What makes a fighting game deep (and why it matters)
Sign up for my Patreon to keep up on my work! Paying subs also get to ask me questions.
Hey folks! This is a long one that I’ve been working on for a while, so I recommend either saving it until you’ve got some good deep focus time to chew on it, or picking at it in small chunks.
Also, check the Patreon post for a Sunny-cat video for the book bounty! I’m getting more videos of him these days than the others because he’s kinda ridiculous, but I will endeavor to spread the action around a little more. Look at him in his little box!!!! Anyway, buy a copy of From Masher to Master 2 and we’ll keep up the cat video streak.
If there’s one thing fighting game players love to do, it’s argue about fighting games. Which games are better or worse, which games are worth playing, which games are hype and which games are sleep.
In the process of arguing about why their particular favorite fighting game is objectively the best, fighting game players often bring up game depth as an indicator of the game’s worth or appeal; to play a deep game is considered virtuous and valuable, its competition more worthy and its competitors more impressive and capable. Yet the deepest games are generally not the most successful or popular, and people rarely agree on what makes a deep fighting game.
This is because fighting game players generally do not understand what makes a fighting game deep, or how that depth shapes the player’s experience of playing the game. So, since I’m tired of seeing people have wack arguments about this kind of thing, I figured I’d lend some perspective. I got to do a lot of early-stage fighting game prototyping and iterating during my years on 2XKO as a designer and producer, and I learned a lot about how to design for different kinds of depth in the process. I also learned a lot about how the different sources of depth affect the holistic experience of playing a fighting game at different levels of player skill, and how the overall shape of a game’s depth can attract or alienate its players. Let’s talk about it.
Where fighting game depth comes from
First, let’s start with a couple basic definitions: Fighting games as a genre are generally characterized by, well, fighting; specifically, real-time fighting between at least two characters who are moving, attacking, and defending on a single shared screen. The players control their characters, telling them to move, attack, and defend on the screen, and the outcomes of those decisions are determined by a varied set of skill tests. Depth, in the context of a fighting game, refers to the set of strategically distinct decisions that the players get to make during the course of the match, and the skills tested by each decision, both physical and mental. In general, the more stuff there is in a fighting game to master, and the wider the variety of skills a player must refine in order to master that stuff, the deeper the game. (This does not scale linearly, mind you, but we’ll keep it simple for now.) Let’s dig in.
When you and your opponent pick your characters at character select, you’re choosing a list of attack, defense, and movement animations that create hitboxes and hurtboxes on the screen at certain times, with damage numbers assigned to those hitboxes. The first layer of depth comes from comparing your animations with their animations to see which of your animations beat theirs, usually by hitting them before they can hit you, or defending the attack and recovering before they do, or moving your hurtboxes out of the way, as well as the damage reward you get for winning. In a well-made fighting game, each action beats some others, loses to some others, and pays out a reward that is based on how often the action wins or loses compared to the others, so a fast jab will win frequently but do little damage, while a big heavy punch will lose often but pay out much more when it wins.
This layer of fighting games is often referred to as “Rock-Paper-Scissors”, though more specifically, it’s “Weighted Rock-Paper-Scissors”, because the reward for winning an interaction is not equally valuable the way it is in a traditional RPS game; imagine Rock-Paper-Scissors where winning with Rock is worth a dollar, Paper five dollars, and Scissors ten dollars, and that’s closer to the dynamic we’re working with. So, not only do we have the basic prediction gameplay of Rock-Paper-Scissors, where we’re choosing options against an opponent and trying to figure out how the previous choice influences their next one, we also get different values to each option when it wins, giving the players more information to predict in the form of how greedy or conservative they want to be with their choices. The weighted RPS is thus a complex test of pattern recognition, where both players are competing to discern their opponent’s thought process and cash in with higher-value reads as often as they can.
Picking characters also creates another layer of depth; since not every character has the exact same moveset, the weighted Rock-Paper-Scissors interactions are different per character matchup, and so each player has to learn the game from the perspective of each character they want to play. A fighting game is thus not just a single game, but a different game for each character matchup; the “game” that the players are choosing to play in any given match is first determined at character select. This is also where the first layer of “metagame” depth comes into play, as players can choose characters based on their understanding of whether any given character is strong or weak overall, or strong or weak against the opponent’s character pick. Players can additionally choose to just focus on one character, or play multiple characters, and get different outcomes based on how the game rewards them for playing broad or deep. If the game offers any other pre-game decisions that further change how each player plays, like input methods or different sets of game mechanics (such as 2XKO Fuses or CvS2 Grooves), then these also create additional matchup permutations that players have to learn how to play effectively.
When we take the Weighted Rock-Paper-Scissors combat dynamic and add Character Matchups, what we get is a heavier test of a player’s knowledge. In a game with only one character, both players simply have to memorize a single table of win/loss/draw interactions to play the game fluently, but now that we have a different table for each character matchup, each player has to memorize every single table for every single character matchup to play completely fluently (or, at least, every single table for the characters they choose to play). For most games, this takes a lot of time and effort, and the players who put in more effort here are rewarded with a better idea of which tools are more or less effective in specific matchups, as well as the insight to understand when an opponent is making suboptimal choices based on their lack of knowledge about a specific matchup. Character Matchups are primarily testing each player on their memorization of the game’s details; players who have diligently studied the interactions between both characters in a match are more capable of making informed decisions in gameplay than players who haven’t.
