The Smooth Ball analogy (fighting game balance and creativity)

I’ve been around the FGC for a long time and have seen lots of different types of game balance across a multitude of different fighting games, many of which will target different styles (rushdown vs defense heavy, air vs ground heavy, etc). I was heavily invested in fighting games back when I knew patches would never happen, and I played and studied intensely many of the games that brought the FGC into the internet era, where patches are not only necessary, but expected. I’ve been around people who think you should only buff the low tier and never nerf the top tier, as well as players who think nerfs for any strong strategy should have happened yesterday.

I want to bring all these ideas together into an analogy, which I will call the Smooth Ball analogy. Maybe it will give you some perspective on why fighting games can be fun even though they aren’t balanced, and why balance in games can be thought of in many different ways.

Imagine the summation of all tools for all characters in a fighting game can be mapped to the contours of a ball. These are the tools you will be using to damage your opponent’s character, to resist damage being thrown at you, and generally to add spice and flavor to your game so that players want to pick it up and play it. Are you picturing this ball in your mind? Okay, good.

What does this ball look like for your theoretically perfect fighting game? Before reading the rest of this post, give it 20 seconds of thought, and come to a conclusion.

Lots of people will say that the ball should be perfectly smooth and shiny, like a marble. I think this is probably most people’s first thought about what a “perfect” ball looks like. This is very aesthetically pleasing and “pure” in a mathematical sense, so you might think that a designer’s goal is to create a game that is close to a perfect sphere as possible. No sharp edges anywhere (wouldn’t want anyone to get hurt). We can roll it on a table in a straight line and it will reach its destination super predictably. A paragon of design that will have players playing this fighting game for years to come, knowing all their decisions were projected against a perfect sphere.

That is… until you realize that this ball is incapable of damaging someone. Remember, your goal is to do damage to someone (or prevent against damage) by using this ball, and unless you throw it really hard at someone’s forehead, nobody’s going to be getting hurt from a perfectly round metal ball being rubbed up against their skin. You might look at the ball and say “that sure is pretty”, but nobody will play this fighting game. No moments of unpredictability can happen as you roll the ball down the table. Every match feels and looks the same, no matter what angle you hold it.

So, okay, the designer decides to weaponize this ball a bit. We’re playing a fighting game, let’s add the capability to hurt someone. They add a small imperfection to one side of the ball so that if you rub someone’s skin against it, they might become mildly irritated. The clear strategy is to now use this one added tool, since it is the only thing that can hurt someone, but you don’t get too much mileage out of this.

Fair enough, let’s add a few more imperfections. Heck, let’s take one side of the ball and add a spike to it, that’s gonna hurt someone. But like before, we can’t just add one spike, or else it becomes the best strategy. Let’s add one spike to each area (ie, each character) of the ball. Let’s make it the same height, though, so that no matter which side you pick the ball up, you can stab someone the same way. Again… we haven’t really improved our situation, but at least one thing is for sure; our ball is becoming farther and farther away from the perfect sphere we started with.

Let’s cut to the chase here. After playing around with this for a while, you start to realize; let’s just add tons and tons of spikes all over the ball, of varying heights and angles. Now, no matter how you pick up the ball, you will be able to find a small area of the ball that can really hurt your opponent, often in unpredictable ways. They will have to creatively use different sides of their ball to shield themselves from this damage. This can clearly go too far, though. If you add the spikes “randomly”, you run into the risk of introducing one particularly dangerous area that has such a sharp spike that no other contour of the ball can defend against it.

So again, I ask the question… what would the ball for your theoretically perfect fighting game look like?

To me, this answer lies at the crux of both fighting game balance, and fighting game character design (ie, creativity you allow within the game engine). Think of fighting game balance as the distribution of spikes around the ball, and fighting game creativity as the length and number of the spikes you give the players.

A perfectly smooth ball has no spikes at all; this game might be perfectly balanced (equal distribution), but has virtually no creativity (height of spikes is 0). To be blunt, this is a crappy game. It might seem philosophically pleasing, but there is nothing worth exploring here. The ball’s function has been decided by the designers and there is nothing to discover.

Let’s talk about a game like SFV; I think it is kind of golf-ball esque (if the dimples on a golf ball were protrusions instead of indents). Decently balanced, but very little creativity allowed. And here is an important part; when the ball is still close to smooth, design imperfections are all the more obvious. If a character like Balrog has overwhelming mixup spikes, or Guile has overwhelming space control spikes, other characters are immediately frustrated because it is so blatantly clear that there are no imperfections on their side of the ball that they can use.

