A bunch of different videos have popped up lately in fighting game circles, about “Emergent Gameplay”. I’ve watched a number of them and they’re grasping at concepts they can’t totally describe. They use a lot of vague terminology and almost say what they want to, but not quite. The gist is, old games had Emergent Gameplay, new games don’t, but why?
This video by Brian F kicked off the conversation, featuring Kayin Nasaki’s tweet on the subject, using the word “Depth” in the title, and “Emergent Gameplay” in the thumbnail, and the script of the video.
It was followed up by Maximillian Dood’s video here:
And LordKnight commented on Max’s video here, from more of a competitor’s perspective:
Emergent gameplay is the wrong term here, because everything not specifically enumerated by the developers is emergent (props to Brian F for pointing this out, even though he immediately jumps to Emergent = Unintended). While unintended gameplay can be a hallmark of fun games, a game doesn’t need to have unintended tricks to be fun. Everything that emerges unintended in a game could also be deliberately designed and balanced by an aware game developer (this is the story of many features in Skullgirls and Project+ for example).
The concept that is being danced around here is “depth”. Every time Max says “loose” he means “deep” or more specifically “nuanced” in my current terminology. Allowing a move to be canceled at any time gives that move nuance. Allowing a move to lead into a lot of other moves gives that move nuance. These things create additional game states, creating more possibility space in a game. There are more raw states. Literally more things can happen. As more variables can be carried over through this canceling process, such as inheriting velocity, invincibility frames, hitstun, etc, the more things can happen in a game. What’s important isn’t whether those things are intended or not, it’s how many things you can do.
Allowing you to use more moves as assists creates a wider possibility space. Allowing you to use meter on a bunch of different things creates a wider possibility space. Having a lot of different moves creates a wider possibility space. The possibility space of a game gets bigger as there is more more more. More moves, more cancels, more techniques, more system mechanics, etc.
LordKnight’s Take & what it means for depth
LordKnight is approaching this from a competitor’s mindset, as well as a more experienced player’s mindset. As a competitor, it doesn’t matter what’s in the game; it’s your job to accept it and figure out how to win, and not to complain. Competitors don’t care if the game is deep or not, they only care about how to use the tools at their disposal to win.
As an experienced player, LordKnight is also tempering Max’s takes by pointing out that a lot of the depth of a game isn’t known when the game comes out. It can be easy to dismiss a game as lacking depth on day 1 because people don’t really understand it, which explains the cycle of complaining about every new game as lacking depth, because we’re comparing a game at the end of its lifecycle to a game that just started theirs, and 5 years down the line there will be a bunch of new techniques that we didn’t initially know about. Even though these possibilities might exist, they’re not actually depth until people discover them. Further, just because there might be a lot of possibilities, they might not always be useful. You might be able to cancel a move at any point, but does that actually add interesting decisions to the game, or is it just an additional combo option, which means you (or more likely some better player) spends more time in the lab for a bigger combo, and then you do that until the heatdeath of the universe? This is what I call Relevant Depth, the possibility space that is actually represented in a game. Relevant depth filters the possibility space for states that players actually know about and can perform, as well as the states that players consider viable to make use of.
The other thing LordKnight indirectly points out is that even though the possibility space of a game might get bigger as there is more stuff, the game might not actually get more deeper. For example, in games like DBFZ, you might be able to cancel everything, but it all leads into the same routes regardless, and you can convert off of anything. This means the differentiation between moves gets weaker, because from a given move you have most of the same followups, so moves are only differentiated in terms of what they do in neutral, not as much in what combo routes they allow or whether they allow combo routes at all. This is what I call Differentiated Depth, or Absolute Depth, the amount of depth left in the game after you filter the possibility space for all the states that are actually distinct from one another and judge their relative level of distinction.
Once you filter a game’s possibility space for the relevant depth, and the differentiated depth, you have what I just call Depth. As LordKnight points out, you can’t really know the true depth of a game when it comes out, because it gets deeper (or shallower) as players make more discoveries, and you can’t just call every decision branch in a game depth, because many of those possibilities are deleterious to other possibilities, either through overlapping in differentiation, or stifling the relevance of other decisions. This is why the filters of Relevance and Differentiation are important, because pure volume of stuff and nuance alone won’t tell you everything. Possibility space, filtered through Relevance and Differentiation is the Effective Complexity of the game, how complex it is to play for actual humans.