Crucially, this is a very different kind of skill than the pattern recognition involved in Weighted Rock-Paper-Scissors; matchup study is first and foremost studying the game itself, rather than studying the opponent playing it. There are also many different ways to study matchups; a player might begin by just dissecting attack interactions in Training Mode to see which tools are good or bad in any given situation, but as characters and games get more complicated, players will often find that they’ll need to seek out skilled players of every character in order to develop a more complete understanding of how each character matchup works. This means that players can differentiate from each other on how well they understand certain matchups; it’s not uncommon for high-level players to win bad character matchups frequently, simply because they’ve put more effort into studying their bad matchups than their opponent has in studying their good matchups.
So far, our fighting game is Character-Specific Weighted Rock-Paper-Scissors, and we can see that we’ve already got a lot of depth, and we haven’t even started on the stuff that is unique to the medium of video games! Now it’s time to look at the screen.
The next layer of depth comes from the screen itself; namely, the space that both players are fighting in. Remember how much depth we got when we took the set of Weighted RPS interactions and multiplied it by the number of Character Matchups in the game? Well, it turned out that we kind of over-simplified the situation, because the Weighted RPS interactions are actually different based on screen position. Just comparing attack win/loss outcomes based on attack startup time would make sense if both characters are just standing next to each other all the time, but because both characters are capable of freely moving around the screen, those interactions will work out differently based on each character’s screen position; a standing jab will beat almost anything point-blank, but lose the moment the opponent is out of reach.
Of course, fighting games do all kinds of things to reward players for getting advantageous screen positions; if you corner an opponent, then they cannot run away, and you’re free to take whatever position you want. Many games will also reward players for cornering the opponent with additional combo potential, or even a win entirely, like the Ring Outs of Virtua Fighter and Soul Calibur. Every single strategically distinct screen position is a different gameplay situation, and the task of understanding the game is complicated significantly now that space is involved, because now the players have to understand the Character Matchup Weighted RPS based on a specific screen position, and play around the option to move around the screen to escape unfavorable positions and find advantageous ones. What’s more, moving around the screen takes time, and all of this is happening in real-time, so players are essentially doing a lot of fast geometry to figure out when and where the two characters are going to collide on the screen. Our fighting game is now Character-Specific Weighted RPS Across Space and Time.
This is a lot of math! The Weighted RPS part is basically statistics with a little psychology, and the Character Matchup study is largely memorizing a bunch of basic arithmetic to figure out which options win or lose, but when we add the spatial and temporal elements of a real-time video game, we have to do a lot of real-time geometry to figure out which Character-Specific Weighted RPS game we’re playing. It is the element of playing the game through space and time that makes fighting games qualitatively different from other comparable competitive games, like Poker or Chess or Pokemon. The math itself is relatively easy, for the most part, but having to do it as correctly as possible and faster than your opponent does is the hard part.
However, it’s not all math. Adding the element of time, in particular, introduces an additional layer of depth that creates distinctly different skill tests related to how the players control the game in real time — in other words, how effectively they execute commands in response to the game outcomes. Where board games and card games usually progress by having each player take a turn that advances the game state, fighting games are real-time, and must be operated in real-time. When a player makes a decision in a real-time fighting game, they have to physically perform the input that executes that decision in real-time, and their ability to make correct decisions is now more difficult because they have to both know the correct decision and do the correct decision. What’s more, the difficulty of performing that decision in real-time is often related to the reward that the decision offers if correct. It’s less of a mathematical test, and closer to playing a sport or musical instrument, where the players have to use their bodies to manipulate an input device in order to perform very specific, intricately timed combinations of movements in real time. Before, we were describing a game that felt like it could be played by a robot, but now that physical movement is involved, this game starts to feel distinctly human, as all the real-time math we’re doing has to be expressed through our bodies, each of which are different from each other through nature and nurture.
When we add real-time execution to the decision-making games mentioned at the beginning, we now have an additional layer of execution performance to test the players on, where some players may know the optimally correct decision in any given situation, but cannot perform it accurately in time. This means that every decision a player can make in a game is now modified by a chance they fail to perform it, and players can differentiate themselves over how effectively they perform higher-value actions that require more complicated movements and timing windows, and whether they choose to spend more of their practice time on execution or matchup study. We’ll call this an Execution-Based Character-Specific Weighted RPS Across Space and Time. (Of course, since that’s a mouthful, we can also just call it a fighting game.)
The real-time element is also where fighting games lose a lot of would-be fighting game players, because they want to do all the beautiful math thoroughly and completely without having to play under the pressure of another player doing math a little less thoroughly but a lot faster. To be good at a fighting game typically looks less like doing the math perfectly, and more like doing “good enough” before your opponent does, which has the effect of curtailing some of the depth on any given skill check, and replacing it with real-time cognitive management skills (commonly referred to by the “mental stack” analogy. You do not have time to look up the appropriate-speed move in the middle of a game of Tekken, and so you instead have to prepare your mind ahead of time to respond to the situation with something fast enough to get the hit. In other words, it is not enough to simply know, you have to be able to do, and the kind of physical and mental training required to do competitively is a different set of skill checks entirely.
If I were to sum up the depth that is created by adding all these different layers of gameplay together, it’s that we have an extremely complicated set of game states that have to be acted upon in real time with physical inputs, and so the challenge at hand is to outperform our opponent across a series of interrelated tests of game knowledge, quick decision-making, and physical execution. This, too, is a mouthful, so people in fighting games often refer to this set of tests as the Head, Heart, and Hands; I first encountered this while working with Seth Killian on Project L, but others have described different permutations of this framework as well. It is depth that is created by making a game where humans have to do a lot of different difficult things very quickly, and the ways in which a game challenges its players to do those things creates space for different humans to rise to the challenge in different ways.
