Trying to make everything "fair" will just result in everything becoming average-and that's no fun! Average just means mediocre, bland, and unremarkable.
Sakurai out here low-key telling you that if you think a Smash character inherently sucks it's a literal skill issue, and bringing the receipts to prove it.
for online casual play? sure. for offline competitive play? absolutetly not! Smash is NOT a well balanced game when there are only like 5 characters that have a realistic shot at winning a competitive tournament
Same with thinking a character is over powered. One of my friends is constantly complaining about pyra and mythra being overpowered purely because they are my best characters. I should probably show them this.
@@Yarharsuperpirate i'll foretell the prophecy of your future if you choose to go down this route your friends will think you're a complete douchebag and they won't ever want to play with you again. switch off the top tier, dude
Especially when you consider that they don't push an update every week or something, the way MOBAs like League of Legends do. Not saying they managed to get those rates first try, but... in a few tries.
I honestly want to see what the data looks like if you take the top 1000 or so players instead of everyone. The vast majority of players play very very casually, so its much more likely that any given character will be roughly balanced with any other given character. (in fact, the fact that there is as much as a 4% difference between the min and max rates WITH that in mind is surprising.) Higher level players will be much more likely to bring out the imbalances in the game, and hi-light potential weaknesses in that balancing.
@@Anonymimus Honestly, I think it's more difficult for a primarily team based games to be balanced because of so many more factors to consider, think of it as the currently trendy 'three body problem'. (I'm assuming Sakurai's numbers are from 1v1 play, and this isn't to say that THAT is easier either 😅)
@@deefdragon Yeah, but then it becomes more of a philosophical question. Do we balance the game for most players or balance it for the best players? When I was in that top 1% in various games, I thought I wanted balance for the best players at first, but after a while, I realized that relearning the game every 2 to 8 weeks is actually less fun than trying to eek out that last 1% efficiency with the "S/A Tier" character(s). But even then, there's still a middle ground somewhere. We don't want to see 100% meta knight vs. meta knight again either. Just a question of how many characters are S/A Tier vs. "Viable outside of the top 1%".
@@deefdragon You can find or curate a good chunk of that data if you really wanted to. If you're looking for top-level play, there's plenty of tournaments where they're gonna be showing up. You can find player & character standings over years of development this way.
It *is* worth noting that, when you look at truly optimal play, like you see in tournaments, the disparities between characters become much greater due to players being able to better make use of advantages while mitigating shortcomings, and a lot of the cast gets filtered out at that level as a result. However, balancing a game based on tournament-level play is pretty pointless, given that they represent only a fraction of a percent of the playerbase.
I am *fascinated* by the SSBU data he mentioned, I understand that he's probably not allowed to share it at all, but it's still really interesting, just to know that the data is out there at all.
dead by daylight, is a game that is not being thoroughly playtested by the developers at all, and they determine buffs and nerfs on things depending on its pick and success rate. so the data is out there and sometimes developers can be more open about revealing this data. its just not a sureway to balance a game if all you do is look at statistics instead of actual player experience.
@@SobmicSSBB Player character win-rate doesn't really fall into personal data and wouldn't require special agreements as such (in most countries at least, can't claim to know all their rules)
Consider also that if he said “this character wins the most” people would start jumping to that character and ditching whoever the “least winning character” revealing anything about the data to the player base would influence the players and thus change the data
"Internet is some sort of echo chamber" Internet people: Agreed Jokes aside, having a relationship with the interner as a dev must be really hard, not only you have to deal with a fandom with mixed understanding of the thing, you also deal with people who dont care and dont like your type of thing yet keep talking, and the trolls have to be the most annoying part to deal with if you are not able to tell whats trolling and whats not, like, i wonder if Sakurai truly thoight people wanted Shrek in the game.
I think that, for his mental health, he learned not to care what other people say on the Internet. That's the kind of thing that makes people quit gaming industry as a whole
There are way more games with unbalanced asymmetrical design than there are balanced asymmetrical designed games. The reason most games try to balance with similarity is because it's easy. There's a phrase "it's like comparing apples and oranges" for a reason @soundrogue4472 @@soundrogue4472
@@sammer1097 tier lists are generalized opions, it is not the absolute truth and they change more often than not. all a tier list tells you is the general percieved strength or weakness for a particular character, it is not the absolute truth, it is just the generalized opinion of the competitive community.
@@dunnoausernamedunno5533 lower level doesnt guarantee balance. I might actually have a completely different tierlist than high elo because easier character probably perform much better the question is just that if character A beats B every time in low elo but B beats A every time in high elo, would that be a 50% winrate according to this statistic or does the winrate shown take into account the different skill brackets
Sakurai made me ask myself here why do I enjoy character reveal trailers a lot, even though I likely won't play the character much? Because what I really like is fighting against variety, even if my main is usually static.
This video is really important for design in general (I think our natural biases need to be accounted for when designing playable character strength), but I find it interesting that Sakurai points directly to win rate data for Smash Bros.' case. It does beg the question of "how weighted is this win rate data?" - if people at highest GSP (so basically just high ELO) end up weighted higher for win rate results on a character by character basis. Ultimately we are not privy to this raw data, and probably will never be, but this matters - some characters noob stomp by nature, some characters are much better if you're very technical, not to mention the wide range of players in not only skill level, but age and how much time they regularly spend playing as opposed to just hopping on occasionally. This is why in competitive singles Pokemon, when they use win rate based tier shifts, they weight the data for usage based on who uses that Pokemon (with higher level players indicating "more" usage), and they count win rate for low ELO and high ELO (cutoff dependent on the tier and committee) as separate metrics. Notably, back in Ultimate's earlier days, we saw at the invitational that top players exploited Bayonetta in a very similar way to WiiU/3DS/4... and then she got a harsh nerf. This precedent leads me to be wary of the weight of this data. Though in fairness to them, 1) 4 may have not gotten the same degree of playtesting, 2) there's only a couple top players that helped playtest and all in Japan's scene, where afaik Bayonetta was less prominent, and 3) it was ultimately still pre-release with a lot more testing to be done, even though it was pretty close time-wise. That being said, law of high numbers also plays a factor here. There are a lot of Smash Bros. players so the amount that 1% win rate matters depends on the margin of error at play - if there's a smaller sample size, then 1% could simply be margin of error and it's hard to tell. At the astronomical sample size that Smash Ultimate's online match count over its entire *lifetime* ... 1% does matter a lot more, and it could be said that some characters are better than others and will win more consistently. Though that's a bit of a nothingburger for diverse character design - in fact, the fact that no one single character has a larger gap with a major outlier is actually fantastic news. I know people say Ultimate's roster is surprisingly balanced for how large the roster is, and I think this data is still proof of that. But it's worth scrutinizing the little details we *are* privy to, because data can be skewed in isolation :) I appreciate Sakurai's willingness to share this data, and this video is valuable for way more reasons than this one detail
Remember that balancing a game around competitive players risks alienating the vast majority of players unable or unwilling to play at that level/style. Plenty of developers have killed/harmed their games by balancing them right into mediocrity at the behest of comp players and their internet sycophants.
