lets make a 75+% good matchup chart

[zackattack_056]

I know it sounds impossible, but I am working on a spreadsheet that show all character matchups out of ten rounds.

It will take a while. when I get some more basic things entered, I will encourage support, even on one section.

These entries will be like a chart where it goes as winsof(fill in top characters name)/win of(leftcharacters name)
These character results will be created by doing level 9 AI vs. Level 9 AI battles with two stock on omega stages.
Mii gunner default, Mii swordfighter default, and Mii brawler default get their own slots
This means no equipment and 1st slot moves

This spreadsheet will be posted later, but you can in the meantime tell me results. I already have most Mario results, so don't
Do his matchups

I will input more soon, but in the meantime you can input your opinions on this and your results if you agree to my ideas.
This is also possibly helping for tiers, since 99% of them have too many flaws to be credible

This level 9 AI thing is for fair game, equal skill, to do this.
I have a copy if the spreadsheet gets trolled, so don't even try it.

This system is by no means perfect, but this is the only way for truly equal character matchups. For instance, mario had 6-4, the 4 being bowser. So yeah

NOTICE: I need you to read ALL of my posts before a reply

 

[Poxnixles]

CPUs can't possibly know ATs and real ins-and-outs of the characters they're assigned.

 

[zackattack_056]

I dint say it was perfect, just AI vs AI

As far as AT, I don't know what that means

 

[zackattack_056]

If not then what do you want me to do for equal "talent" because tournament tier lists can vary from skill of only specific characters but thanks for your input at least

 

[Indigo Jeans]

AI vs. AI is a terrible way to go about this, simply because AIs will never be able to perform a character to the fullest of its abilities. For example, the Rosalina AI is absolutely horrid and would probably end up losing well over 80% of its matches, which would create a matchup chart proving to be terribly inaccurate when compared to even the most mediocre human Rosalinas.

 

[zackattack_056]

If replies lack, be patient. I am forgetful and I sometimes forget things. Thankfully, It won't take long for me to be reminded of it

Not me, I suck at rosalina

 

[zackattack_056]

True, but how else should I go about it? I can fix the chart to work, I just need to know. For instance, average ten peoples for glory % out of ten matches and write down the character fought against to find specific %'s. Actually, if it took less work(avg. 5 instead?) That might not be so bad

 

[Indigo Jeans]

Individual character MU charts are normally done by receiving input from the character mains and deliberating for a week or more then coming to a consensus, also taking into heavy consideration how those who are considered as the characters' "top players" feel about and perform in the MU. Entire MU charts are normally done by the specific game's back room.

Also, it isn't a good idea to double post so much. You should edit your posts instead.

 

[zackattack_056]

Ok thanks for that, working on stopping that

So you mean like how me playing as villager and can't beat little mac, so little mac is a good pick against villager?

Also, I noticed people put rosalina on smash 3ds tier list tops, even though they need To take screen size (being able so actually pay attention To just you)

Also just random, has anybody beat level 9 classic mode with villager? I've almost done it, but so far just 8.5 has been completed. Also, is global smash power flawed?

 

[DanGR]

Welcome to Smashboards.

A tier list decided by CPU matchups would be a CPU tier list, and nothing more. More traditional tier lists like the ones you find scattered throughout smashboards use character matchups decided through theoretical analysis of character vs. character interactions guided by humans using the characters in question. This means human reaction times, timing, and physical limits apply to these matchups. Of course, CPUs don't possess any of these and many other human traits that heavily influence matches; therefore, CPU matchups are irrelevant.

 

[Sleek Media]

Guys, chill out. He just wants to make a spreadsheet. It's not a scientific journal.

 

[zackattack_056]

I agree with the cpu vs cpu being flawed, but stop saying it! also, while it is, this is for even matchups, not weither skilled or not. Remember, the AI is set to level 9, so the most strategy possible while being even is possible. The only other way to improve AI is amiibo, which actually does "learn" but it isn't out yet, and even then not all will be released. Human players aren't usually even, so how would I judge it? The usual way is percise, but it would take even more effort and it would need to be voted upon.

+sleek media thanks, but I was trying to make a matchup chart to work with for a tier list

 

[~ Gheb ~]

http://smashboards.com/threads/result-based-matchup-chart.367712/

 

[Shaya]

The results may in fact be interesting, but they could very well be inconsistent still.
AIs in Brawl "learned" from human play patterns, if you sat in a game spamming falcon punch on falcon as your only move for long enough, you'd be finding your CPU Falcon's using a lot more falcon punch.
This "feature" is being hyped up as an amiibo thing, but I'd assume it still exists for regular AI (why wouldn't it? Maybe due to technical limits on 3DS), meaning that the same 10 matches per character (and for this to cover every match up you're looking at 25,000 matches, that's insane) would have different results on another person's 3DS, hell even your own 3DS after a period of time.
Another terrible issue is the fact that I'm pretty sure (like in the past) when you pit together two CPUs, the game decides randomly which will have the more "dominant" AI given to it. Maybe different in Smash 4, but this in itself kinda invalidates any data collection ability.