The critical aspect of fighting game depth, then, is not found only in the degree to which a single skill check is linearly tested, but in the varied nature and distribution of those skill checks that give players the room to find different ways to pass those checks, different ways to practice the game in order to improve at those checks, and allow different players to disagree about how best to play the game. Put simply, you cannot be great at a fighting game without developing your head, heart, and hands, but you cannot be great at each aspect without infinite time. The challenge to the player is thus to develop yourself as best as you can with the time that you have, and the challenge to the game developer is to build a fighting game which tests the players in ways that achieve their goals, whatever they may be — which we’ll get to next.
How depth affects the fighting game player experience
I mentioned in the introduction that fighting game players like to argue about depth as a virtue of a great fighting game, and now that we’ve identified where the depth comes from, it feels kind of funny to argue that a game is better because it requires more work to be good at. But it’s true! Deeper games require more time and effort to master, so it feels more rewarding to be good at them, and people generally think it’s cool to spend a lot of time and effort getting good at a thing. Think about all the videos you’ve seen of some person you’ve never heard of, effortlessly spinning on their head, or kickflipping a skateboard over a huge set of stairs, or answering trivia questions on Jeopardy; all of these make us go “Wow.” The further a person is from our understanding of what the average human being is capable of, the bigger the “Wow,” and the more effort it takes to master a thing, the further we get from the average human being.
However, there’s more to depth than just the “Wow” factor. While most people who play fighting games don’t get anywhere near experiencing a game’s full depth, the players who do go deep will have very different experiences based on the degree of depth, the different decisions and skill tests that create that depth, and the ways players can spend their time mastering that depth. So now let’s talk about how depth influences the experience of playing the game itself.
At the most basic level, people buy video games because they want to experience something, and people who like video games that give them lots of things to learn often like fighting games precisely because there’s so much stuff to learn; depth is basically the content in the game they’re consuming. However, for players who don’t really like learning stuff, but might put up with doing the minimum amount required to get to the part where they can beat someone else, more stuff to learn is a detriment to their experience, because it gets in the way of the part where they get to beat someone (and, indeed, if they’re playing against people who like studying and practicing, they’ll probably lose). This is where complaints about “sweats” and like 99% of the stuff on ScrubQuotes comes from; the players who want to have an uncomplicated time playing a game will lose to the players who enjoy studying the game and breaking it down to understand how to play it better.
In other words, these players want a game that feels rewarding to play and win (so they can feel good about winning), but they also want it to not reward the things they don’t feel are fun (studying and practicing). They might like the fact that they’re playing a game that is respected as deep and worthy of mastery, but they don’t like the kinds of things they have to do in order to master the game themselves. If the game were less deep, they might feel less pressure to practice it, but they’d also feel like the wins were less meaningful because they didn’t require as much skill. In other words, their problem isn’t the amount of depth in the game, it’s the design elements that created the depth, and the different ways players must spend their time practicing in order to master that depth.
For example, let’s compare a typical Tekken game to, say, Marvel vs. Capcom 3. Tekken is a one-on-one game compared to MvC3’s teams of three, with tagging and assists; Tekken’s individual characters have well over 100 different moves, while an individual MvC3 character usually has a couple dozen; Tekken movement is limited mostly to sidestepping and backdash canceling (also known as Korean Backdashing), while Marvel vs. Capcom 3 characters have to use the entirety of its very large screen to jockey for position; Tekken primarily rewards players for knowing matchup frame data and determining their opponent’s patterns based on that matchup knowledge, while in MvC3 frame data doesn’t really matter (because assists and chain combos just kind of wash over all of it) and the primary reward is located in the knowledge and execution of team-specific combos and setplay.
Despite the fact that they’re both extraordinarily deep fighting games, they could not be more different in terms of where the depth comes from, and this shapes the player experience rather profoundly. A new Tekken player is going to spend a lot of time learning their character’s moves, and then every other character’s moves, in order to understand the basic flow of combat and the risk/reward calculus in any given situation, and before they have a decent base of information built up to intelligently play from, they’re just kind of guessing based on whatever they see in front of them — kind of like playing Poker without actually knowing how the different hands work. At first, the novice Tekken player is mostly just learning their character against other players who are also learning their own characters, and the contest is in figuring out who knows their own character better, but as the novice progresses towards intermediate-level, they’re tested on understanding matchups and figuring their opponents out with limited information, which is often far more grueling in comparison to simply learning how to do your own stuff.
Meanwhile, a new MvC3 player is immediately forced to pick a team of three characters, which is pretty challenging when they don’t know anything about the game or its systems, and learn how to play all three characters (each with three different assists) at once. The typical learning curve usually involves learning a basic combo or two with each character, a couple uses for the different assists in neutral, pressure, and combos, some additional combos using the TAC and DHC mechanics to safely switch between characters, and one or two uses of X-Factor to start comebacks. Since MvC3 combat interactions are typically not determined by frame data but by screen position — like using a projectile assist to lock down an opponent in neutral and then crossing them up with the point character to get a hit — there’s much less emphasis on learning each individual property of a move, and greater emphasis on learning how each character works in conjunction with each other to control the screen and deal damage. There are also significantly higher upfront execution challenges in MvC3, since the input buffer is far less forgiving than Tekken’s, and the effective control of a full team requires many more inputs than piloting a single Tekken character does.