Important Statement #1: Games that aim for smooth balance often tend to provide the opposite, because as soon as the smallest unintended or undesigned imperfection is found, there is nothing left to counteract it; the rest of the sphere has been polished to a shine. One or two small mistakes can throw the balance of the ball so wildly off and no amount of trying to roll the ball in a straight line will fix this. It’s why, I feel, so many people are frustrated at SFV’s balance of the top tiers. It’s actually not really that badly balanced of a game (ie, many characters contain a somewhat powerful spike), but the length of these spikes is disproportionate, and most importantly, the number of spikes is so low that you don’t have to look too long at your own spike to see that it’s not going to win in a measurement contest.

Let’s take a look at a game like… KI Season 2. This is a game that, I think most will agree, had hundreds of spikes sticking out all over the place. And when you have a spiky ball, doing damage to your opponent is often extremely easy (and fun!). If one spike starts to fail, no problem, you have 50 other spikes of various lengths and sizes to check out. So is the solution just make your ball super spiky? I think to some extent, yes, but it’s not the full solution, since powerful spikes can still emerge from this ball and become degenerate, both on offense and defense.

Important Statement #2: Just buffing the low tier while keeping the top tier identical is not a viable strategy for game balance; the goal needs to be to limit the effectiveness of the powerful spikes. And often it’s not much, it can be just shaving a little bit off the tip while keeping the spike in its same place and orientation.

Important Statement #3: Fighting game balance is harder to decide the spikier your ball is. To put it another way, when you give characters many tools to overcome other character’s many tools, sometimes it takes years to decide which area of the ball is the spikiest. But most importantly, it can actually cover for some bad design decisions. Marvel 2 is a broken game, but because its ball was so spiky, we ended up with all sorts of glitches, strategies, and flavors of the eventual top tier characters that people played it for 10 years. Marvel 3’s most dominant current player uses point Chun-Li, a character deemed incredibly outmatched by better characters at the game’s launch, but RyanLV found an underexplored topology of the ball, because it was possible to do that. Vergil is probably too good of a character (that is, it was a bad design decision), but there are enough spikes on the ball for other characters to mitigate that bad design decision in the long term.

Now, let’s compare this to KI Season 3, a game which I believe is moving closer to what I would call good fighting game balance; a well-distributed but very spiky ball. Some of the sharpest edges from S2 were shaved, but because KI’s design philosophy was to fill the ball with spikes, there are still plenty of spikes to go check out. Some players are mad that their favorite spikes were shaved down, while their most disliked spikes for their opponents were untouched; one approach is to get mad at the designers, but the real answer here is to look for new areas on the ball with spikes people haven’t checked out yet. And I will confidently say that KI is the most underexplored fighting game I have ever seen that saw at least 3 years of tournament play. There is LOTS we haven’t checked out yet.

But the most important thing is that we CAN check it out, because the KI ball is spiky. Using a spiky weapon that might hurt you just as much as the opponent is a skill that the best fighting game players should love to be measured by. And I worry that as people incorrectly measure the distribution and length of the spikes (because, say, they are only concerned with a local maximum rather than a global maximum), and demand that the ball be polished and buffed to a shine to eliminate its imperfections, we will end up with not only a game not worth playing, but worse… a game whose slightly overlooked imperfections become glaringly obvious, like a fingerprint on the back of your new iPhone. A game so carefully designed to be pure that its one area of impurity can only be the sole focus of your attention.

I would encourage you to think about this the next time you hear someone throw the word “yolo” or “cheap” around, or when someone says they want to nerf all the top tier or buff all the low tier, or that something “shouldn’t be in a game.” Are they saying this because the spike is egregiously long and disproportionate to the rest of the ball? Are they zoomed in too close to a local maximum? Are they trying to polish the ball into something that will only show its imperfections faster and leave you with less creativity to overcome the designer’s mistake?


Very interesting, and well thought out theory. It’s interesting how you used a spikey ball as a way of explaining things.

What about in the case that the ball is TOO Spikey? Doesn’t that hurt the game as well? So there has to be a happy medium for balance and such.