@@CyclonSilver I don't know what games this person is referring to, but I heard Overwatch was a game that went downhill due to balancing around competitive and its league. You can try looking into that.
@@WildspeakerYTit doesn’t account for skill level or other factors that took place in every ongoing match. There’s too much nuance for any of that to be meaningful
Sakurai-san won't reveal which fighters have the highest and lowest win rates, and I think it's for the best. It's his assurance that all players will have a chance, no matter what fighter they choose, is most important for Smash Bros.
It's a good time when Sakurai posts something that even the wider smash community should definitely take notice of(such as his video showing the beta of smash 64). And I for one, appreciate that Sakurai is focused on making a game rather than an "esport", that's really lacking in this genre. But going by the purported win rates, it seems it's still doing pretty alright! Also what people don't get about "tier lists" in regards to smash ultimate tier lists, they are all done under a specific ruleset with specific stages and conditions that in reality the majority of players don't play under, and in fact those who do are the overwhelming minority. You have people who play this game who don't even grab or shield, so it's not just "items" that I'm talking about here. That's probably why characters like Ike and Ganondorf you can probably expect to stay in roughly the same mold they are in rather than something balanced more for esports competitive play, because a lot of people get bodied by those characters regardless.
But then quick play and elite smash become a toxic mess since the balance caters to casuals too much even though they don't want to learn how to play the game
Couldn't have put it any better! eSports aren't everything, even just super serious competitive players take up a small minority like you said. Samurai and his team understand that it's important to never lose the forest for the trees. Even hinting at it when he mentioned echo chambers 😅. At it's core, a game should be fun, and not everyone really cares about frame data or even have the time to. But it's okay if you do, everyone has their own approach.
It's a shame people fall for misrepresented statistics so quickly. It's true that for mediocre players the game is well balanced, which is great for the average player. But for good players or competitive there's a clear divide that you simply can not deny. At the same time, with asymmetric characters it's also entirely unavoidable, so he still did an insanely amazing job.
A perfect video as always. Something I really appreciate here is that Sakurai points out that the internet can often be an echo chamber. For this game especially, people tend to constantly focus on the viability of a handful of characters because they're already being talked about, leading to a perpetual cycle of mostly the same few characters being perceived as "the best" or "the worst" while the vast majority of the cast is overlooked. It's really a shame.
My problem with the game balance personally is that many moves don’t “feel right”. Hitboxes are too small or too big, frame data don’t match the visuals, a strong move doesn’t have enough recovery frames, etc. It makes certain characters frustrating to play as or against.
The parts about the internet being an echo chamber, and proving your point by having an actual match feel like he's directly talking to the fanbase rather than just giving development advice.
Thanks for sharing this tip and the survey data as well! This gives me confidence to let uniqueness shine while increasing flaws of each character in my games to balance the game.
At the moment at least, for some patches the range has been as bad as 30 to 70. Excluding some patches where heroes were bugged in a way that gave them a 99% win rate...
Always thought alike. When friends told me that "this character has no good recovery, he's so bad" or so, they were missing out tons of other strengths and the uniqueness that made this character.
Doesn't matter how strong your character is if he's bad in the most essential aspect of the game. If your character can kill at 50% but he can't hit anything, he's just bad
Okay so I can blame certain top players for using their echo chamber to bully people who use unpopular characters they don't like. Glad to have this information even though I know who that one character with a 47% win rate is. (Still makes a difference at top level play, but again the community isn't really known for being understanding and sympathetic a lot of the time) Thanks Sakurai.
I’ve never really thought about balance like this before. You can still have powerful builds or strengths so long as it also has contrasting weaknesses. Problems in games I’ve played arise from adding strengths such as a new powerful gun, but then not enough weakness added
Gonna talk about Turn Based Strategy for a bit, but characters with big strengths AND weaknesses tend to be fun, in my view. It's one reason I like the playable cast of Fire Emblem Engage, they all have very clear things they're good at and bad at from the moment they join. It's also why I find Andy from Advance Wars got a lot more interesting in the 2nd game, where he gained a REALLY strong Super CO Power as a reward for "putting up with" his average-ness long enough to build it. At the same time, I also feel Jake from Dual Strike was a more interesting way to do an "average" character in Advance Wars. Instead of average units across the board, his CO Powers boost movement range of Direct Combat units AND firing range on Indirects, but at less of an effect than the dedicated specialists for both. This lets him use a varied army and flexibly adapt, without feeling too bland. (Though the "firepower boost on Plains" thing was minor enough to not be too relevant)
For how unbalanced Fire Emblem is, especially post-Awakening, it still works because even suboptimal strats and builds feel really strong. Really, FE is a great argument how balance isn't the end-all, with how the people's favourites are Genealogy, Awakening and Three Houses.
@@SobmicSSBB I'd say Fates Conquest's units are shockingly well-balanced for how hard that game is, besides Camilla being too good. (But you kind of need her, and at least she isn't as mindless as Ryoma on Birthright) There's kind of two categories of FE, the more tactical ones, and the "fun to break sandbox ones", which those three fall into but also Sacred Stones. (And maybe Path of Radiance too. It's easy, but everyone's usable.) Engage has my favourite set of characters gameplay-wise though, because of how intuitive everyone's introduction is for how to use them (Etie joins near enemy Pegasi, Chloe near a Mage, Celine with Celica's Emblem to show off warping, etc) I'm also in a massive minority that find 3H's story and characters its only redeeming aspect. I find the monastery too tedious on repeat playthroughs and the chapters aren't fun enough to make up for it. 80% of your time is spent building units rather than using them.