In short, it's a novel idea, and I recall someone having the same idea/making the same topic early on in the Brawl days, but it never eventuated and is ultimately not very helpful for learning anything other than which AIs in the game seem to be more functional than others.

 

[zackattack_056]

Gheb, don't use tournament results, people can main in specific characters and mess it up. At very least, I would do for glory style 1v1 battles, or else a tounament based matchup/tier chart could be screwed up by equal characters barely beating out others on the 1st round. If villager went agianst a spam duck hunt and lost, is that actually fair game since he usually never used an attack? While you are doing kills, barely wins or accidents on a lloid rocket, giving him 0 deaths to villagers 2 deaths, would be unfair to other players.

Ok so I agree, ai vs ai opinions, I need to know WHAT ELSE TO DO, because this is a result based chart, not a consensus based chart, to find how many times out of ten a character would normally win

 

[~ Gheb ~]

Right, I forgot that the lvl 9 CPU has fundamental knowledge of his and his opponent's options as well as brilliant decision making skills, unlike top level players who "mess up" a lot. Sorry about that.

 

[Sleek Media]

Honestly it's too early to make an "accurate" matchup chart. Most people still think Megaman is low tier haha. Just go with whatever you think is fun, and make a note of the criteria you used. Even if it's just AI, we might learn something new.

 

[Wiley]

This may work for finding out who's Amiibo may have better chances... but let's be real, as many people have already (both seriously and sarcastically stated) this isn't enough for "result" based data, as you want. Result based data will be more accurate, when we have more results... more tournaments and testing. Just be patient and stop trying to jump the gun by cutting corners with far lesser versions of each fighter.

 

[Forsage]

Ok so I agree, ai vs ai opinions, I need to know WHAT ELSE TO DO, because this is a result based chart, not a consensus based chart, to find how many times out of ten a character would normally win

Hmm... So you want to determine how many times out of ten a character would normally win? As you've already said, when player skill varies, as it always does, then you can't really extrapolate the number of wins between two specific players to all other players. The best way to go about it then, is to take an average of as many approximately equally skilled players as you can.

Now, it's worth noting that characters can have completely different win rates depending on the skill levels of the players. I'm sure we're all familiar with how OP Kirby is at lower levels due to his invincible rock. At higher levels though, not so much, and I'm sure you'd see a drastic change in Kirby win percentage (at least in previous games, not necessarily Smash 4). So who do you want this data to apply to? You can take results on For Glory for instance, and you'd get a set of data representing a wide range of low to mid level skilled players, only playing on completely flat stages. Alternatively you can take tournament results, and get data representing mid to high level skilled players, playing on all competitive stages.

I, personally, am of the opinion that looking at tournament results is the better option. This is due to both
a) Players wanting to know how good a character is at their full potential, (High tech skill, all stages, no online lag, etc.)
b) A smaller deviation from the mean in mid-high level players then low-mid level players. While there will be exceptional players at tournaments who win despite bad matchups, most players will be approximately the same skill level. You can have horrible players online though facing even a half-decent player who could win with basically any character they choose. This skews your results and makes extrapolation inaccurate.

So, if we're both of the opinion that tournament results are the best, then how should you go about collecting that data? And here's the main reason there hasn't been tier lists or matchup charts constructed yet, a lack of data. The Smash 4 scene has barely begun, and there's only been a few tournament results to look at. With that said, you can certainly begin the process of collecting data.

Now, this is nice and all, but Gheb is basically doing this already. To avoid duplicating his work, which would be a waste of your time, I'd suggest you make a theory based matchup chart, which will likely differ from the actual matchup chart a fair amount. To do this, you can go around the character boards and see who's complaining about who. Because it would be impossible to quantify this kind of qualitative data, you should stick to a simple, Character A has a (good/bad/neutral) matchup against Character B. You can start with every character having a neutral matchup against every other character. Then change the matchups as you discover who's complaining about who.