New Games Are Limited Though
However I think Kayin/Brian/Maximillian have a point! A lot of newer games are designed to specifically limit the range of possibilities from point of conception, and that might be irrelevant to a competitor, who is focused purely on how to use the system to win, instead of caring about the relative complexity of the system, but it’s still important! Street Fighter 4 and 5 were designed with Juggle Potential to make it so combos are guaranteed to be impossible to follow up after a certain point. Every air hit builds juggle points, and every hit of every move has a juggle potential, that won’t allow it to make contact if you’ve exceeded the number of points in its potential. So a move with 2 juggle potential can’t be used on someone who’s taken 3 air hits already, which limits which moves can be used to follow up juggle combos and for how long.
In 3rd Strike, there’s no juggle potential system, just juggle points, so after about 6 juggle points, you won’t be allowed further followup, but you can hit with any move that can make contact, which widens the range of possibilities for combos in 3rd strike. Anime fighters use similarly fuzzy systems of untech timers and hitstun scaling to limit combos, and Blazblue and DBFZ have a hard combo timer of 10 seconds. This can be a stylistic choice, creating a different feeling between different games and thus being more a matter of preference than right vs wrong, but the challenge of game designers is also to create differentiation in these systems so that players make different choices in which moves they use in neutral, and which combo routes they take.
Street Fighter V was more conscious of the frame advantage of each attack than street fighter 4, and limited which moves had enough frame advantage on hit to actually combo into other moves. In Street Fighter 4, many moves had enough frame advantage to combo into each other, so you had a lot of different routes for linking moves together, and SFV tended to pick a couple specific normal moves to link into each other on each character. Street Fighter V went further to limit the range of possibilities in the game by doing things like only letting you use air specials from forward jumps, limiting the minimum air height before they become possible, having limb hurtboxes disappear on move recoveries, rather than track the animation more closely, preventing chip damage from killing with anything but a super attack, and limiting super attacks to only be invincible cinematic moves, rather than having install supers or other more nuanced or utility super moves (which is what most V-triggers ended up becoming).
Older games also had more flexible damage scaling systems, tending to scale per-hit rather than per-move, meaning multi-hit attacks would scale your combo damage more, but also they needed to make the scaling weaker to keep multi-hit moves balanced with single hit ones, otherwise you’d scale yourself out of all your damage after doing a single multi-hit. Per-move combo scaling, as implemented in SF4 meant they were free to nerf your damage into oblivion after a few hits, meaning all the combos possible in the game could only realistically deal a certain maximum amount of damage in SF4 and SFV. They also sometimes clamped combo damage to a minimum of 1, such as SF Alpha 2 and CVS2, so light hits could tack on a lot of damage at the end of a custom combo. Blazblue has a system that lops 40% off your combo damage for everything after the first hit, but doesn’t scale that much per-move afterwards, so single hits can be proportionally strong, while still allowing long combos to deal a lot of damage.
Why bother with depth?
So why do these games have juggle potential and damage scaling at all? Why not just design moves to deal as much damage as they’d like those moves to deal and not deal with all this combo crap? 3d Fighters and Mortal Kombat have strings that deal set amounts of damage, why not simply do that and not allow for juggles or cancels or any of this other stuff? The purpose is to give the player more things to learn about, and more possibilities to explore, because on a basic level, learning about a game is what makes it fun. Fighting Game Developers could very easily just decide, “Okay, these are the only combos you can do” and that’s it. They could have each combo deal a completely arbitrary amount of damage, and then they wouldn’t need to implement all these other systems that limit how long a combo is, or how much damage it does. Instead, fighting game developers create a set of rules that allow for a large number of possibilities, and since this system is governed by rules rather than enumerated by specifically designed states, there may be possibilities in it that the developers didn’t foresee.
This isn’t inherently a good or bad thing, it’s just an inevitability of creating a system of rules. And the volume of things players can learn about a game is governed by how big the possibility space created from these rules is, hence depth. The challenge for developers is to create a set of rules that will allow a large number of things, while making them as different from one another as possible, without allowing some things to be a lot more powerful than the others, and to that end, they frequently implement limitations ahead of time so that the unintended consequences of their rulesets will never get too far out of hand.