@@BigKlingy You aren't the only one in that minority. The monastery is just too much of a slog. I wanted to play through all the routes and experience the full story but after my first playthrough I was already feeling the burnout. The amount of time it took for me to clear every path took me longer and longer mostly because it was hard building up the willpower to see it through to the end and that is mostly due to all the chores you had to do between chapters. Engage's story might have been cheesy as hell but I think it did a better job in the parts that mattered most.
@@Doppel145 I personally think Warriors Three Hopes handles the base much better, but I haven't played it due to going through exactly what you did: I played all 4 routes back to back and got massive Fodlan burnout. In a game without 4 routes that you need to play all of to fully understand the story, the monastery might've been fine. It also would've been fine if the actual maps were god-tier. (e.g I don't like the Somniel minigames in Engage, but I don't care because the maps are amazing)
Though I have to wonder how much of those results come from free for all and item play. Has anyone else noticed the power of spacific attacks is lowered in an item match? Of course all characters are going to be close to 50% if your including those results. It's a shame we will never know for sure if that's the case though.
But I still think that the online match data is a bit biased towards that, online gameplay. I think we can all agree the game plays very differently when there is a lot of input lag involved, hence not allowing each character to be expressed as intended
Not just Smash Ultimate, either. Vampire Savior/Darkstalkers 3 is a fighting game with a much smaller roster and much less balance work done, yet is balanced enough that its "worst" character, Anakaris, is pro tournament-viable as one cult-iconic VSav event proved. I believe the key to facilitating balance in spite of diversity is a strong core design that gives everyone powerful common tools, with the unique traits being largely vectors toward applying those tools.
Not a game designer, but definitely a lesson I learned years ago after believing for a long time that "perfectly average" would be the ideal balancing point. I've definitely come around on that. 3:12 I've gotta say, though, I was definitely curious as to what the collected data looked like and that really helps to understand why the game balance in Ultimate looks the way it does. Obviously, the counter argument is that a character's performance is likely gonna fluctuate a great deal between different levels of play, so this statistical balance does not represent how each character performs at a competitive level. That's how we get the competitive community Tier Lists that we do, after all. But perhaps that's also kind of the point of the video. While balance is a good ideal to strive for, perfect balance probably should not be the explicit goal, even if it'll inevitably upset some people.
Sakurai, if it's not troubling you much, could you speak about designing weapons and abilities for single player games? More specifically, what I would like to be seen addressed is the appeal of weapons and abilities that often causes people to disregard one to favor the other because either the stats are greater or because it looks cooler (specially in action). For example: Whenever I play Kirby, Sword is my preferred ability, while Beam is the one I avoid at all costs. What are your thoughts on that?
really surprising results on the win rates. there are still HEAVY disadvantages for fighters competitively, but globally things are more evened out than i expected
I believe casual play has a heavy role here, which are still the majority of the playerbase(not just Smash, but any game out there has a casual playerbase hundred times larger than the tournament players). Some characters are not that broken if the player doesn't know how to utilize it or is even aware of it. Like, how many games include Luigi as a character and how many include him performing a 0-to-death combo? At high level tournaments, Ganons slow speed will be his quick downfall, but random online matches(with a not to underestimated amount of delay/lag) will include lots of people just running into his oversized sword and get KOed at 60%...
League of Legends/Dota 2’s balancing is insane because not only do they also have to factor in item synergy and matchups, they also have to worry how fair and fun characters are from early to late game.
This video reminds me Square Enix's issues with balancing jobs in FFXIV. Recently, I feel they've begun to homogenize a lot of their jobs, adding similar skills across the board for every job.
The stronger people think a character is, the more people play them in a competitive setting. The more people that play a character in a tournament, the more likely that character is to win the tournament. A tournament where a large portion of the competitors play Mario, is more likely to have Mario in grand finals than a tournament win only one Mario player. More people playing the character also means new strategies and techniques are developed at a faster rate. Steve is a good example of this, as it took a bit before people thought of Steve as a top tier. Once that happened, the rate of new Steve tech exploded. Because that character does well in tournaments, more people will think that character is good. The result is a feedback loop that keeps characters in at the top for a long time, and keeps other characters at the bottom.
Fascinating! It's astonishing that Sakurai-san and his team have managed to establish that degree of balance across all of this game's characters - AND have the data to prove it. That said...I can't help but be curious which character has that ever-so slight advantage at 51.43% wins recorded thus far.
Very interesting points, I know the last bit giving too much info could breach some NDAs and such, but would be interesting to know if that data is from all type of battles, or just 1vs1, if Items were on or off. Probably wouldn't vary too much the win % in either margin, but considering it is just numbers without too much context, feels weird
not everyone play/is interested in playing at the high levels, so makes sense to take into account the bulk of the players rather than the 1 or 2% thats at the top, plus it not only takes into consideration 1v1, but also teams and free for alls.
The win rate stuff gets a bit muddled in any game that has skill-based matchmaking. True SBMM will always draw you toward a 50% winrate, which makes character differences disappear, only because the better characters will be in a higher Elo than worse characters.
The character differences could also balance out the skill difference between players, too. You'd have to be way more skilled as Shiek when fighting against a Ganondorf player who may not be as skilled.
@@yannlindegaard8522 balancing for a team game where small differences in leads and money snowball over the course of 40+ minutes and accounting for all the different interactions between characters and items is a wholly different beast. I don't think you can compare the two.
This man is a game design Demi-god. Every game maker should be lining up around the block to throw money at this man. That being said, I really wish there was a way to dub these in English. I consume most of my TH-cam via audio (listen while working) and I always have to skip his videos when I'm on the clock.
I can see where Sakurai is coming from with the need to avoid leveling everybody to a boring "flatness" of sorts, but honestly, making characters with huge extremes such as Pyra and Mythra, with great everything EXCEPT a horrible recovery, or Ganondorf with absurd kill power and everything else pretty much crap, cannot be the way either, as a lot of stuff ends up being extremely polarising and not fun as well.