To begin with, as a Zelda main, I can attest that she has a poor matchup against Little Mac, based upon the majority of her moves having strong hits, but lots of lag, making them easily punishable by Mac. Even on platform stages its still pretty bad, because Zelda doesn't have good attacks that hit downwards. On the other hand, Zelda seems to have a pretty good matchup against projectile characters like Link or Duck Hunt, due to her teleport being a long-ranged attack against them if they get predictable (which they sort of have to be if they want to spam them). Unable to use projectiles, they seem to have a difficult time adapting.

So yeah, that's my general idea on how to do this. If you do decide to go the theory route, remember to get more than a couple opinions on characters. PM any notable forum posters you see if it seems like they'd have a credible opinion on their main's matchups. I hope I helped, and best of luck!

 

[kenniky]

AIs in Brawl "learned" from human play patterns, if you sat in a game spamming falcon punch on falcon as your only move for long enough, you'd be finding your CPU Falcon's using a lot more falcon punch.

That's actually not true. http://www.ssbwiki.com/List_of_flaws_in_artificial_intelligence#Rumors

 

[zackattack_056]

I think I will do it based off of common results in for glory with specific matchups. Such as saying, I can't beat little mac with villager, so little mac is good towards villager. I've decided this can be the only way that is fair, as a common consensus is based off of experience.

So submit your main and your bad/good matchups instead of cpu lvl 9 results

Ultimately conviced by forsage, shaya, danGR, indigo jeans, and gheb_1 (looked at his chart and tried to find a new way to do it)

 

[Shaya]

In which case, there is no relevance to you doing a chart then. Why are wifi for glory results worth anything? You could have an opinion on the little mac match up as villager, but that's why you go to those individual subforums to talk about it and learn them / advance them. You don't just take a grain of what happens IN LAGGY CONDITIONS as gospel.

TBH, this is quickly turning away from being even remotely competitive. You have two choices, I can move this to the online discussion, where perhaps "wifi match ups" is actually worth talking about, or I close this thread.

 

[Sleek Media]

Why don't you guys just skip the drama and put up a "no new members" message on the front page? Cuz there's no way rookies are going to stick around when you're so hellbent on shutting them down.

 

[Shaya]

Well, there's a big stickied top of the page etiquette message that 'covers' it, but either way you're saying we're meant to cater to people's feeling/emotions beyond logic or sense? How or why should we?
None of us are being rude or belittling (analogies are helpful tools for explaining ideas last time I heard). Blunt/direct/logic. If you have issues with this, you're probably not competitively minded in the first place. Which bodes the question, why are you in this section?

If someone wants to make a CPU match up chart and pass it out with competitive merit, they're kidding themselves. I'm sorry that you think we're ostracizing new members or saying "don't bother posting here", but that seems to be your issue, not ours and if you're going to make a habit of standing up for the little guy (when they don't need to be stood up for) then you're just going to get legitimate ire from me.

 

[Sleek Media]

Wow, you are just so elite. How can we be as competitive and cool as you? Get over yourself. Yes this is absolutely the time to stand up for the "little guy", who for all you know could be a huge contributor one day (and he seems to have the intention). Your attitude is garbage, and you are wrong for posting what you did. Go ahead and mod me if you want. Show the kid how we do things at smashboards.

This is a competitive subforum. A competitive subforum filled with people (including both of us) who know that it's far too early to make tier lists, complete matchup charts, etc. So what are we going to do, not post anything that you can't nerd out over with numbers? You do remember Super Smash Bros Melee, right? The game where we thought Jigglypuff was bottom tier for about five years until oops turns out all our "analysis" completely missed? Or how about Young Link, who rose even later?

What is your argument? The CPU doesn't know enough advanced techs? How long did it take to find wavedashing again? The CPU sucks? How good are our tourney players in their first year vs their fifth? What argument can be made against the suggestions he's made that can't be applied to our current lack of skill?

It doesn't matter how long you wait to start gathering data for this kind of list. No matter what logic you try to apply, whatever you come up with is going to be wrong. The sooner you have the discussion, however, the sooner you can start figuring out what is wrong and why, and before long you have something that actually works a little bit. Even if this guy's list turns out to be completely upside-down, if it spurs us on to learn something new, then it is a worthwhile contribution.

Instead of tearing this guy down and threatening me, why don't you set a positive example for competitive players, and use your knowledge to help him refine it as much as possible?

 

[Shaya]

Wow, you are just so elite.

I've been around for a very long time, and it's not because of this that I try to enact "eliteness", but rather, the experience as a moderator requires for there to be standards that are upheld to avoid us devolving into something not worth competitive players reading or posting in. There's a lot of positive and thought provoking discussion that can exist here, without numbers/nerding it up (why fill your post with so much irrelevant... truth..?)

Yes this is absolutely the time to stand up for the "little guy", who for all you know could be a huge contributor one day (and he seems to have the intention).