The frustration being expressed in these videos is that in the developer’s quest for balance, to avoid anything being too broken, they implement limitations that are very restrictive of the possibility space. Experienced Players frequently express this as saying modern games aren’t designed to be too broken, which is why we get discourse over and over on balanced vs broken. A game with a wider possibility space is a game that is more prone to broken shit happening, because it’s harder for developers to foresee every possibility and safeguard against it.
Broken old games
Older games were designed with fewer of these safeguards, so we see a lot more things happen in them that would never happen in a modern game. Therefore older games tend to have worse balance, and newer games tend to have more restrictive combo and scaling systems to ensure that even if things happen the developers didn’t predict, that it won’t upset the game balance too much, and to ensure that the possibility space of the game is limited enough that they can foresee problems before they happen.
But the cost of this safeguarding is that many people perceive the style of newer fighting games as being less fun. They don’t have the words to articulate this, so it ends up as the dumb discourse of, “fighting game devs should make games broken instead of balanced”, when the real target is depth (and a deep game can be balanced, while imbalance can cost a game a lot of its depth). Players perceive that older games are frequently deeper, which coincides with brokenness, so they think that what they want is the brokenness, when really they want the depth.
And of course, high level players are fine with things being broken, since they’ll just pick whatever’s most broken and try their hardest to win. Improving at a game is always an eternal process of discovery, because even with the simplest competitive games, you can always improve your skills or learn new things, even if the relative scale of depth is a lot lower than more complex games. Chess grandmasters still improve all the time, even if Go is objectively the deeper game. For players more concerned with fun than pure competition, there are differences of degrees. Improvement in more shallow games is less about finding new discrete techniques, and more about squeezing water from a stone, which can be boring.
Some games will have magnitudes more depth, more possibility space, than others, and even if you can’t predict how deep a game will be in 5 year’s time, you can frequently take a good guess based on the range of systems available today. It’s an easy excuse to say that a game will be more deep as more things are found for it in the future, but the number of things found for a game like Super Smash Bros Melee dwarfs that of its sequels and predecessor, Smash 64. Not all games develop the same over time, and while it pays to have some reservations before judging a game, it’s not entirely unpredictable how a game will develop in the future.
Ideally, fighting game developers can thread the needle between creating a large possibility space by creating a large number of characters, moves, and system mechanics that each have a lot of nuance and possible interactions with each other. Ideally they can make all of these elements distinct from one another, without allowing broken shit to peek through and make it all irrelevant. Of course, this problem, like a Gordian Knot, is incredibly tough to tackle, but I’d hope developers give it a shot instead of continuing to limit the scope of games.
This also brings up a question of governance, by the game developers. How will a game developer treat unintended consequences of their rules in future patches or versions of a game? In other video game genres, the typical answer is to always patch out any bug. Bugs are bad by default and anything unintended should not exist. Ideally a good developer will not have such a black-and-white view, because the whole reason they made the game is to allow possibilities to exist as a consequence of the rules they’ve created. Fighting game developers tend to recognize that bugs add additional layers of decision-making to games, they can create interesting choices, and new possibilities, and thus add depth to a game.
A good fighting game developer will ideally try to preserve the range of possibilities in their game over time, while not allowing bugs to create possibilities that wreck the game balance, but it’s MUCH easier to simply patch and nerf whatever is causing issues, decreasing the range of possibilities instead of attempting to create an interesting choice out of an overpowered player discovery. Many games have had their casts nerfed to the point of being less deep and fun overall. As Brian F says in his video, Oro’s fast fall tech had nuance, it gave Oro more options, and even if it went against Capcom’s intentions for the character, it made him more deep without making him too strong. Ideally game patching processes should aim to maximize the depth of the game, rather than reify the developer’s intentions. The other issue with patch governance is that sometimes you might need to target something ruining your game right the fuck now, and you might want to wait a year for players to figure things out and let it rock.
A Shorthand for Depth
It can be difficult to think in abstract terms about Possibility Space, Absolute Depth, and Relative depth. I also have a shorthand for these ideas in the form of 4 criteria for evaluating if a mechanic/character/whatever is deep.
These criteria aren’t perfect, because something might be unique and have a ton of uses, but be balanced in a way that makes it worthless; such as Ky’s Greed Sever in older Guilty Gear games. This move can hop lows, hits high, and leads into big damage combos, but it’s reactable and unsafe on block, so at a high level it’s not very useful as a mixup tool, and it’s relegated to beating low attacks. Or a move might be the most deep move in the game, but it’s so good there’s no point in doing anything else.