1 Like

Yes, this is the crux of FG balance, why it is so effing hard, and why I think most people who think they would do a good job at balance actually wouldn’t (typically because they are looking only at local maxima for their characters, or they have vendettas against certain sides of the ball).

You can have a smooth ball with a small number of glaring imperfections, or place 5000 spikes of random length all over your ball and dare your playerbase to sift through it all and find the best stuff. What typically happens in this last case, though, is that they’ll find a “suitably long” spike and just ride it out because the optimization problem is so difficult that they don’t need to do any better than that; only if the game somehow gets played for 15 years, like for example a game like Melee, will the top players be forced to optimize the spiky ball to the greatest possible effect. Most games don’t last that long.

The “best” possible fighting game is probably a game that has lots of spikes, but a good distribution of them and none of them too overly long. Asking players to optimize this is still very difficult but relatively achievable. Think of a kush ball.

The point I wanted to make with the top post, though, is that the perfect fighting game is decidedly not a perfect sphere, like I think many people believe - some objectively pure game where all decisions are matched against some mathematical perfection [of footsies], there are no 50/50s or guessing [but somehow you still can do damage], etc.

The best games have lots of grime, dirt, and creativity. And because designers are human, the best fighting games often have a couple long spikes to deal with. But the philosophy should NOT change to “let’s remove the spikes” because there are a few abnormally long ones.


I feel as this just doesn’t apply to FG’s but all multiplayer games in general.

1 Like

I’ve personally talked about the balance in KI with Max, and our overall conclusion was that a perfectly balanced fighting game would be boring. A fighting game with nothing but 5-5 matchups across the board would not only be boring, but impossible to do in a game like KI. The aim of fighting game balance isn’t to reach a perfectly balanced game either, otherwise characters become stale, which is what happened during the SFV betas up to launch.

Fighting games need to have high, mid, and low tiers in order to be interesting, especially from a spectator standpoint. How hype was it watching Letalis make it as far as he did at KIWC with Aganos and RAAM, characters that are considered low tier? If you want a fighting game to be interesting at all there needs to be characters with interesting tools, and some characters will just inherently be stronger because of those tools.

Infil should get a Nobel Prize for Video Game Science.
Wait. That isn’t a category for a Nobel Prize? Well make it one, damnit! @Infilament deserves a prize!

On a related note, I initially thought of something completely different before I finished reading.
See, I saw the title of Smooth Ball, and had the idea that the smoothness of the ball CAME FROM the different types of characters.
The idea was that every character would start at the center, in the core of the globe. Then each character would move out in different directions from the other so we’d have characters opposite eachother in style on other sides of the globe. This would also include putting characters opposite of the characters they have good and bad matchups with.
The idea was that eventually each character would be opposite of the characters they were good or bad against.

My idea wasn’t as smart or well explained as yours, so thanks for that :sweat_smile:

1 Like

Interesting…very insightful, but I wouldn’t expect any less from you Infil.

I don’t know if this will help out as well, but I do know this as the nature of spheres… if you were to theoretically saturate the surface of the ball with spikes, you could in theory be right back to having a smooth sphere again, just a larger one than the initial sphere with no spikes.
I suppose in game term you could say too many tools to help fill in weaknesses and you could also be right back to having that perfectly balanced snooze fest.

this analogy is unbelievably smart, i love it.

and yeah, i think that when it comes to what the perfect ball is, i think KI is probably the closest i’ve seen a game hit it in a while, next to a beast of a game like Tekken.

so a thing i’m thinking is: you can approach each separate game like its own smooth/spiked ball, and understand something about each game by its singular ball texture. what if we took this analogy and used it for the sake of balancing an individual game, where each character is represented as a different ball? rather than KI being represented as a spiked ball, it is instead more like a shelf that holds 27-30 balls of varying textures and color – texture representing that character’s “gimmicks” or advantages, and the color of the ball representing the type of character they are (rushdown, zoning, mix-up, etc…)

this might be going off into a weird tangent, but i could also see balance, in terms of character usage, working a bit like bowling balls. individual bowling balls have their own weight, size, & finger grip style (the character), and of course there’s always the person rolling the ball that factors into the effectiveness of the shot (the player).

if your ball is too light, you’ll have a harder time knocking stuff down, but you might hit things more reliably even if they don’t get knocked down as hard. (underpowered)
if your ball is too heavy, you might have a hard time actually using the ball, but the ball itself might knock stuff down more effectively without you needing to work as hard. (UP, in a different way)
fast-forward a few months, and someone brings their shiny new state of the art bowling ball that can physically split into two half-balls in order to easily deal with 7-10 splits.

