#combat design

Two to five minutes for a single battle is actually a really long time. Most turn-based combats in games take seconds, not minutes, to complete and that's on purpose - it helps keep the combat flow from getting old. I suggest you do some more technical research on your subject. I suspect you'll find your results surprising. Here's what I mean.

What I think you should do is record yourself while you play through each fight in some of the turn-based combat games you think have good flowing combat (e.g. Final Fantasy 6, Chrono Trigger, Expedition 33, whatever) using capture software like OBS studio or even just with your phone while playing on your console. When you're finished, write down how long the fight took from beginning to end. Write down how many actions you took and choices you made. Write down how often the enemies got a chance to attack and what they did. Write down how much damage you dealt and how much you took. Write down as much as you can about the numbers, the timing, and what happened. Break the entire combat encounter down as systematically as you can. Repeat this process for several battles so you have data across multiple fights for multiple games. How does a boss battle differ from a regular battle? Does a scripted battle differ from a boss battle? Analyze them too. Write this all down and then look for commonalities across the different types of battles and games.

Repeat this process for games that don't feel as good, where it started to get old for you. Record all those data, then start comparing. How are the less good fights? What's the time per battle differential? Are you making similar choices or different ones? Try to figure out what technically makes it feel "old".

This is actual video game design research. You need real data to make your comparisons, not just some vague ideas about how things work. You need real numbers to model your own systems after, not just assumptions. Take the games you're researching and actually break them down with real data, and then use that real data to draw your conclusions. The results may surprise you.

[Join us on Discord] and/or [Support us on Patreon]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Short questions: Ask a Game Dev on BlueSky

Long questions: Ask a Game Dev on Tumblr

Found this old GDC talk id been meaning to look up. A while back, I joked that most combat-heavy videogames are basically just Skylanders, and a couple people, both of them Skylanders fans, seemed a little incredulous at me. Well, I was referencing this talk, so hopefully if you watch it you'll understand where im coming from. If not, hopefully you'll enjoy learning a little bit more about how one of your favorite games was made.

(this showed up when looking for a Sonic Frontiers gif)

A couple of combat design concepts for my portfolio! I think that every fighting game should have a really powerful old man actually. ID under cut

I recently was absorbed into Person 4 Golden now that it got released on steam. Since the game is a few years old at this point I’d normally say that this is ground covered by people better and smarter than myself and not write about it, but when fighting the first boss a moment occurred that completely changed my perspective on video game bosses as a whole and that revelation is just too intense not to write about! I’m of course talking about the moment Shadow Yukiko tries to summon her prince charming for a second time and the spell fails. I want to take a look at why a boss trying and failing an attack is probably the single most amazing synthesis of story and gameplay I’ve seen and why it tells you more about that bosses character than anything else could.

For those of you who haven’t played Persona 4, I’ll quickly run down the setup to this boss. I wouldn’t consider it a major spoiler since it is only the first real boss of the game. Anyway enter Yukiko, a girl who is born into a family that runs a famous inn and will be forcing her to carry on the family business when she’s older. She’s not at all into this life and this manifests as the thought that she needs to find a prince to whisk her away and get her out of this life. The boss of this section of the game is her inner feelings manifested into a monstrous phoenix in a cage. Sometime into the fight this monster uses the summon ability introduced earlier via normal enemies to call forth a prince monster that looks like a puppet and which fights alongside her. However, when you defeat it or it runs away, she attempts to resummon it only for her to completely fail and we get a bit of dialogue along the lines of “Why? Why won’t he come!” After you defeat the boss and Yukiko solves her inner turmoil she learns she needs to seek independence by herself and not wait to be passively be saved by someone. 