Well ain't this a unique perspective on things? Plus, that data towards the end is really interesting; it further hammers it in to me that tier lists really don't matter... well, atleast it's what I think.
Sakurai is right, win rates aren't everything; for example, that 51.43% win rate character, can the 47.18% character win against them? Is it an uphill battle where 51.43% has it easier? How much easier? Even in the worst case, how often does this match up happen? There's at least several dozen more factors. Balance is WAYYY complicated, especially with a cast so large.
Brigid was the character who immediately showed their balance design principles shifting. Overwatch was balanced around the idea of every character having strong, and weak points to bring to a match. And was designed around having to swap if someone wasn't performing well. Brigid was the first "Good at everything" character and instantly ruined that whole balance idea
Yup, this is really evident every time they've reworked a character's kit. Nearly every time the uniqueness gets pared back in favor of creating something that is neither strong nor weak against anything in particular. Just cause a bunch of comp players demand it doesn't mean it's a good idea.
Man I was just wondering about how big competitive game devs feel about the balance discourse online. Like, maybe you're trying your best not to tell them about that sleeper pick that destroys the "OP" meta strategy and shut everyone tf up.
I wish I could see more details in the data. Win rate on its own isn’t that helpful of a statistic without a sample size for the character. Still though, having the win rates as close as they are is a pretty big achievement
I was expecting this for years, is an invaluable advice and guidance like this right now... How to balance everything, making it unique. I'm feeling more relaxed, I wasn't that off track in what I do and, after all... I wasn't the only one who was going through the same questions. My efforts in these years have not been in vain. Thank you, Mr. Sakurai.
why do i remember Sakruai sharing this whole online winrate data thing before? i specifically remember this whole "the overall winrate is roughly 50% among all fighters" thing.
So, the win rate is basically the whole coin flip paradigm people been discussing, where one side is more likely to win than another, but by a small margin.
Tier lists and metas are for competitive play. They mean absolutely nothing when it comes to casual, which is what the win rates point out. When a game is put in a situation of "how can I do my best with the best?" Then tier lists are important. But again, that means nothing when your goal is to have fun lol
The funny thing is that while winning rates may seem levelled across the board, there might be cases where a fighter has a bigger percentage against a specific opponent, but a lower percentage against another. Still part of the push & pull philosophy, but be careful if only considering averages when balancing. Those are my two cents on this topic. Many thanks for the video, Sakurai!
Many people mix up 'poor designed' with 'being bad'. I think Ganon is an example of a well designed bad character. Everyone loves playing him despite that he sucks
I kinda feel the title is a bit misleading. For a sec i thought he was going to talk about how some people thinkg something being a 6/10 is acceptable while others thing is worst than being a nad product, but then again, the video is in the [design specifics] Although it would be interesting to see what japanese game designers think of the game rating system.
![Meteos [Game Concepts]](http://i.ytimg.com/vi/snHecpmPeJs/mqdefault.jpg)
![Meteos [Game Concepts]](/img/tr.png)
![Using Parameters to Establish Characters [Planning & Game Design]](http://i.ytimg.com/vi/zwiS1L6QVY0/mqdefault.jpg)
Sakurai calling internet an "echo chamber" must be one of the most accurate descriptions when an opinion comes to mind.
Hello soul of cinder, hows thine ash?
Twitter
reddit
Degenerates* 👀
@@ACDCvirusLiterally ANY social media platform. Twitter or not.
Sakurai out here low-key telling you that if you think a Smash character inherently sucks it's a literal skill issue, and bringing the receipts to prove it.
you don't even compete what are you talking about
for online casual play? sure.
for offline competitive play? absolutetly not!
Smash is NOT a well balanced game when there are only like 5 characters that have a realistic shot at winning a competitive tournament
Same with thinking a character is over powered. One of my friends is constantly complaining about pyra and mythra being overpowered purely because they are my best characters. I should probably show them this.
@@Yarharsuperpirate i'll foretell the prophecy of your future if you choose to go down this route
your friends will think you're a complete douchebag and they won't ever want to play with you again. switch off the top tier, dude
@@Yarharsuperpiratethey kill at 90% with an Up B. Shut up
1:42 "Settle it... IN S M A S H" ~Sakurai
How about Kid Icarus Uprising?
@@MasterKnightDH OPs line was an ad for Smash on Wii U and 3DS from 2014.
Wow, that margin between the top and bottom character's winrate is impressively small. The mark of a well balanced game, right there.
Especially when you consider that they don't push an update every week or something, the way MOBAs like League of Legends do. Not saying they managed to get those rates first try, but... in a few tries.
I honestly want to see what the data looks like if you take the top 1000 or so players instead of everyone. The vast majority of players play very very casually, so its much more likely that any given character will be roughly balanced with any other given character. (in fact, the fact that there is as much as a 4% difference between the min and max rates WITH that in mind is surprising.) Higher level players will be much more likely to bring out the imbalances in the game, and hi-light potential weaknesses in that balancing.
@@Anonymimus Honestly, I think it's more difficult for a primarily team based games to be balanced because of so many more factors to consider, think of it as the currently trendy 'three body problem'. (I'm assuming Sakurai's numbers are from 1v1 play, and this isn't to say that THAT is easier either 😅)
@@deefdragon Yeah, but then it becomes more of a philosophical question. Do we balance the game for most players or balance it for the best players? When I was in that top 1% in various games, I thought I wanted balance for the best players at first, but after a while, I realized that relearning the game every 2 to 8 weeks is actually less fun than trying to eek out that last 1% efficiency with the "S/A Tier" character(s).
But even then, there's still a middle ground somewhere. We don't want to see 100% meta knight vs. meta knight again either. Just a question of how many characters are S/A Tier vs. "Viable outside of the top 1%".
@@deefdragon You can find or curate a good chunk of that data if you really wanted to. If you're looking for top-level play, there's plenty of tournaments where they're gonna be showing up. You can find player & character standings over years of development this way.