obviously, at first glance, this is overpowered AF.
so there’s two things to do here: you weaken the overpowered ball and bring it more in line with a standard bowling ball (which is balanced, but isn’t as interesting), OR you can go the other route and give your ball something else crazy that it does to make it more effective, which risks making the entire game Marvel 2 Bowling (a LOT of fun, but maybe difficult to learn, and leaves you asking a lot of questions).
additionally, even if the other ball is overpowered as hell, if it’s not so overpowered that the better player with the regular bowling ball can still win, then is the “OP” bowling ball really so bad, as much as it is just Different?

in the end, the thing that really matters is what you want to play, right? if the game is Neo Bowling 3080 X, and you want to play Neo Bowling 3080 X, then you balance the game around it being that. if the game is NB3080X and you’re afraid of it being too crazy so you balance it like regular bowling, then it just becomes regular bowling.

… i think i lost myself, but yeah, great thread lol.

The math term for this is “isosurface” in case you were curious. :slight_smile:

Actually I don’t think having all your spikes end up roughly the same length is that bad for balance, because the decisions to make within the game end up being the juice that prevents it from being boring; you still have many spikes to choose from at any given time and they interact in interesting ways with each other. I think that’s actually probably the “perfect” FG… many, tall spikes that can still somehow form a perfect sphere at the tip. It’s just not going to be practical to make such a game, so some spikes will be proportionately too big as a necessity of humans making a creative product.

1 Like

It’s a great analogy and a terrific discussion. Fair warning - I’m going to indulge in a pretty long ramble that may only be interesting to me.

Everybody probably has a slightly different idea of what the ball represents - for me my first thought was that the ball was more like a globe representing the whole game, serving as a map of the characters and their abilities. A perfectly smooth ball would have no differentiation between characters, which makes it pretty obvious (at least to me) that the game would be boring. You might also imagine a ball for each character.

And of course we can all create perfectly balanced games in our heads. Just imagine a game with one character. These games have existed, starting with Karate Champ you could play competetive fighting games where there was zero difference between characters. Even SF 1 you could only pick to play as Ken and Ryu. The fight was perfectly balanced. And yes, you can have fun for hours playing mirror matches. But does anyone believe the FGC would benefit from reduced character variety? There’s no variety unless you start roughing up the surface of the ball.

Another thing you can play with in your mind is the size of the ball. Are you holding it in your hand and able to freely examine it, or is it huge like an asteroid or a planet? I’m a biologist and I tend to think of fighting game mechanics like genetics or cell and molecular biology. You can see the outcomes of what you are talking about but you often can’t see the direct interactions that lead to the outcomes. So you have to make up a model in your mind (like the ball analogy). The thing about the model is it’s certainly not precisely correct. It may just be adequate to explain what you are seeing. But you might later see something else that can’t be explained by your model. So over time you learn more and have to adjust your model for new observations. If you look at the balance discussions and buff/nerf arguments they are really just people arguing over their different models of how the game works. Some people will defend a model even long after it’s been demonstrated that the model doesn’t adequately explain what’s happening. The point is, balancing the game is even more challenging because you don’t ever have a clear and direct “perfect” picture of what the ball actually looks like. And even if you do, it’s tough to convince people that your model of the ball is the correct one.

These kind of analogies also help you (or me, at least) think about what part of the game is fun for you. My ideal play space is where neither me nor my opponent has a perfect understanding of the ball. What’s fun is challenging the other person’s model of how the game works and having them challeng yours and then seeing if you have it right or more right than they do. Adjusting the model as you learn more (i.e. Improving). And of course you can have a model of how the game works and then adopt different strategies built around the same model. All if this highlights that I think for most of us, the game has to be complex enough that it isn’t possible to hold it in your mind completely - otherwise we have a perfect model and we know what to do all the time to win or avoid losing (tic-tac-toe). Think of the arcade game Joust for example. You could call that a one on one fighter with very simple rules - be higher than your opponent when you collide on screen. Anyone can understand that game perfectly. But it was made more challenging by the controls, which were floaty and unintuitive. That’s one way to make a game interesting - with an execution barrier. I’ll leave that thought for now to spare you another wall of text.

