What’s the deal with hitboxes and hurtboxes?

Paid supporters of my Patreon can email me questions that sometimes turn into essays like this one

Hi Pat,

I’d like to know more about hitboxes and hurtboxes. At what point in character design do they come into play? What are the trade offs and balance considerations? Do top tier characters tend to have a certain profile? Do you know ahead of time which characters are top tier?

Thanks!

Boxed Out

Aha, another fine opportunity to don my fighting game design hat! Let’s break this down a bit.

Fighting game characters are collections of animations. Most game characters are collections of animations, but fighting game characters have more animations than most game characters except for maybe character action games, and those usually have only a couple playable characters, while fighting games have a whole bunch. You can kind of think of a fighting game character as a deck of cards for a trading card game, except each card is a unique animation.

Those animations are, first and foremost, intended to communicate the fantasy of combat; you want to see a cool character doing cool things on the screen, and the animations are those cool things. Press a button and a cool thing happens — but what happens, specifically, to your character and the opponent’s, is based on the rules of the fighting game, and the hitboxes and hurtboxes are part of those rules. Fortunately, the rules are simple: If one character’s hitbox contacts another character’s hurtbox, then the character that was hit leaves whatever animation they were in and goes into a reeling animation (also called a “hit reaction”).

The timing and positioning of these hitboxes and hurtboxes during any given animation define the actual gameplay, but since people who play fighting games want to see cool characters doing cool things and not collections of green and red boxes moving around, the games do not show this information during a match. Instead, each animation is responsible for communicating its combat information in space and time — when the attack is hitting, where it’s hitting, and where the character exists in space while doing the animation — without directly drawing the boxes on screen in order to maintain the combat fantasy. Imagine a version of Chess that wasn’t locked to the grid, and each piece’s movement was just kinda vibes; that’s more or less how a fighting game works. As players, we look at the game’s animations, see them interact with each other to get a sense of how they work, and when the outcome breaks our expectations, we get mad (“How did that hit me?!”) and screenshot the offending moment for social media approval.

Check the quotes for some excellent examples.

This is, fundamentally, the main reason why fighting games are hard to make: the combat rules are generally dependent on the aesthetic presentation. I like to describe it as playing real-time Magic: the Gathering, except each card’s function is determined by the details of its art, rather than the art simply being a tool to help us imagine ourselves being cool sorcerers instead of nerds who spent all our money on cardboard pictures.

The order of operations in any given game dev process can be different, but generally, when a team is working on a character’s attack design, they’ll start by brainstorming different kinds of attacks that the character looks like they should be doing. Concept artists might draw out specific poses that make the character look cool as hell, animators might be testing out those poses and movements to see how to make them work in motion, and game designers might be working out their own attack prototypes to determine what kind of moveset the character will need for their gameplay to function. Hitboxes and hurtboxes are assigned when a move gets implemented into the game, and tweaked as needed after playtesting.

Historically, though, fighting games weren’t a particularly thoroughly “game-designed” genre; it was mostly just the animators making moves, and if the team had a battle planner they’d be assigning collision volumes wherever it made visual sense*. After all, these games did not start out as venues for professional competition, they started out as a fun way to spend a dollar at an arcade. Only as the genre grew did the competitive integrity of the game become a major concern, and that led to the emergence of a set of game designers who specialized more in the gameplay outcomes of combat design; designing characters to have moves that yielded interactive and rewarding gameplay when facing another human being.

(*If you read these two interviews with Akira Nishitani (Shmupulations and Capcom), battle planner on Street Fighter II, three things become clear: one, that the character animation was completed before the battle planning started; two, that most of the dev time was spent simply getting the game to work, never mind work well; and three, that the primary use case they were developing for was actually single-player, at first. The team would brainstorm move names, the animators would draw animations that fit those names, and the actual combat rules didn’t really get implemented until after animation is done. So while there was some intentional combat design in there, it wasn’t design-first; the design had to work around the available animation.)

Sentinel’s finger extends just far enough to get caught by Captain Corridor assist at pretty much any range but full screen. Stop snitchin.

The location of hitboxes and hurtboxes, and the timing in which they appear and disappear, is essentially equivalent to the rules of a Magic: the Gathering card; they define the risk and reward for any given move, and thus dictate its function for the player, and counterplay for the opponent. The baseline expectation is that the gameplay outcomes match the visual information; when a character begins a punch, they’ll start by going into a pose that is different from their standing idle, and the hurtboxes will be assigned to match the space they occupy; when the punch is fully extended, the hitboxes will be assigned to match the space on the screen that the punch covers, and the hurtboxes will be changed to match the new pose of the attacking character; as the punch recovers, the hitboxes disappear and the hurtboxes change again until the attacker returns to idle.