It *is* worth noting that, when you look at truly optimal play, like you see in tournaments, the disparities between characters become much greater due to players being able to better make use of advantages while mitigating shortcomings, and a lot of the cast gets filtered out at that level as a result. However, balancing a game based on tournament-level play is pretty pointless, given that they represent only a fraction of a percent of the playerbase.
"The true metric is how the character feels"
I am *fascinated* by the SSBU data he mentioned, I understand that he's probably not allowed to share it at all, but it's still really interesting, just to know that the data is out there at all.
That's what the data collection we keep blindly agreeing to when we accept terms of service is for.
dead by daylight, is a game that is not being thoroughly playtested by the developers at all, and they determine buffs and nerfs on things depending on its pick and success rate.
so the data is out there and sometimes developers can be more open about revealing this data. its just not a sureway to balance a game if all you do is look at statistics instead of actual player experience.
@@SobmicSSBB Player character win-rate doesn't really fall into personal data and wouldn't require special agreements as such (in most countries at least, can't claim to know all their rules)
Consider also that if he said “this character wins the most” people would start jumping to that character and ditching whoever the “least winning character” revealing anything about the data to the player base would influence the players and thus change the data
"Internet is some sort of echo chamber"
Internet people: Agreed
Jokes aside, having a relationship with the interner as a dev must be really hard, not only you have to deal with a fandom with mixed understanding of the thing, you also deal with people who dont care and dont like your type of thing yet keep talking, and the trolls have to be the most annoying part to deal with if you are not able to tell whats trolling and whats not, like, i wonder if Sakurai truly thoight people wanted Shrek in the game.
I think that, for his mental health, he learned not to care what other people say on the Internet. That's the kind of thing that makes people quit gaming industry as a whole
The fact that he managed to gain such good balance with only asymmetrical characters truly astounds me. Well done Sakurai
dude you've got a scary ass spider on your profile pic wtf get rid of it
I don't have a profile Pic, Just the letter N@@MatrixEvolution17
You have a large cast of asymmetrical characters which unironically make balance; if you don't then you end up like Overwatch.
Yo same, I saw your spider on your logo and it made me flinch.
There are way more games with unbalanced asymmetrical design than there are balanced asymmetrical designed games. The reason most games try to balance with similarity is because it's easy. There's a phrase "it's like comparing apples and oranges" for a reason @soundrogue4472 @@soundrogue4472
Sakurai doesn't believe in tierlists
Completely missed the point he was making.
And he shouldn't. Tiers are poison.
Tier lists are biased.
@@sammer1097 tier lists are generalized opions, it is not the absolute truth and they change more often than not.
all a tier list tells you is the general percieved strength or weakness for a particular character, it is not the absolute truth, it is just the generalized opinion of the competitive community.
Didn't he think Ganondorf was the best character in Brawl or whatever?
For as massive of a roster Smash Ultimate has, it's amazing how well balanced it is even if it isn't perfect.
This just in:
Statistically speaking,
your Smash main is mid.
And when everyone's mid, no one will be.
@@dunnoausernamedunno5533 lower level doesnt guarantee balance. I might actually have a completely different tierlist than high elo because easier character probably perform much better
the question is just that if character A beats B every time in low elo but B beats A every time in high elo, would that be a 50% winrate according to this statistic or does the winrate shown take into account the different skill brackets
Sakurai made me ask myself here why do I enjoy character reveal trailers a lot, even though I likely won't play the character much? Because what I really like is fighting against variety, even if my main is usually static.
This video is really important for design in general (I think our natural biases need to be accounted for when designing playable character strength), but I find it interesting that Sakurai points directly to win rate data for Smash Bros.' case.
It does beg the question of "how weighted is this win rate data?" - if people at highest GSP (so basically just high ELO) end up weighted higher for win rate results on a character by character basis. Ultimately we are not privy to this raw data, and probably will never be, but this matters - some characters noob stomp by nature, some characters are much better if you're very technical, not to mention the wide range of players in not only skill level, but age and how much time they regularly spend playing as opposed to just hopping on occasionally. This is why in competitive singles Pokemon, when they use win rate based tier shifts, they weight the data for usage based on who uses that Pokemon (with higher level players indicating "more" usage), and they count win rate for low ELO and high ELO (cutoff dependent on the tier and committee) as separate metrics.
Notably, back in Ultimate's earlier days, we saw at the invitational that top players exploited Bayonetta in a very similar way to WiiU/3DS/4... and then she got a harsh nerf. This precedent leads me to be wary of the weight of this data. Though in fairness to them, 1) 4 may have not gotten the same degree of playtesting, 2) there's only a couple top players that helped playtest and all in Japan's scene, where afaik Bayonetta was less prominent, and 3) it was ultimately still pre-release with a lot more testing to be done, even though it was pretty close time-wise.
That being said, law of high numbers also plays a factor here. There are a lot of Smash Bros. players so the amount that 1% win rate matters depends on the margin of error at play - if there's a smaller sample size, then 1% could simply be margin of error and it's hard to tell. At the astronomical sample size that Smash Ultimate's online match count over its entire *lifetime* ... 1% does matter a lot more, and it could be said that some characters are better than others and will win more consistently. Though that's a bit of a nothingburger for diverse character design - in fact, the fact that no one single character has a larger gap with a major outlier is actually fantastic news. I know people say Ultimate's roster is surprisingly balanced for how large the roster is, and I think this data is still proof of that. But it's worth scrutinizing the little details we *are* privy to, because data can be skewed in isolation :)
I appreciate Sakurai's willingness to share this data, and this video is valuable for way more reasons than this one detail
Remember that balancing a game around competitive players risks alienating the vast majority of players unable or unwilling to play at that level/style. Plenty of developers have killed/harmed their games by balancing them right into mediocrity at the behest of comp players and their internet sycophants.
bro i aint readin allthat bro cmon man... take a shower
@@KnoxEmDown I have opinions on that, but mostly I'd really like to hear examples. Not even ironically, please educate me.
@@CyclonSilver I don't know what games this person is referring to, but I heard Overwatch was a game that went downhill due to balancing around competitive and its league. You can try looking into that.
the win rate stuff is very interesting
It's really not. The stats for the whole playerbase does not prove balance between the characters.