Ideally a move should be unique enough that it has a reason to exist in a character’s kit and it’s not outperformed by other moves. It should have a variety of different applications, instead of being designed to do 1 thing only. It should have nuance to how you can apply it, such as having different effects based on how you input it (chain timing, how long you hold the button, inherit velocity, etc), or when it hits (like meaties), or what area it hits (like clean hits), or a number of other factors. And it should have synergy with the character’s other moves to form a cohesive whole.
This too is something that’s changed over time. In older games, moves tended to be more multi-faceted. You might have multiple anti-airs that have different tradeoffs between them in terms of difficulty to execute, startup, hitboxes, and recovery. You might have multiple pressure routes or ways to extend pressure, and many modern games compress these types of options into 1 option to rule them all. A move might serve multiple purposes, such as Makoto’s s.MP, which is effective as anti-air, pressure, and whiff punish. Or Urien’s Aegis Reflector, which is pressure, combo extending cancel, unblockable setup, you can combo throw into it, it has a crazy number of connections to your other moves. Of course, older games also leave a massive amount on the cutting room floor, because they’re not really designed for characters to have balanced movesets, or even for the game to have a balanced cast of characters, but what matters is that what’s left over is more than what’s left behind.
Ideally in future discussions of fighting game design, we can avoid vague language like, “loose”, or if we’re going to coin terms, we could at least be more discrete about what they mean than just random examples of stuff we think fits. Ideally we can target the thing we actually want to target, instead of getting hung up on side-effects, like a game being broken or having unintended mechanics. And hopefully we can distinguish from a competitive player’s perspective and more of a game designer’s perspective or a general fan’s perspective.
I would like to see future fighting games target depth more (tbh, the goal of this blog is to help all games target depth more). This means game designers accepting the risk before release that the game might be broken, which means careful patch governance to target time-out infinites with easy starters and not necessarily a lot of other stuff the playerbase might discover. This means game designers having to think a lot harder about their game design as the game evolves, rather than simply patching out everything unintended. We see this same conflict in speedrun communities, where developers are frequently the enemy of speedrunners, continually stunting the routes of various games. Hopefully we’ll see more games in the future embrace depth more instead of attempting to limit what’s possible at the early planning stages.
Thanks for this write-up, both the clear headed analysis and the fact that this is an actual written blog instead of a twitch stream ramble cut-and-pasted to youtube.
One addition I would like to add about “patch culture” that I think is relevant to how players engage and think about these games is how it warps conversations about issues that arise within the games. For instance when Killer Instinct (xbox one) was still receiving patches there was no end to players complaining about various changes that had occurred or changes that ought to occur. This was all while the developers were very open about their balancing (writing detailed patch notes with reasoning for their changes, and doing live streams explaining these things). Once the post-release support for the game was ended and players realizing they could not depend on the developers to make their problems go away suddenly started speaking more positively of the game, and it became tolerable to participate with them.
I see the same dynamic with Starcraft Broodwar vs Starcraft 2 – one game hasn’t had patches in over two decades and everyone who plays it firmly believes it never will, the other has had pretty frequent patching since release and conversations very often return to how the developers ought to fix this or that issue with the balance. More recently Blizzard let SC2 go over a year without a patch and it seemed like the community might start to believe the game was finished (though they would frame it as “abandoned” or “dead”), but recently there was a patch and now we’re back to arguing about how there are too few protosses winning top level tournaments so they need buffs vs there are too many protosses in the highest ranks on ladder so they need nerfs.
While I appreciate patches for balance in games I do think this style of long-term patch culture has a negative effect on competitive players willingness to knuckle down and overcome the various problems their opponents bring to the table. I think it actively stifles productive conversations between players, and it also diminishes advancements in understanding of the game: When we look at old matches from broodwar we note how differently people played then, versus when we look at old matches from SC2 we note how different that patch was. (This seems even more egregiously pronounced in League of Legends, but I mostly concern myself with 1v1 games so this is mostly just assumption)
This is one of the main reasons why I tend to favor older games, and while it does relate to this general idea people have of older games allowing for a certain type of player engagement it usually gets lost by the wayside in favor of people wanting to talk about how ‘roll canceling and wave dashing was not intended by the developers and that’s cool’
If there was more of a focus on how players shape how a game is played versus how patches shape how a game is played I think we would get better conversations, though I have several issues with how discourse happens with regards to competitive games – and fighting games especially – that likely would need to be addressed first, but I have gone on long enough.