So its pretty clear the designers have baked extremely important narrative details into this fight, but most of the time when you see this its the villain's strengths being turned into their primary abilities. Why is it so important that this boss tried and failed to do something? Well when we’re talking specifically about villains, we’re talking about people who we know have flawed views of the world. These characters are defined by their flaws first and foremost and in most media these flaws directly lead to their undoing. Voldemort in Harry Potter didn’t anticipate Harry’s mother’s love protecting him and accidentally turning him into the Horcrux that would be his undoing. President Snow expected the fear of the Hunger Games to keep people in line not lead to rebellion. The Joker in the Dark Knight didn’t anticipate the two ships not immediately blowing each other up because he believed people were inherently corruptible. Villains by there very nature are meant to have flaws that directly lead to their downfall and this is often how stories present moral lessons. 

However in games we find this far less often and its often reserved to finding the weak point is an enemy’s seemingly impenetrable armor. The twist identity of Doc Oc in the newest Spiderman game for example is one of the best interpretations of the character as a whole yet the fight devolves into you just damaging him in generic ways over and over with cutscenes showing you trying to disable his arms. His mechanical arms are certainly showcased to full effect in this fight, but not his superiority complex placing himself above people he wants to “help” in a perversion of Spiderman’s ideology. Arguably the best antagonist in Bioshock offs himself and you end up fighting against a guy who uses too siphons too much plasmids to himself to try to become super powerful. In this boss fight we do see the plasmids as the weak point you use to defeat him but only after a full phase of him wrecking house using said plasmids to good effect. In less story focused games, villains often boil down to really bad people who want power and thus it feels a bit counter-intuitive for the player to defeat them by powering up yourself (but you’re the good guy so don’t worry). Video games in general are less likely to present a boss’s failing during the battle.

When movies have villains who are defeated directly because of their flaws it provides a solid moral framework, but when video games fail to do that the moral lesson can fall flat. How do we know Doc Oc’s idea of thinking of those he saves as lesser is bad? Well he says it in a cutscene and he’s the bad guy so case closed I guess? In Yukiko’s case we know for a fact that her idea of a prince coming to save her is incorrect because we’ve observed visible effects on gameplay. This is far more impactful than if she just said it because in order to win the game we as players are hyperfocused on observing the patterns of gameplay that will lead to our victory. In other words, when something happens that helps or hurts our chances of winning that’s something we care a lot about immediately. 

The other reason this boss design is effective is because it gives us a better understanding of the character in a diegetic way which is basically a fancy way of saying its the “show don’t tell” rule applied to video game writing.

Now when making a boss fight like this you can either approach it from a bottom up or top down approach (to borrow magic the gathering design terminology). Top down would be writing the character first and then designing the fight to fit the character, but bottom up would be making the fight first and then writing the character. As always a bit of both is probably best, but lets imagine using bottom up to write a Pokemon gym leader villain. Many joke that the gym leaders are supposedly elite trainers but only use one type of Pokemon and therefore open themselves up to having clear weaknesses. This fact would actually work well if we wanted to write gym leader characters whose flaw is that they are rigorously set in their beliefs and won’t change. For instance the Fire gym leader would be always passionate and excited but not know when to calm down and rest and the Water gym leader could be always calm and not know when jumping into immediate action is necessary. In these cases, them having only one type of Pokemon is symbolic of their unchanging natures. Perhaps when you meet them later in the game they’ve learned and have other types of Pokemon?

It's generally easier to prototype and develop ranged combat that feels good than it is for melee.

Melee combat lives and dies by the animations and systems we build. If my character swings her sword in a horizontal arc, it creates very different expectation for a hit reaction than stabbing or chopping. A player will probably realize near-instantly if the reaction animation doesn't match the type of attack properly and it will make the combat feel cartoonish, breaking immersion. The melee animations are also performed extremely close to the camera, so the player gets full view of all of the details. We need many different attacks and many synchronized hit reactions in order for the melee combat to look right.

Ranged combat, on the other hand, does not usually carry this kind of action-reaction animation expectation. Shooting animations are generally simple and small motions. Playing a hit reaction can be as simple as a flinch or just falling over. Hit reactions are limited to hit location and typically don't need to take the motion of the weapon into account. Enemies that play hit reactions are also usually smaller on screen and further away, meaning less overall player scrutiny. We can sometimes just get away with turning on ragdoll and not even need a specific hit reaction.