@@R-son-fm1wn it does prove it when your game is for the whole playerbase of people playing the game, and not only the ones in tournaments
@@WildspeakerYTit doesn’t account for skill level or other factors that took place in every ongoing match. There’s too much nuance for any of that to be meaningful
Sakurai-san won't reveal which fighters have the highest and lowest win rates, and I think it's for the best.
It's his assurance that all players will have a chance, no matter what fighter they choose, is most important for Smash Bros.
It's a good time when Sakurai posts something that even the wider smash community should definitely take notice of(such as his video showing the beta of smash 64). And I for one, appreciate that Sakurai is focused on making a game rather than an "esport", that's really lacking in this genre. But going by the purported win rates, it seems it's still doing pretty alright!
Also what people don't get about "tier lists" in regards to smash ultimate tier lists, they are all done under a specific ruleset with specific stages and conditions that in reality the majority of players don't play under, and in fact those who do are the overwhelming minority. You have people who play this game who don't even grab or shield, so it's not just "items" that I'm talking about here. That's probably why characters like Ike and Ganondorf you can probably expect to stay in roughly the same mold they are in rather than something balanced more for esports competitive play, because a lot of people get bodied by those characters regardless.
But then quick play and elite smash become a toxic mess since the balance caters to casuals too much even though they don't want to learn how to play the game
Couldn't have put it any better! eSports aren't everything, even just super serious competitive players take up a small minority like you said. Samurai and his team understand that it's important to never lose the forest for the trees. Even hinting at it when he mentioned echo chambers 😅. At it's core, a game should be fun, and not everyone really cares about frame data or even have the time to. But it's okay if you do, everyone has their own approach.
It's a shame people fall for misrepresented statistics so quickly. It's true that for mediocre players the game is well balanced, which is great for the average player. But for good players or competitive there's a clear divide that you simply can not deny. At the same time, with asymmetric characters it's also entirely unavoidable, so he still did an insanely amazing job.
A perfect video as always. Something I really appreciate here is that Sakurai points out that the internet can often be an echo chamber. For this game especially, people tend to constantly focus on the viability of a handful of characters because they're already being talked about, leading to a perpetual cycle of mostly the same few characters being perceived as "the best" or "the worst" while the vast majority of the cast is overlooked. It's really a shame.
i wonder if there's someone out there who genuinely thinks melee only has like 3 characters
My problem with the game balance personally is that many moves don’t “feel right”. Hitboxes are too small or too big, frame data don’t match the visuals, a strong move doesn’t have enough recovery frames, etc. It makes certain characters frustrating to play as or against.
This is exactly why less balanced games tend to be more interesting.
The parts about the internet being an echo chamber, and proving your point by having an actual match feel like he's directly talking to the fanbase rather than just giving development advice.
Basically, he's saying that its best to experiment and make up your own conclusions.
Thanks for sharing this tip and the survey data as well! This gives me confidence to let uniqueness shine while increasing flaws of each character in my games to balance the game.
He has to be talking about Helldivers lol
3:23 that one part sakurai said makes totally sense.
"Stop arguing on the Internet and bring back duels"
- Sakurai, essentially
Settle it in Smash!
Winrates ranging from only 51 to 48. That is actually very impressive balance. I play a lot of Dota and in it it the winrates range from 40 to 60.
At the moment at least, for some patches the range has been as bad as 30 to 70. Excluding some patches where heroes were bugged in a way that gave them a 99% win rate...
That's just what matchmaking is supposed to do
Always thought alike. When friends told me that "this character has no good recovery, he's so bad" or so, they were missing out tons of other strengths and the uniqueness that made this character.
Doesn't matter how strong your character is if he's bad in the most essential aspect of the game. If your character can kill at 50% but he can't hit anything, he's just bad
but overall a character can still balance out to be bad
Okay so I can blame certain top players for using their echo chamber to bully people who use unpopular characters they don't like. Glad to have this information even though I know who that one character with a 47% win rate is. (Still makes a difference at top level play, but again the community isn't really known for being understanding and sympathetic a lot of the time) Thanks Sakurai.
Probably Sheik
I’ve never really thought about balance like this before. You can still have powerful builds or strengths so long as it also has contrasting weaknesses. Problems in games I’ve played arise from adding strengths such as a new powerful gun, but then not enough weakness added
Gonna talk about Turn Based Strategy for a bit, but characters with big strengths AND weaknesses tend to be fun, in my view. It's one reason I like the playable cast of Fire Emblem Engage, they all have very clear things they're good at and bad at from the moment they join. It's also why I find Andy from Advance Wars got a lot more interesting in the 2nd game, where he gained a REALLY strong Super CO Power as a reward for "putting up with" his average-ness long enough to build it. At the same time, I also feel Jake from Dual Strike was a more interesting way to do an "average" character in Advance Wars. Instead of average units across the board, his CO Powers boost movement range of Direct Combat units AND firing range on Indirects, but at less of an effect than the dedicated specialists for both. This lets him use a varied army and flexibly adapt, without feeling too bland. (Though the "firepower boost on Plains" thing was minor enough to not be too relevant)
For how unbalanced Fire Emblem is, especially post-Awakening, it still works because even suboptimal strats and builds feel really strong.
Really, FE is a great argument how balance isn't the end-all, with how the people's favourites are Genealogy, Awakening and Three Houses.
@@SobmicSSBB I'd say Fates Conquest's units are shockingly well-balanced for how hard that game is, besides Camilla being too good. (But you kind of need her, and at least she isn't as mindless as Ryoma on Birthright)
There's kind of two categories of FE, the more tactical ones, and the "fun to break sandbox ones", which those three fall into but also Sacred Stones. (And maybe Path of Radiance too. It's easy, but everyone's usable.) Engage has my favourite set of characters gameplay-wise though, because of how intuitive everyone's introduction is for how to use them (Etie joins near enemy Pegasi, Chloe near a Mage, Celine with Celica's Emblem to show off warping, etc)
I'm also in a massive minority that find 3H's story and characters its only redeeming aspect. I find the monastery too tedious on repeat playthroughs and the chapters aren't fun enough to make up for it. 80% of your time is spent building units rather than using them.