However, allowing aesthetic aspects of a game, like the details of an animation’s spacing and timing, to define combat outcomes is a pretty dicey prospect. Animators are the performers behind a character’s movement (and mocap actors, if the game uses motion capture), and their job is to deliver visually compelling motion; asking them to also be responsible for the nitty-gritty of PvP combat interactions is a bit much. The fighting game animators I worked with in my time were all extremely skilled professionals, but I’ve only ever worked with one who I would consider competitively skilled at fighting games, largely because he grew up playing them. (In general, I find that if you’re drawn to fighting games for the aesthetic presentation of character combat, then the nuts and bolts of the combat rules are often less interesting to you; conversely, if you’re really into the PvP game design side of fighting games, the aesthetic considerations are often lost on you, which explains why most indie fighting games are good at one or the other but not both.)

When the hurtboxes fit the animation too closely, though, you get other issues. For example, look at these two screenshots of Makoto walking in 3S:

While the hurtboxes do match the visual expectation, the dramatic difference between the width of her hurtboxes (the blue squares) mean that an attack can whiff or hit depending on how far into her walk cycle animation she is, which is probably not a gameplay detail either player would be expecting to make decisions around. In other words, while the visuals make sense, it doesn’t match how players are thinking about the game while they’re playing it, leading to a different set of unexpected outcomes that is also jarring. Having your game generate a bunch of social media posts consisting of screenshots of seemingly egregious hitbox/hurtbox interactions is annoying, but seeing moments where both players are confused about what happened because too-detailed hitboxes and hurtboxes led to something unexpected is also not great for gameplay clarity. (Personally, I don’t mind this kind of stuff in the games I play, but it’s not for everyone.)

So the placement and timing for hitboxes and hurtboxes aren’t just dictated by the animation visuals, but also by the designers’ intended combat outcomes. A character will have a somewhat-visually-arbitrary hurtbox that shows up in front of them before they swing because the attack is meant to have some additional risk of getting counter hit in startup, or a lingering hurtbox in front of them after the hitbox disappears because the attack is meant to be whiff punished in recovery. The hitboxes go in front of the hurtboxes to make the attack more likely to win against a simultaneous attack at range; if the hurtboxes are close to the hitboxes, or even in front of them, then the attack is more likely to lose or trade. This way, the designers can relieve the animators of being responsible for both the move’s form and function, and the designers can be nerds about the attack properties while the animators can focus on making the characters dance. It’s still a collaborative process between the two, of course, but this way the responsibilities and skillsets are more sensibly distributed.

See how much further Bruce Lee’s red hitboxes extend past the blue hurtboxes? This flying kick is pretty strong.

These days, games generally have combat standards defined early on in the development by a lead or principal combat designer that delineate the design space for attacks: attacks should never be faster than X frames, low attacks should have hitboxes that go no higher than Y coordinate, heavy attacks should have extended hurtboxes for Z frames before and after the attack to create more whiff punish potential, and so on. When an attack has low startup and recovery, lots of active frames, and big hitboxes that extend past its hurtboxes, it’s generally going to be very good, and if the inverse is true, it will be very bad. For a combat designer, the ideal situation is that the combat standards for attack timing and spacing are so clearly defined that the tradeoffs and power budget for any given move are very understandable, all the attacks in the game feel like they fit together seamlessly, and balance issues are easily spotted early on in design and playtesting. If the combat standards are good, a game designer can design an entire moveset by just writing down spacing and timing numbers in a spreadsheet and not have to worry about its balance. The more tightly defined the combat standards are, the easier it is to build characters quickly, meaning it’s easier to deliver more characters and hopefully a successful product.

However, because fighting games are primarily a genre focused on delivering an animation-based combat fantasy, having rigidly defined combat standards can compromise the aesthetic quality of the combat and make it less compelling. The vast majority of players do not care about the rules of combat, they just want to see two cool characters hit each other with cool moves, and they won’t stick around to understand how beautifully consistent and interactive the game’s combat is if the attacks look boring and same-y across characters. After all, there are only so many ways to animate someone throwing a punch to the head or a kick to the body. The point is the fantasy, not the rules that simulate the fantasy, and the more the combat feels like rules, the less evocative it will feel to players, especially the ones that are mostly just here for a good time.