If we compare the new player experience between both games, then, the Tekken player will generally have an easier time playing other people and learning through mashing early on, while the new MvC3 player will often find themselves losing to people who might have less live PvP experience but a lot more lab time. Once a MvC3 player can pilot their team reasonably competently, though, they get to have a lot of fun playing to do their team’s bullshit against other players trying to do their own team’s bullshit, and the nature of the 3v3 team compositions offers a really exciting amount of personal variation, where each opponent can feel like they play completely differently due to the amount of team-specific work that goes into learning to play a MvC3 team. Meanwhile, the mid-level Tekken player often finds themselves in a pit of despair as they try to play “good Tekken” against a bunch of players who are mashing suboptimally and getting away with it. If a Tekken player tries to learn MvC3, they may find themselves turned off by the emphasis on upfront labbing and execution practice compared to Tekken. Likewise, a MvC3 player might find the task of memorizing the attack interaction tables to be a drag compared to the satisfaction they find in freely experimenting with a MvC3 team to find its possibilities, and in teaching their hands to perform new tasks that allow them to control the team far more effectively. (A Tekken player once described it to me as “competitive real-time flash cards.”) Both the Tekken player and the MvC3 player are nerds, of course, but the flavor of nerdery is different, and different nerds prefer different flavors. When depth comes from something you don’t really like, then it just kinda feels like work.
Creating room for variance within deep fighting games
Now that we can see where depth comes from, and how players practice in different ways to master the skill tests within that depth, it’s time to take a look at where the game gets its variance from. When players dismiss games for being “random” or “nerdy”, they’re usually referring to games that have high and low amounts of variance, respectively; games with high variance feel like the outcomes of a match don’t feel tied to each player’s individual skill, while games with low variance feel like pre-determined, linear skill comparisons with no room for exciting upsets or comebacks. When Guilty Gear creator Daisuke Ishiwatari wrote in the pre-release messaging for Guilty Gear Strive that previous GG games were like Shogi, and Strive more like Mahjong, the difference he was describing was between games with low variance and high skill determination, and games with much higher variance; older GG games simply have more stuff to master that allows the experienced player to win far more consistently without having to deal with predicting an opponent’s guesses.
Variance is essentially the properties of a game’s design that allow a game to create different outcomes within the same play space; each fighting game has different sources of variance, and different degrees of variance within a match. Sometimes it comes from actual randomness; stun values in Street Fighter 2, or Faust’s item tosses in Guilty Gear, are outside the control of either player, and can lead to dramatically different outcomes even if the rest of the match is identical on an input-by-input basis. Most of the time, however, it’s tied to human elements; extremely difficult execution challenges create variance by being droppable even at the highest levels of play, unreactable guessing situations (like a strike/throw mixup) force both players to essentially guess their way out, and so on.
Get Patrick Miller’s stories in your inbox
Join Medium for free to get updates from this writer.
The sources of variance create different effects upon the game itself as well, since players in a competitive game are competing to reduce variance as much as possible so they can reliably win, and the way a player plays around variance is shaped by where it comes from. When variance comes from the game itself, it feels like both players are basically playing statistics around a random number generator (kind of like playing around drawing from a deck of cards); when it comes from head-to-head hard reads, then it’s an exercise in chasing your opponent’s pattern recognition; when it comes from execution, it feels more like a sport, where the players are grinding to reduce their own execution failures, pressure their opponent’s ability to execute in stressful situations, and make intelligent risk/reward decisions based around the probability of execution failure in any given situation.
Additionally, the degree of variance itself also influences the game experience; if gambling on a hard read could get you 10% of your opponent’s life bar, the room for the variance to change the overall outcome of the game is much lower than if gambling could get you 40%. Games with very low variance largely feel like linear tests of skill; if I am better at this thing than you are, I will win every time. There was a side tournament at Evo Japan a couple years ago for a shoot-’em-up training tool, where players competed head-to-head on how many times they could press a button within a certain amount of time; this game is a completely linear skill test with zero interactivity between players and minimal variance outside of basic athletic factors like physical exhaustion, so we would expect the player who is better at pressing buttons faster to be the winner every time. Older fighting games, like Capcom vs. SNK 2 and Marvel vs. Capcom 2, often have much lower degrees of variance than newer ones, because a lot of the skill tests in those games are in execution and reaction more than hard reads, and so the player who grinds their hard stuff is just going to win most of their games over the players who don’t. Newer games are typically designed for a much higher degree of variance, so that the gap between players of different skill levels feels less intimidating and insurmountable (and therefore more approachable to the lower-skill players that constitute the majority of a fighting game’s potential market), but this often makes the game feel far more volatile to the more heavily invested players, who want to feel that their time spent is consistently rewarded with wins against less-invested players.
If we return to the Tekken and MvC3 comparison, both games have a moderate amount of room for variance through both hard reads and execution failures, but the players looking to reduce that variance and win consistently have to grind the other core skill checks — memorizing flash cards for Tekken, and labbing setups and combos for MvC3 — if they want to lower the risk of losing to players who guess right. Of course, they can also get better at guessing right; players can stabilize a game’s volatility in hard read situations by leveling up their pattern recognition skills to quickly profile an opponent’s likely future decisions, and develop their risk/reward calculations to reduce the risk of catastrophically bad reads. But no one can guess perfectly every time, especially in a short tournament set where players do not have much time to figure each other out, and so players have to grind the lower-variance skill checks, like game knowledge and execution, in order to stabilize their game performance as much as possible against the more volatile parts of the game. If you want to get good at these games, you’re going to be spending a lot of time doing that stuff, and if the game you want to get good at involves doing a bunch of stuff you don’t want to do in order to get good at it, you’re either going to have to deal with it or find a new game.
How deep is too deep?
At this point, we may be tempted to ask: If we think deep games are cool, and being good at deep games is noteworthy and valuable, why aren’t we all playing the deepest games out there? Shouldn’t developers be making games that are as deep as possible? And yet, we know that people generally don’t flock to the games with the deepest degrees of mastery. And now that we have a basic understanding of some of the sources of depth in fighting games, we can accurately describe why this is — both from the player’s perspective, and the developer’s.