@@BigKlingy You aren't the only one in that minority. The monastery is just too much of a slog. I wanted to play through all the routes and experience the full story but after my first playthrough I was already feeling the burnout. The amount of time it took for me to clear every path took me longer and longer mostly because it was hard building up the willpower to see it through to the end and that is mostly due to all the chores you had to do between chapters. Engage's story might have been cheesy as hell but I think it did a better job in the parts that mattered most.
@@Doppel145 I personally think Warriors Three Hopes handles the base much better, but I haven't played it due to going through exactly what you did: I played all 4 routes back to back and got massive Fodlan burnout. In a game without 4 routes that you need to play all of to fully understand the story, the monastery might've been fine. It also would've been fine if the actual maps were god-tier. (e.g I don't like the Somniel minigames in Engage, but I don't care because the maps are amazing)
Though I have to wonder how much of those results come from free for all and item play. Has anyone else noticed the power of spacific attacks is lowered in an item match? Of course all characters are going to be close to 50% if your including those results. It's a shame we will never know for sure if that's the case though.
But I still think that the online match data is a bit biased towards that, online gameplay. I think we can all agree the game plays very differently when there is a lot of input lag involved, hence not allowing each character to be expressed as intended
Not just Smash Ultimate, either. Vampire Savior/Darkstalkers 3 is a fighting game with a much smaller roster and much less balance work done, yet is balanced enough that its "worst" character, Anakaris, is pro tournament-viable as one cult-iconic VSav event proved. I believe the key to facilitating balance in spite of diversity is a strong core design that gives everyone powerful common tools, with the unique traits being largely vectors toward applying those tools.
Somebody send this video to the FFXIV job balancing/design team.
The only job I played and truly enjoyed was red mage, because it had a difference compared to the other classes.
"Average and mediocre are the same thing." Somebody send this to helldivers2 balancing team.
Not a game designer, but definitely a lesson I learned years ago after believing for a long time that "perfectly average" would be the ideal balancing point. I've definitely come around on that.
3:12 I've gotta say, though, I was definitely curious as to what the collected data looked like and that really helps to understand why the game balance in Ultimate looks the way it does. Obviously, the counter argument is that a character's performance is likely gonna fluctuate a great deal between different levels of play, so this statistical balance does not represent how each character performs at a competitive level. That's how we get the competitive community Tier Lists that we do, after all.
But perhaps that's also kind of the point of the video. While balance is a good ideal to strive for, perfect balance probably should not be the explicit goal, even if it'll inevitably upset some people.
I always forget and get surprised how deep Sakurai's voice is 😂
That win rate scale is probably closer to just how often the characters are played.
This reminds me of a video I've watched from a channel called Bricky where he talks about how fun should always be prioritized before balance is.
The fact all 85 fighters win rate fall within 4% of each other is INSANE.
Sakurai, if it's not troubling you much, could you speak about designing weapons and abilities for single player games? More specifically, what I would like to be seen addressed is the appeal of weapons and abilities that often causes people to disregard one to favor the other because either the stats are greater or because it looks cooler (specially in action). For example: Whenever I play Kirby, Sword is my preferred ability, while Beam is the one I avoid at all costs. What are your thoughts on that?
I used to never play Little Mac because everyone said his recovery sucked, then I found out that he’s crazy good at control
And that's why an echo chamber of negativity can lead to false expectations. Try something yourself and decide if your playstyle fits it!
What do you mean with "control"?
This is such a fantastic video! Love the section regarding win rates. The visual effects and sounds reminded me of Wii fit
amazing insight and advice! thanks sakurai!
It's very important thing to consider. Thank you for sharing!
Why does it look likes he's getting younger
So that Ratio could be more about skill-gaps and player demand on characters.
If people want to prove their point, they're best off settling it in an actual match, I'd say."
*Based*
Settle it in Smash!
This kind of time-tested topic is my favourite. Feels like it applies to creative development outside of games! Thanks for sharing your insight!
really surprising results on the win rates. there are still HEAVY disadvantages for fighters competitively, but globally things are more evened out than i expected
I believe casual play has a heavy role here, which are still the majority of the playerbase(not just Smash, but any game out there has a casual playerbase hundred times larger than the tournament players). Some characters are not that broken if the player doesn't know how to utilize it or is even aware of it. Like, how many games include Luigi as a character and how many include him performing a 0-to-death combo?
At high level tournaments, Ganons slow speed will be his quick downfall, but random online matches(with a not to underestimated amount of delay/lag) will include lots of people just running into his oversized sword and get KOed at 60%...
Wow the balance is incredible!
Absolute Banger video milord.
The first smash bros came out when I was 9 years old. I'm now 34. This man looks 15 years younger than me. HOWWWW?!?!
League of Legends/Dota 2’s balancing is insane because not only do they also have to factor in item synergy and matchups, they also have to worry how fair and fun characters are from early to late game.
This video reminds me Square Enix's issues with balancing jobs in FFXIV. Recently, I feel they've begun to homogenize a lot of their jobs, adding similar skills across the board for every job.
The stronger people think a character is, the more people play them in a competitive setting.
The more people that play a character in a tournament, the more likely that character is to win the tournament. A tournament where a large portion of the competitors play Mario, is more likely to have Mario in grand finals than a tournament win only one Mario player.
More people playing the character also means new strategies and techniques are developed at a faster rate. Steve is a good example of this, as it took a bit before people thought of Steve as a top tier. Once that happened, the rate of new Steve tech exploded.
Because that character does well in tournaments, more people will think that character is good.
The result is a feedback loop that keeps characters in at the top for a long time, and keeps other characters at the bottom.
Sakurai really just addressed one of the most controversial aspects of Super Smash Brothers like it was nothing.
Fascinating! It's astonishing that Sakurai-san and his team have managed to establish that degree of balance across all of this game's characters - AND have the data to prove it.
That said...I can't help but be curious which character has that ever-so slight advantage at 51.43% wins recorded thus far.
This is why I play as every character when I play smash and online I set the settings to all random
Very interesting points, I know the last bit giving too much info could breach some NDAs and such, but would be interesting to know if that data is from all type of battles, or just 1vs1, if Items were on or off. Probably wouldn't vary too much the win % in either margin, but considering it is just numbers without too much context, feels weird
That is why Gannon will never get buffed.