Really, it comes down to the fact that animation and system requirements of making a good melee combat system work are more expensive than making a good ranged combat system work. Since the act of shooting and the act of getting hit are significantly distinct from each other, ranged combat is much easier to get working than melee combat where the hitting and the getting hit are inextricably linked. Combine this with the fact that melee combat is, by necessity, up close and immediately visible and you have a noticeably higher bar for acceptable quality from a melee combat system.

[Join us on Discord] and/or [Support us on Patreon]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

Scaling an encounter to support different numbers of players is generally a question of resource allocation. If we have a small number of these encounters in the project, we can do most of this by hand and spend more effort on each one. World of Warcraft's old raid system where raids supported only 10 or 25 players are examples of such where tuning and mechanics were modified to specific player numbers. However, doing this for one to eight players (or, as it stands now in WoW anywhere from 10 to 30) means eight balancing passes for every encounter. That's far too much work to do in any reasonable amount of time. This means we need to think about building a system of rules to do this for us quickly and easily.

Usually when we're talking about scaling an encounter, we have a number of dials we can adjust, as you said. Here are a few of those numbers

Enemy health

Enemy damage per basic hit

Enemy base attack rate

Enemy special attack cooldown

Number of targets an enemy's normal or special attack can hit

Whether/when an enemy can use a particular move/mechanic

Amount of navigable space for an encounter

Enemy enrage timer

Debuffs on the players

Some of these are more easily scaled, like health, damage, debuffs, etc. while others like specific attacks and cooldowns often require more careful balancing. For enemy health, we consider the amount of damage per second we expect per player and multiply by how long we want each enemy to live, then give a little extra buffer (e.g. additional time for players dodging enemy attacks). For damage per hit, we consider how much health and damage mitigation players should have and how many hits we want them to be able to take before death. For debuffs and cooldowns, we consider what percentage of the encounter time we want the players to have to deal with debuffs or cooldown abilities. It really helps to think of the players as a scalable percentage for the enemies to affect, rather than specific numbers.

If player healing is a role that needs to be handled, it makes tweaking the encounters more difficult since we cannot always expect there a healer to be present. In such situations, we can potentially provide an NPC healer or other environmental elements like interactables or consumables with cooldowns that provide enough healing for the fights. Our goal is usually to make healing and health management an additional task to be considered by the players during encounters, even while solo.

There's a lot more depth we can go into on this - tailoring action set piece raid boss fights in World of Warcraft is very different from doing a broad tuning pass for a squad with variable size in a game like Call of Duty Outbreak - but this should explain the broad strokes of how a system designer approaches this kind of thing. We have a set of different numbers we can adjust (and set up formulas for), and we have a holistic view of the players as a collective of effectiveness that we can modify on a percentage basis instead of an individual player basis. Then, if we have the time, we establish the kind of experience we want the players to have and tune toward that experience.

[Join us on Discord] and/or [Support us on Patreon]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr

The hardcore theorycrafters can get pretty dang close to our internal formulas, but that's not necessarily going to stay the case. We will make adjustments to the formulas for various reasons (e.g. we want to tune crit up across the board, we want to tune this specific ability down, we're raising the level cap in an expansion, we want to change things up for this season, etc.) which can often invalidate previous assumptions.

In a lot of ways, the theorycrafters are actually the mirror opposites to designers. We designers start with the math and use it to try to craft the intended player experience. Our goal is to create a combat experience that is fun and engaging for the player. The theorycrafters start with the player experience and try to figure out all of the math that drives it. Their goal is usually to optimize their combat performance, even if it means sacrificing the engaging gameplay to do it. It's a pretty major distinction - lots of people can understand what it means to break down an experience into the component mathematical parts, but it is far rarer to find people who can use those math components to craft an interesting experience.

[Join us on Discord] and/or [Support us on Patreon]

Got a burning question you want answered?

Short questions: Ask a Game Dev on Twitter

Long questions: Ask a Game Dev on Tumblr