As players, we may think deep games are cool, but whether we want to actually play those deep games depends in large part on where the depth comes from. We understand that the depth in a game comes from all the different little skill tests contained within it, and the effort that we put in to master those skill tests. If you really like competitive flash cards, or real-time geometry, then playing games built around those skill tests will be a natural fit for you. However, other players will treat those tests as a necessary chore they’ll do begrudgingly in order to get to the fun part. And since fighting games get depth from multiple different kinds of skill tests, each additional type of skill test increases the likelihood that you’re going to find something that you think is a chore.
We see this most commonly with the fighting game players who come from turn-based one-on-one games, like card games or board games, who are initially excited by the prospect of playing a game that has a similarly rich possibility space, with room for personality and mastery, but look a lot more compelling as real-time video games. However, these players are frequently turned off by the parts of the game that test real-time hand-eye coordination and input accuracy. Similarly, the relaxed execution requirements and rewards for newer fighting games often turn off the fans coming from older fighting games, who were drawn to the genre because of its emphasis on physical dexterity, and don’t like how newer games force players into more math-based skill tests. If you want a brainy duel, you’ll go play Chess; if you want a physical contest, you’ll go train Boxing; fighting games are closer to Chessboxing, where you have to be down to go deep in both body and mind, and the potential audience for that is a bit narrower than just one or the other. And if we took Chessboxing and added trivia questions to it, like real-time Combat Jeopardy, we would get unimaginable levels of depth, but the set of players who would be down for this would be even smaller, and while it would be cool to watch, most people would not finish watching a Combat Jeopardy top 8 and immediately start grinding for the next event.
Indeed, if you add too many diverse skill checks to a game, then the skill expression of each individual skill check can decrease as the different skill checks interfere with each other, making the game starts to feel more like a cursory test of a couple dozen different things than an exercise in depth of mastery. If a game offers you a rich set of compelling characters worthy of mastery, and then accidentally gives one of them a one-button unblockable infinite combo, well, the game is now One-Button Unblockable Infinite Combo: The Game (until that character gets banned, anyway). Part of figuring out how to play a fighting game is assessing the kinds of skill checks that you must master in order to succeed as a player, and if the game feels like it’s misleading you about the nature of those checks, or undercutting the value of going deep on one skill check by giving you an easier alternative to pass it, then the players will feel that their time investment has been disrespected. A compelling fighting game not only has to combine multiple kinds of skill tests for depth, it also has to make those skill tests feel related to each other, so that a player feels like mastering all of them feels engaging and worthwhile, rather than a collection of arbitrary tests that just happen to be connected by the fact that they’re present in the same game.
Increased degrees of depth also make it harder for any given match to be competitive, since there’s a wider range of skill levels available in the game. More depth means more ways for a skilled player to crush a less-skilled player, while a narrower range of skill reduces that possibility. Anyone can play Rock-Paper-Scissors and have a reasonable chance at winning no matter how big the experience gap is between players, but as the games we play get deeper, the harder it is to find a similarly-skilled partner to play against, leading to more matches that often just feel like foregone conclusions. These games may still feel satisfying to play; winning because you worked harder than the other player rarely feels like a bad thing for the winning player. But even if the game itself is deep, your experience of that depth in an average match might feel extremely limited unless you’re able to reliably find people who will take you as deep as you can go.
If a lot of games lead to one-sided stomps, the players getting stomped often will get tired of the stomps and eventually stop playing, and the people watching the game will go to something more volatile and suspenseful. And if you want to play a deeper game with your less-experienced friends, it’ll be harder to play with them and have a mutually satisfying time when the skill gap is so pronounced that the games themselves don’t feel competitively exciting. This is why I like to say that it’s easier to make friends with fighting game players than to make friends into fighting game players; a game that pushes your buttons in all the right ways for you, probably won’t hit your friends the same way, and so while groups of friends will often start playing a fighting game together, it’s very common for only one of them to stick with it while the rest fall off.
Depth thus creates tension for fighting game players, as a game needs enough of it to feel like it rewards a player’s dedication to practicing and playing, but not so much that it scares off the players who are just testing the waters and not yet ready to commit. A successful competitive fighting game needs to balance these two factors just right. Sometimes, this requires decreasing the sources of depth and compressing the range of player skill, as we saw in the Strive shogi-to-mahjong example, but sometimes this involves hiding sources of depth from players so as to avoid scaring them off until they’re ready to go deeper. Let’s wrap this essay up by talking about how modern fighting games are built to be just deep enough.
Depth trends in modern fighting games
Now that we’ve established a functional understanding of where fighting game depth comes from, and how its sources of depth shape the overall player experience, we can look at modern fighting game design trends and evaluate their intention and impact from a more analytical perspective. Modern fighting games are both simplified and more approachable than previous generations, while also deeper and more taxing in other ways, and it comes down largely to the changes that their developers have made to the sources of depth in order to attract and retain a larger audience while maintaining a competitively engaging degree of depth for highly-invested players.
Since the tension between depth and approachability affects whether players buy and play a game, it also naturally affects the way developers make the games. Modern video games take a lot of time, effort, and specialized skill to make, and any game made by paid professionals is going to need to at least sell well enough to pay them to make the next game. The more people working on a game, the bigger the budget, the more copies the game needs to sell. So, the developer needs to make the game deep enough to feel rewarding, but not so deep that it alienates the players who are first and foremost looking for an uncomplicated good time. This is where the fighting game as a product that people spend money on comes into conflict with the fighting game as a game or sport that yields rich interactions between its players.