I’m getting lost in my own head, but to bring this back to KI I think the games development cycle has required people to constantly adjust their models as characters and mechanics are added, patches etc. For someone like me who doesn’t ever expect to have the game “solved” and in fact would prefer to play an unsolved game, this works out just fine. I think for most players below the highest level this is probably true. But you can see the challenge it has made for the higher level tournament players. They generally have a preference for a stable, “solved” game where they can then develop their own optimal stratagey and execute it consistently. For those of us who expect and even enjoy seeing new things every time we play the game (and who are familiar with being told “adapt and get better” and work at doing this when we encounter something that disrupts our model) it just seems like the pros are not following their own advice. But if you think about the time and energy they have put in to fine tuning their model of the game and then accept that someone deliberately changed the game to make it different from their good and accurate model (rather than just discovering something they didn’t know before) you can see why that might throw them for a loop - even though they follow the “level up” mentality of improving the model as you learn the game. So thinking about it this way, it’s not entirely fair to just tell them to “deal,” like the rest of us do.

TL;DR thanks for the great analogy. I’ve really come to enjoy trying to think about these theories for games - despite plateauing as an intermediate level player when I actually play the game.


A wonderful anology their infil. :clap: :clap: :clap: :clap: :clap: :clap: :clap: :clap: :clap: :clap:

Fighting gamr balancing is very difficult to do. I always beleived it was soley behind the fact that you have to understand what this character’s tools are vs everyone elses tools. Their strength and weaknesses vs the others’ strength and weaknesses.

However a game that has wonderful mechanics and things like in KI’s case; No matter who you pick you can certainly win because you can overcome the weaknesses of your characters.

However another metaphor I think would also fit this is people have their own personal spheres , with various spikes, blunts, and smooth areas. This ball represent the player’s skill. If your opponnet’s ball is to rough for you, you can use the sides you have to defend against it, Or you can shape your ball in the case of a strategy to deal with that opponent’s ball.

In KI’s case this where all the different strengths of your combo-hits, counter-breakers, juggles, and moves your characters have allow you to do this.

1 Like

I got nothing to add, i just wanted to leave a little more than a simple like to this theory and discussion. Carry on.

Thats avery good theory. Sadly I’ve got nothing to add to at the moment. still this is a good read.

Heheheh… smooth balls…


I’ve got a rep 19 character in For Honor which I’m grinding for a bit of mindless fun. Why is it mindless? Because in ball terms (lol) each hero has about 1 or maybe 2 spikes that stick out and that’s it. There’s a base spikiness of general offensive options which all heroes share, but each one is defined by one overwhelmingly good spike, e.g. Warden shoulder vortex, valk sweep unblockable, shugo’s charge, etc. Hence, it’s boring.

Lol. Yeah I think for games that aren’t one v One fighters there are some additional elements. The environment, the level design, the presence of teammates and the way you work together etc.

As a professor of mine liked to say in grad school - “all models are wrong; some models are useful.” :slight_smile: I think you’re probably very right on the mindset of some higher-level players. They’ve found a model of the ball that they think accurately describes the game (in truth I think they’ve probably mostly found a “good enough” spike orientation that Infil mentioned), and it hurts to have that model proven (or made) invalid through character additions or balance changes. But the model was never “right” to begin with - all models of complex systems are inherently wrong; the only question is if the model is useful. Understanding that and adjusting and iterating your model to make it more useful is a huge part of what makes the game fun in the first place.

Sidenote: I really, really love these FG theory conversations. Thanks for the topic @Infilament :smile:


Yes, which is why the 4v4 modes are more interesting as they present a set of constantly changing circumstances: hostile environment, NPCs (captains/soldiers), ledges, outnumbered situations both in favour and against, etc.

1 Like

Fantastic, fantastic analogy and analysis

Hey @Infilament , I’ve got a question:
In regards to this analogy, what would it be one if the spikes… started developing spikes?
My guess is something along the lines of finding a way to use something not as intended that actually becomes a pretty big part of the meta, like Quick-Scoping, Strafe Running, Rocket Jumping, etc.
Although sometimes these extra spikes on other spikes may block the spikes around them and make them less lethal (or as a tangible example, these meta tactics may render something designed for that purpose obsolete even though it uses a character/weapon who wasn’t originally purposed for that role).

I’d like to hear your thoughts on this, if that’s alright.