", "immersive" ANYTHING!
But deep Deep gameplay will always refer to length and breath of the mechanics, and though yes..can take 20-30 minutes to master...it will offer a MUCH bigger symphony of stimulating experience then what angrybird EVER could.
If someone can't, direct them to a different genre.
The article refers to the simplicity of the inputs required from the player, and I don't think that complexifying the inputs would necessarly make the gameplay deeper.
Take the simple Reloading mechanic in a competitive FPS, which is normally one button press. The player still have to figure out when and where it is optimal to do so. It refers to his map knowledge, his prediction of enemy movement, his analysis of the situation, and more... Even with only one single button press, you can expect a new player to spend quite a while to master it perfectly.
I do not think that you have to ask the player to "do a Shoryuken to reload" to add a depth to that mechanic.
I like how you quoted halo a lot in this article, as it is one of the few shooters that really has satisfying open battle ( aka not purely scripted) combat, a lot of people attributed this to the AI. which partially is true.
however the enemy design is so clever and the enemy hierarchy so well constructed that the combat feels like a tactical game (also 30 seconds of fun or The Illusion of Intelligence by bungie is a great resource)
I also think that this is one of the reasons I did not enjoy halo 2 as much as I did one. I loved the E3 trailer, it showed complex situations and interactions and Cortana gave you warnings which was really amazing, but in the game when the Elites were replaced with the Brutes a lot of the tactics got lost.
the brutes did not have a very interesting attack pattern (imho), if the damage received was too high, charge.
which was also very annotying to me as a player. for me the most fun I had in halo is when I was able to really set up a fight, example in the second level when you have to resque the marines there are a few great moments, where you take over a position, and then have to defend it with any means,
usually a quite clear way you can do this is defined (sniper rifle is handed off or happens to be there) but you can ignore i and through clever warthog placement get very different fights.
also the one encounter by the cliff gave me a lot of fun, you ahve to extract marines from a underground structure, and placing my warthog intelligently in that fight really helps.
again it is almost a tactic fight not so much a scripted event or a rollercoaster ride, you ahve to plan your fights and evaluate on a per case basis how to engage you enemies, and there are many possible answers.
again thank you for the article.
Other than that one mistake in your diagram this is an excellent article. Hope to read more.