So far, we’ve mostly been talking about fighting games as activities, in the same way that we’d describe games and sports. Chess and Tennis exist primarily as sets of rules, and buying those games simply means buying the pieces and tools required to play the game. A fighting game is also a set of rules, but those rules are encoded in software alongside a lot of art and sound to be played on a computer or game console, and people buy the game not just for the rules, but for the rest of the experience — single-player stories, combo trials, online multiplayer, whatever — because they want to have a good time. Even though the fundamental nature of a fighting game, as a game, is closer to Chess or Tennis, the ways in which fighting games are manufactured and sold, and the people who are buying them, have more in common with buying a single-player game or a movie. The money that people spend on a fighting game is considered “entertainment,” and their expectation is to have a complete, entertaining experience before moving onto something else. Compare this to buying a Chess set or a tennis racket, where the primary value of the equipment is in using it, and while you may not buy either item thinking it’ll be a lifelong obsession, it’s a tool that enables you to play the game whenever you want, and its value is tied to how much you use it over the lifetime of the equipment, rather than as a consumable experience to be enjoyed and then set aside.
In practice, most of the people who buy fighting games will buy them to use them as disposable entertainment; they’ll open up the game, play through all the story content, maybe play some games with their friends, and put it down within a couple weeks once the next thing comes out. When fighting games sell millions of copies, the vast majority of those millions are not going to play the game for more than a couple hours. But they spent just as much on the game as the players who will play for thousands of hours (well, outside of season passes and DLC characters, anyway), and their sales subsidize the experience for the players who stick around for the long term. Meanwhile, the players who keep playing the game become its long-term source of content; they’re the players who keep the netplay modes populated with opponents, the ones who stream the game and compete in tournaments that keep the game in the spotlight, and continue to market the game through their play so that when the next one comes out, it’ll have an existing audience ready to repeat the cycle all over again.
Developers thus have to split their time and focus to satisfy two distinctly different kinds of player; the one looking to have a good time before moving on demands maximum satisfaction with minimum time investment, and the one looking to spend several years and thousands of hours mastering the game to the fullest demands a rich, deep game worthy of their time. Both players are important for a fighting game, but both players are served very differently. The entertainment player treats their time as an expense, and every hour spent learning to play a fighting game is an hour they could spend on something more exciting and immediately gratifying, so developers chase this audience with simplified combat design to make it easier for them to pick the game up and feel engaged, and expensive single-player narrative content in order to satisfy their itch for easily consumable entertainment. The lifestyle player, meanwhile, treats the cost of the game as simply the entry fee for what may be a very long journey; their primary investment is the time and effort of mastering the game, and the depth that turns away entertainment players is the content that keeps lifestyle players around for the long haul.
When we look at modern fighting games, especially the recent versions of franchises with long histories like Street Fighter, Tekken, and Guilty Gear, we can see a couple notable common trends between them. Combat feels very close, often uncomfortably so. Mixup situations often fundamentally rely on either hard reads, like guessing on a throw, or very explicit reaction checks like reacting to an overhead. Individual character matchups are heavily streamlined, and decision-making in combat is often shaped more heavily by playing around the game’s global systems than it is character-specific tools or nuances. And much of a character’s power in the game is tied to meters and resources that affect the risk and reward of any given decision more heavily than, say, strict execution or reaction checks. While each of these trends are often criticized for decreasing the depth of the game and the room for a player to outplay each other — and make no mistake, they do indeed have this effect — they are also done thoughtfully, with the goal of still retaining enough depth to make the game worth playing. Let’s go through each of these examples to see how they work.
When it comes to movement and spacing, the trends are pretty easy to see. Guilty Gear Strive slowed its airdashes down and reduced the screen size from Xrd (which was itself in a smaller screen space than XX); Street Fighter 6 introduced Drive Rush, which allows a character to zoom across the screen for a minor resource spend, as well as other systems like Drive Impact and Drive Parry which make it harder for characters to control space at midrange with pokes and projectiles; Tekken 8 reduced the efficacy of its backdash canceling tech to lower players’ rewards for disengaging movement, and which instead pushes them to interact with the more volatile sidestep system. Even more chaotic fighting games have seen similar trends; Melty Blood: Type Lumina retained a lot of its trademark air movement, but gated some of them behind a meter mechanic and put the characters in stages that are smaller than Actress Again: Current Code’s with attacks that are often bigger. Team games like 2XKO and Marvel TOKON Fighting Souls are far more grounded than many of their genre predecessors, and cram more characters with bigger attacks into smaller spaces. Even though many of these games aim to offer movement that feels good and satisfying from a control perspective, players have less meaningful room to navigate around the screen, and do so more slowly, decreasing the room for players to express their skill by outmaneuvering each other to land the hit.
The reasons for kneecapping movement are pretty clear to anyone who has spent substantial time playing older fighting games; getting your ankles broken by your opponent’s superior footwork feels like going up against the speed specialist in any battle anime ever. They’re airdashing over your fireballs, baiting and whiff punishing your pokes, crossing up your blocks and dodging your throws, and you lose without ever feeling like you got a chance to fight. When movement is strong in a fighting game, the more skilled mover gets to dictate where and when the combat happens, and that feels real good when you’re the stronger player and real bad when you’re the weaker player. As a source of depth, it is both highly satisfying and highly polarizing, and if you’re trying to make a game to appeal to a general audience, the polarizing stuff is generally the first to go. If you lose an interaction because you picked a button that was slower than your opponent, or because you guessed wrong on a throw, then you can remember to pick a faster button or guess differently next time, but if you lose an interaction because your opponent is better at moving around the screen than you are, your only choice is to play around the movement deficit in the moment and grind out your movement skills in the lab when you get home — something that the sickos usually love, but the less-invested aren’t as excited about.