PREACH
this is why i tell people overbalanced games suck, when every option is the same, then there is no option, cough cough halo infinite.
Any character can be good I’ve always said this glad to hear it officially confirmed
I hope the data doesn't include people who self-destruct at the start of a match for some reason.
If you just aim to make them all fair, what you've done is make them boring. I love unique fighter mechanics.
That margin is crazy, I can't begin to imagine how much work came into balancing the game
Are the win rate's based on play at all levels? Low level play might be skewing the results a tad.
not everyone play/is interested in playing at the high levels, so makes sense to take into account the bulk of the players rather than the 1 or 2% thats at the top, plus it not only takes into consideration 1v1, but also teams and free for alls.
The win rate stuff gets a bit muddled in any game that has skill-based matchmaking. True SBMM will always draw you toward a 50% winrate, which makes character differences disappear, only because the better characters will be in a higher Elo than worse characters.
The character differences could also balance out the skill difference between players, too. You'd have to be way more skilled as Shiek when fighting against a Ganondorf player who may not be as skilled.
dota and league both have sbmm but one game has champ wr ranging from 40 to 60% whilst the other keep them within 48 to 52%
@@yannlindegaard8522 MOBAs don't count because they're team games.
@@yannlindegaard8522 balancing for a team game where small differences in leads and money snowball over the course of 40+ minutes and accounting for all the different interactions between characters and items is a wholly different beast. I don't think you can compare the two.
I wonder why that data is secret. It seems like something that would be able to be posted freely.
Thank you Sakurai San!
This man is a game design Demi-god. Every game maker should be lining up around the block to throw money at this man. That being said, I really wish there was a way to dub these in English. I consume most of my TH-cam via audio (listen while working) and I always have to skip his videos when I'm on the clock.
I can see where Sakurai is coming from with the need to avoid leveling everybody to a boring "flatness" of sorts, but honestly, making characters with huge extremes such as Pyra and Mythra, with great everything EXCEPT a horrible recovery, or Ganondorf with absurd kill power and everything else pretty much crap, cannot be the way either, as a lot of stuff ends up being extremely polarising and not fun as well.
I imagine these characters are not as crazy to fight against when there's stage obstacles and items to use against them.
I like different match-ups in Smash Bros.
Awesome video!
Well ain't this a unique perspective on things? Plus, that data towards the end is really interesting; it further hammers it in to me that tier lists really don't matter... well, atleast it's what I think.
Ultimate rock paper scissors
Sakurai is right, win rates aren't everything; for example, that 51.43% win rate character, can the 47.18% character win against them? Is it an uphill battle where 51.43% has it easier? How much easier? Even in the worst case, how often does this match up happen? There's at least several dozen more factors. Balance is WAYYY complicated, especially with a cast so large.
This is why I stopped playing Overwatch, the game became boring for the sake of balance.
Brigid was the character who immediately showed their balance design principles shifting. Overwatch was balanced around the idea of every character having strong, and weak points to bring to a match. And was designed around having to swap if someone wasn't performing well. Brigid was the first "Good at everything" character and instantly ruined that whole balance idea
Yup, this is really evident every time they've reworked a character's kit. Nearly every time the uniqueness gets pared back in favor of creating something that is neither strong nor weak against anything in particular. Just cause a bunch of comp players demand it doesn't mean it's a good idea.
Came here to say this
Man I was just wondering about how big competitive game devs feel about the balance discourse online. Like, maybe you're trying your best not to tell them about that sleeper pick that destroys the "OP" meta strategy and shut everyone tf up.
Average and mediocre being the same thing is true…
But I feel like calling something mediocre is a bigger insult than average lol
That's the point. He's trying to help you avoid "average is good" by making you associate it with the disgusting "mediocre".
I wish I could see more details in the data. Win rate on its own isn’t that helpful of a statistic without a sample size for the character. Still though, having the win rates as close as they are is a pretty big achievement
I was expecting this for years, is an invaluable advice and guidance like this right now...
How to balance everything, making it unique. I'm feeling more relaxed,
I wasn't that off track in what I do and, after all...
I wasn't the only one who was going through the same questions.
My efforts in these years have not been in vain.
Thank you, Mr. Sakurai.
God I watch these videos and I still have a tiny hope in the back of my head that he’ll announce a new smash character lol. Miss those days.
why do i remember Sakruai sharing this whole online winrate data thing before?
i specifically remember this whole "the overall winrate is roughly 50% among all fighters" thing.
So, the win rate is basically the whole coin flip paradigm people been discussing, where one side is more likely to win than another, but by a small margin.
This proves why tier lists and metas should never be taken as gospel.
Branch out and experiment with everything and you'll have more fun!
Tier lists and metas are for competitive play. They mean absolutely nothing when it comes to casual, which is what the win rates point out. When a game is put in a situation of "how can I do my best with the best?" Then tier lists are important. But again, that means nothing when your goal is to have fun lol
The funny thing is that while winning rates may seem levelled across the board, there might be cases where a fighter has a bigger percentage against a specific opponent, but a lower percentage against another. Still part of the push & pull philosophy, but be careful if only considering averages when balancing.
Those are my two cents on this topic. Many thanks for the video, Sakurai!
you forgot to mention that payable characters should be more powerfull if you want good sales
overwatch devs could learn something
I know, right?
So.... how about that Ganondorf? O.O
were the win rate precents among casual matches or elite rank? :O
Many people mix up 'poor designed' with 'being bad'. I think Ganon is an example of a well designed bad character. Everyone loves playing him despite that he sucks
I wonder if the character choice and winrates change on the region.
jesus these win rates are insane
Great insight from the great Sakurai
Great insight as always Sakurai-san 😊
I love this.
Confirmed: Sonic win rate is 51.43% and Little Mac is 47.18%
Thanks Sakurai
Sonic has the highest win rate!? This is truly a W.
The prophet hath spoken.
I kinda feel the title is a bit misleading. For a sec i thought he was going to talk about how some people thinkg something being a 6/10 is acceptable while others thing is worst than being a nad product, but then again, the video is in the [design specifics]
Although it would be interesting to see what japanese game designers think of the game rating system.