Thanks again for posting
LikeLiked by 1 person
Thank you for your praise, I’m happy to add to the conversation!
This article isn’t really about patch culture, but my 2 cents is that if patches are doing their job (improving the balance of the game), then patches should be smaller and less frequent over time.
Have you seen the most recent SC2 patch notes? They’re like, 5 changes. Pretty crazy! Seems like almost nothing.
Old games giving you a consistent base to improve on certainly helps, but old games could certainly use patches sometimes too, like 3rd strike, which really could have used a final balance pass with a decade of hindsight. Having active development certainly does create that toxic culture of “nerf this, change this” and people refusing to deal with the cards they’re dealt and instead just complaining. It would be nice if we could just try to do our best on whatever the current patch is.
I think that a lot of games other than fighting games also tend to patch too heavily, completely changing the direction of characters, or nerfing them into the dust. I think that the massive changes to SC2 over time represent poor initial direction, rather than an inherent fault with patch culture. We can see similar in versions of fighting games over time though, 2nd impact & New Generation were very different than 3rd strike. Similar from Guilty Gear Sign to Rev2.
I’ll consider writing a more full patch culture article in the future.
I really like this article and tbh it’s pretty validating cos there’s a lot of stuff I was hoping to see in the discussion but don’t have the credibility to declare
To add, specifically to the FG space, I kind of think the slower relative neutral speed and overall game speed of modern games has a pretty large impact on things like expressiveness at lower levels. If the game is sped up it raises the floor of a lot of moves and lets you do a lot more silly or aggressive stuff for no reason in neutral even against much stronger players. You can get away with randomly jumping a lot more in older games or throwing “slow” moves out because without any anticipation they’re simply too difficult to reliably react to. You can reliably form a low to mid level gameplan in some games based on simply rotating between different unexpected “bad ideas” or using your one freebie to throw an extra thing on the opponent’s mental stack, inherently increasing the viability of other options. It almost feels like “reactable” is treated as a binary yes/no property of moves where the asterisk is in earning the right to get away with a move by overloading your opponent’s stack (i.e charge dust) as opposed to being owed the move until the opponent can take it away.
I’m not 100% sure where I’m going with this but I think it’s something along the lines of balancing around human limits and other human elements like anticipation and pressure and not just the vacuum of the individual game elements on a spreadsheet or whatever. I might be full of crap anyways because I usually don’t play enough to get my reactions totally on point in anything, so I’m mostly going off how easy it is to trick experienced players and I don’t have many concrete examples of what I mean. I could probably make a chart though.
To similarly attack the “expand design space by being too overwhelming to optimize” angle, driving up uncertainty in general seems like a good way to create a larger and deeper pool of possibility. What gets lauded about say, +R or whatever other old guilty Gear is the jazzy nature of a hundred systems smashing together at high speed making reliable conversions and damage optimization realistically more difficult. Then of course someone pulls out the concrete example of Testament or whoever existing and easily converting to an optimal combo which crushes the elegance of the rest of the system and, with the same poetic crassness, the ability to have a conversation about it and how it’d be replicated. I don’t think I finished the Romolla video so maybe it’s in the back half but even her discussion seemed to be dominated by focusing on the extremes of uncompetitive knowledge checks and game-defining anchor moves or whatever just ignores the design direction of the rest of the game the most. (I also remember some early sentiment in response to the new direction for Strive being “well whoever they accidentally make the least like the stated goals will be really good” so this kind of thinking will always bubble up)
This part I think expands outside of fighting games a bit more easily as long as the game is in real time and not like unlimited time turn based like RISK or something
I dunno I’m pretty amateur hour and haven’t put in the work on more complex designs yet (also actually balancing around real time factors is a huge headache for my current workflow for example but I think I can grow out of that one) plus I’m not terribly focused so I’m not sure the value of forcing someone to read my thinking out loud.
The article covered the things I’ve managed on my own more thoroughly than I’ve really thought about but it does make me feel like I’m on the right track with some general philosophy stuff so thanks for that much.
Brilliant article, was a joy to read.