Even games not traditionally known for rich, expressive movement, like the Street Fighter series, have gone through similar kinds of streamlining by reducing the variety of ranges at which players can engage in meaningful combat. Projectile attacks have gotten weaker in every successive version of Street Fighter, and characters are not differentiated nearly as much by their effective range as they used to be — both through their attack design conventions, and their global systems (dashes, parries, that kind of thing). The footsies game of spacing and timing normals is still present, but it’s not as important as it used to be, because each game’s global systems typically offer players a number of tools to go through pokes (Focus Attack, Drive Impact, etc.) instead of having to outpoke the opponent. If you’ve ever played an older Street Fighter game against a player much better at footsies than you, it can feel similarly unwinnable as playing against someone with better movement than you in an older Smash or Guilty Gear. By giving a player tools that force the footsies player to guess or react, the game essentially decreases the relative reward for going deep on the character-specific knowledge, spacing nuances, and razor-thin reaction checks within footsies by adding different skill checks to the interaction that make it easier for newer players to parse. Put simply, it’s not that the footsies skill checks aren’t there, it’s just that players have more tools for circumventing footsies against players who are better at it. If you like the linear-but-highly-interactive focus mode created by a game with strong footsies, though, having to deal with the parts of the game that take you out of it feels like the game doesn’t want you to play the parts of the game you like the most.
Mixup situations have also likewise been simplified to make the challenges less complex and more obvious. (This pattern is, of course, different depending on the game, but if you compare any modern version of a game to its predecessors, you’ll generally find the trend holds.) Ambiguous crossups are generally rarer, defensive mechanics that do not need to account for left/right (or even any direction) are common, and both universal and character-specific mixup tools push more on hard reads (like a strike/throw) and reaction checks (overheads and stuff like Drive Impact) than they do fast four-way mix or layered projectile setplay. Characters are less likely to have truly oppressive mix compared to older games, where you often had to deal with setups that forced you through a couple layers of reaction and knowledge checks before you could earn the privilege of guessing your way out. (Indeed, character-specific stuff in general, like hurtbox variance, wakeup times, character weights, etc. have been de-emphasized to make combat more consistent and reliable, but also less nuanced and differentiated from matchup to matchup.)
Like the changes to movement and footsies, this presents the defending player with a reasonable possibility of simply guessing their way out, whereas before they likely would have had to lab the mix themselves to be able to practice dealing with it. In the moment, this gives the defending player a clearer idea of what they could do differently to avoid getting hit by that mixup the next time they’re in that situation, and reduces the gameplay value of lab time for both the player doing the mix (since the reward is less consistent) and the player getting mixed (since the skill required to escape the mix is lower). While I’m sure the multi-layered mix sounds like a nightmare to many, the satisfaction of playing my way out of impossibly oppressive setups is directly proportional to the nightmarishness of the mix, to the point where I’ll get annoyed when Xrd Dizzy players don’t execute their setplay properly since it means I can’t do the good thing I practiced.
Perhaps the one place where contemporary fighting games are getting more complicated than their predecessors is in resource management. It seems like everything these days has a meter that governs how much fun you’re allowed to have; SF6’s Drive Meter, Tekken 8’s Heat System, Marvel Tokon’s Assemble Gauge, Melty Blood: Type Lumina’s Moon Gauge, City of the Wolves’s REV Gauge, the list goes on and on. (Even Guilty Gear Strive introduced additional resource management by giving the Burst Meter other uses, though the Wild Assaults ended up getting rolled back, and I think Xrd still exceeds Strive’s resource complexity thanks to YRC and Blitz Shield.) How on Earth did we end up turning a video game genre known for its in-your-face visceral combat into a real-time bean counting simulation?
From a developer’s perspective, gating power behind resource spends is convenient because you get to give characters powerful tools that can only be used in specific situations, or for a specific number of times in a match. If a combo, for example, is gated solely by an execution skill check, then the player who masters the execution is unequivocally stronger than an identical player who has not. If a combo is gated by a specific situation — say, a certain character matchup, or a counter hit-only combo — then the player’s ability to access that power is also situational and possibly niche enough that it’s not worth mastering until a player is well past the point of diminishing returns. But if that combo is gated by a resource, the developer does not need to worry about tuning or balancing the combo itself because the availability of the resource required to perform it becomes the most significant tuning lever. And with so many games fielding multiple resources in play, like Super Meter and Drive Meter in SF6, players have to be far more cognizant about resource management in any given situation. The question is no longer just “Can you kill off this touch?” but “Can you optimize for damage/corner carry/knockdown off this touch with any given combination of resources?” and each player has to be constantly monitoring their resource spend in the background to ensure that they’re getting the optimal value for each play.
The increased emphasis on resource management as a complicating factor that overlays the entirety of the fighting game does a lot to restore depth lost from the simplification and streamlining of the other historical fighting game depth sources listed above, and in a manner that is more explicit and apparent to newer players than stuff like movement, spacing, and oppressive mix. It’s also a game design language that non-fighting game players tend to understand more easily. But it does add emphasis to yet another kind of math to the mix, rewarding the players who are down for real-time spreadsheet management and taxing the players who are not. For some, the new-and-improved meter management mini-game is a welcome change to further reduce the fighting game genre’s emphasis on physicality and practice, and orient its skill checks closer to other comparable competitive game genres; for others, it feels like it takes away from the parts of a fighting game that feel most like a fight. The physical element of fighting games brings the genre a beautiful parallel to the actual practice of combat sports that few other genres can claim, but most people who buy video games aren’t looking for a physical contest to begin with, so de-emphasizing it in favor of more easily readable decisions — strike or throw, spend or save, etc. — brings the genre into a zone more comfortable to people who play other kinds of competitive games and thereby opens up the potential market to a broader set of buyers.