Making Chipp’s sweep also slightly above the ground so it can crush lows is most likely something done “because it looked cool”.

This gets especially dicey once fighting games move outside the realm of punching and kicking and into more abstract kinds of combat, like weapons and magic and stuff. Swords are a bit harder to see move through the air than a punch or a kick, so we use visual effects like blade trails to communicate their space, and that blade trail becomes the visual indicator for where and when an attack is active, but it rarely maps 1:1 for aesthetic reasons, so there will be times where an opponent just walks through your blade swipe and nothing happens. A sword leaving extended hurtboxes in recovery doesn’t even really make sense from a logical point of view, but it’s necessary to create whiff punishability, which is an important part of fighting game counterplay. The further we get from punches and kicks, the more abstract the combat behaviors tend to be, and the harder it is to cover all those situations with combat standards. Do this long enough, and that’s how we end up with Axl’s chains having extended hurtboxes but Johnny’s Mist Finer stance having almost none.

Bruh.

On one hand, this creates a wider range of differences between the characters that makes them all stand out more from each other, which is really important for a compelling fighting game, but it also means that players have to learn a lot more stuff in order to play the game effectively. Personally, this is the stuff I like; if you ever hear old people like me complaining about modern fighting games (like I frequently do), much of the disconnect between older and newer generations of fighting games is that the combat is too rigidly defined in newer games, meaning there are fewer esoteric things to find and the gameplay generally becomes predictable and understandable very quickly. Classic Guilty Gear games having different wakeup timing per-character doesn’t start out as a venue for characters to express power, it’s just a function of the characters having different wakeup animations that fit the character fantasy, but they quickly come to affect power in unintended ways that I find far more interesting than keeping everyone operating in the same box. As a player, I do not like it when the combat feels “designed”, because nothing that happens in those games feels novel to me; understand the rules and you understand the game. When a fighting game’s visuals dictate the rules, you end up with a lot more weird unexpected stuff going on, and the experience of playing the game feels more like exploring something new and unusual.

In practice, I think what usually happens is that the combat standards are effectively defined early on in character development, but once the character roster gets further along, the team has a better sense of where the combat standards are too restrictive, and later characters get more toys to play with as the team gets more comfortable making them, so the combat standards have to be updated to accommodate for the gradual evolution of the game. This is healthier, I think, and better for the game overall; the standards exist to give character combat designers an early framework to operate in, they build towards those standards and test the game out, they learn the reality of the game beyond the framework, and then adjust the framework to fit their new understanding of the game, which then yields a different set of characters. However, this is far less efficient than simply defining a standard once and going with it, which I think is closer to what we see in most modern games these days.

As for character tiers; a character can certainly be top tier based on their hitboxes and hurtboxes, but it’s usually not just that alone. Sagat in Capcom vs. SNK 2 has many of the best buttons in the game, but his comboability off those buttons isn’t great, and if he had lower damage and HP (like he does in CvS2 EO) then he wouldn’t be so strong. Most traditional 2D fighting game characters tend to have something like 40–70 moves or so, and most fighting games have at least 12–16 characters, so if you think about the matrix of interactions that each character’s moveset can have with another character at any given location in space and time, it’s dense enough that it is pretty hard to predict with any certainty who is going to be stronger or weaker by numbers alone. This is especially true for games where attacks can combo or chain into other attacks, since now you have to account for the power of each attack based not only on whether it is strong or weak by itself, but what kinds of rewards and subsequent situations it leads to next. Yeah, if you have a character with no extended hurtboxes, big hitboxes, and minimal startup or recovery time, then they’ll probably be really good, but it’s pretty rare to see an outlier like that in the modern era.

You wouldn’t expect this to be one of the best buttons in fighting games, but you wouldn’t expect this to be 7f startup 8f active either.

Anyway, if there’s any takeaway to this one, it’s that fighting games are extremely difficult to make, and it’s no wonder that most of the teams making them are the ones who have been making them for decades. Pretty much every other game genre has gone out of their way to decouple aesthetics from gameplay outcomes, especially PvP genres, because it introduces so much more work for the developers. But if you love them, they might very well be the only things that ever hit quite like this, and it’s because of all this work we do.

Thanks for reading!

pm

*Bluesky | read my words*

*Twitch | watch me play*

*YouTube | watch me yap*

*Patreon | support my work*

*itch | buy my books*

💪😎👍❤