From a craft perspective, though, it is pretty cool to see a lot of experimentation here in some of the more mainstream fighting game series, since for a while it felt like most of the innovation here was coming from 2000s/2010s-era French Bread and Arc System Works. UNI is probably the single most interesting fighting game to come out of the last fifteen years or so, and a lot of the modern meter stuff feels like it takes inspiration from GRD. So while the meter stuff might feel a little heavy handed with this generation, I do have hope that it’ll continue to be refined in future generations.
Designing for “deep enough”
This essay has thus far been a relatively dispassionate (for me, anyway) description of fighting game depth; now, I will editorialize a little bit. As someone who loves old fighting games, only plays old fighting games, and spent his 30s working on a “new” fighting game, I am cursed to understand the reason behind each generation’s combat design trends even though they mostly yield games I don’t really want to play. I like to say that if you came up on the fighting games from the late ’90s into the mid-’00s — Melee, CvS2, Guilty Gear XX, 3S, MvC2, KOF2K2, Tekken 5.1, Virtua Fighter 4 and 5, etc. — then you arrived in time to fall in love with a genre that had just recognized its true power, only to get spooked and turn away. When a game’s possibility space exceeds its developers’ ability to understand it, it is the players who are entrusted with discovering the truth of the game, and that game might end up being not very good, which is a huge risk for anyone who funded the game and needs to make its money back. In order to make more consistently sellable products, fighting game developers learned to build games that operated within a narrower set of more clearly defined skill checks so that the range of player outcomes was far more predictable and knowable upon release, and thus easier to ensure that the game is behaving as intended.
For example: I mentioned earlier that for a developer, gating power behind resource management instead of other skill checks is convenient because it lets you design for a wholly knowable budget; it’s far easier to tune a resource-gated tool for power since you control how the meter is gained, and you know the upper and lower bounds for how much meter a player can access in a single game. If a player can do sixteen Drive Rushes in a single combo, that isn’t something earned by the player for some specific kind of excellence, it’s a bug in the meter system that is readily fixable.
On the other hand, designing combat around difficult execution checks and complicated interconnected system mechanics is fundamentally a bet against humanity’s dogged determination to be better than you thought they could be. It is one thing to know that the Daigo Parry or a Sako Combo are theoretically doable in a game, but tuning around those checks is to ask yourself how far you want to allow your players to go in service of victory. Access to Drive Rush is limited by the upper bound of obtainable Drive Meter in a game, but access to a one-frame link is limited only by your ability to execute that link when it arises, and the upper bound is established by the players who play it — the players whom game developers are inevitably going to underestimate. Everyone who knows Third Strike knew that Chun-Li’s SA2 is parryable, but no one knew for sure that people could do it while one tick of chip damage away from being knocked out of a tournament in front of a cheering crowd, and then punishable with a combo that takes advantage of an obscure property of neutral jump parry attacks, until Daigo showed us all at Cal Poly Pomona in 2004.
As a player and developer, I think there is a qualitative difference between the experience of playing games that are built around a clearly defined sandbox, and the games that are built (usually unintentionally) to allow humans to be as good as humans can possibly be; the former feels like I’m trying to have fun playing in the deliberately curated sandbox that the developer made for me, and the latter feels like the developer is handing me some tools and eagerly watching to see what kinds of wonderful, awful things I can do with them. When the overall possibility space of a game is clear and obvious from the outset, there are fewer moments of wonder, no room to marvel at the things we might never have imagined people can do, because there’s nothing left to the imagination. Building a fighting game to be competitively interesting for a console generation, maybe jazzed up with a patch and a new character every now and then, until people have had enough and are ready to move on to the next one feels predictable and disposable in a way I find personally unpleasant as someone blessed to get into fighting games in a time of games that can be played forever. To me, the way to respect your players’ time is to give them a game worth spending their lives on, a game they can make a part of themselves, rather than something to be set aside when the next one comes out.
Unfortunately, selling someone a game that can be played forever isn’t a great business model. There is no question that this is the correct approach from a sales and popularity perspective; statistically speaking, most of the people reading this do not align with my tastes. (Otherwise I wouldn’t have had to write this.) As I mentioned earlier, too much depth can make it harder for people to have enjoyable matches across large skill gaps, and most people aren’t looking for a game to play forever, they’re looking for a game to play for a couple dozen hours at most. But the fighting games I love feel like a challenge to its players to become so good that the game feels like it cannot contain us. I like it when the developers underestimate the player than overestimate us, even if the latter is more accurate, because the former gives the players room to excel beyond expectations and the latter simply feels like a test of whether the developers’ work performed appropriately.
However, the nice thing about forever games is that you don’t need to make that many of them, and even if we only get one every couple generations or so, that’s still a decent yield in the long run. But I encourage you to pick up a forever game if you don’t have one, and let it serve as a grounding element to save you from the highs and lows of whatever new stuff you’re playing. Vampire Savior and Super Turbo are on the younger side of Millennial, Marvel vs. Capcom 3 is almost old enough to drive, Fist of the North Star old enough to drink. And wherever your personal tastes lie in fighting games, I hope this essay will let you argue about them more deliberately.
Thanks for reading!
🙏,
PM