So what? We all started somewhere, and we aren't telling this person to stop/leave and never come back. If you're interpreting otherwise you have serious issues with your comprehension.

Your attitude is garbage, and you are wrong for posting what you did. Go ahead and mod me if you want. Show the kid how we do things at smashboards.

How is telling someone to go to the villager board to discuss match ups they "play" wrong? Why is saying that using wifi for glory results as a match up chart is wrong for the competitive section of the board that is dedicated NOT TO FOR GLORY OR ONLINE PLAY AT ALL (and hence "if they want me to" I can move it there)? Every character board discusses match ups, so perhaps he isn't even aware that this is the case. My tone may not come off as pleasant, but that is nothing of concern to anyone wanting to get anywhere competitively. You're here for a reason, to learn and to get better.

Instead of tearing this guy down and threatening me, why don't you set a positive example for competitive players, and use your knowledge to help him refine it as much as possible?

I'm pretty sure that's what every post in here replying to him has done, except you don't seem to understand this and have garnered an emotional frame to your argument. "Tearing him down" is an emotional perspective and was hardly the intent; if he is upset then I will apologise and clarify to him personally without any issue. You can harp on "THE MAN" all you want, but I'm telling you now, you're only wasting everyone's time (your posts have been significantly less constructive than any one 'tearing him down'). I'm not sure how much you want to stand on a soap box before you get the point, but the threat is 'I'm not going to sit around and let you soap box'. We both know how your post was framed to get the "threat" response you so yearned for, and considering the length of time you've been on these boards apparently, you would know what the smart way to deal with clashes in ideologies are -

If you want to continue this conversation, feel free to private message me.

 

[Sleek Media]

I'm pretty sure that's what every post in here replying to him has done, except you don't seem to understand this and have garnered an emotional frame to your argument. "Tearing him down" is an emotional perspective and was hardly the intent; if he is upset then I will apologise and clarify to him personally without any issue. You can harp on "THE MAN" all you want, but I'm telling you now, you're only wasting everyone's time (your posts have been significantly less constructive than any one 'tearing him down'). I'm not sure how much you want to stand on a soap box before you get the point, but the threat is 'I'm not going to sit around and let you soap box'. We both know how your post was framed to get the "threat" response you so yearned for, and considering the length of time you've been on these boards apparently, you would know what the smart way to deal with clashes in ideologies are -

Projecting much?

It's very simple. You're either interested in helping, or you're not. I'm happy to contribute Mario and Mega Man information for now. What will you be contributing?

 

[Shaya]

Not anymore than you.

If people want my perspectives/match up results based off of Wifi stuff? Sure, I play enough of multiple characters, but I tend to do so in their respective character boards. But as I said, in the case of it going the way he wants it too, it is likely better off in the online discussion, as that is where there is most relevance to it; we otherwise do have projects running (as have been noted by others) that correlate to his goals but with live tournaments in mind. I want to clarify that I agree I may not have worded that as nicely as I could have, but even in hindsight my attitude/justification doesn't change, just the perspective.

 

[Ryik]

Competitive matchups cannot be accurately determined by AIs. Truly, a level 9 CPU is very bad compared to a competitive player. This also applies to For Glory. Competitive matchups should not be centered on low-level or even mid-level play; high level play is what makes it competitive, and both For Glory and AIs do not meet that criteria. At best some matches in For Glory would be legitimate, but the matches that are not would taint your data set.

You could certainly make a CPU matchup list, as well as a "layman's matchup list", (what Shaya calls a wifi matchup chart) but neither would be a true competitive matchup list. Such lists are determined by tournament results or by high level players yatta yatta Shaya has said this already. My point is that this simply isn't competitive.

Gheb, don't use tournament results, people can main in specific characters and mess it up. At very least, I would do for glory style 1v1 battles, or else a tounament based matchup/tier chart could be screwed up by equal characters barely beating out others on the 1st round. If villager went agianst a spam duck hunt and lost, is that actually fair game since he usually never used an attack? While you are doing kills, barely wins or accidents on a lloid rocket, giving him 0 deaths to villagers 2 deaths, would be unfair to other players.

Ok so I agree, ai vs ai opinions, I need to know WHAT ELSE TO DO, because this is a result based chart, not a consensus based chart, to find how many times out of ten a character would normally win

Tournament results are the results of players adept in their characters going up against players adept in their characters and seeing who comes out on top. Human skill along with human error trumps computer skill and poor decisions any day. If they came out on top by a narrow margin, that doesn't make the result any less legitimate, it just means the matchup advantage is smaller, according to the track record of which we are basing the matchup chart on. For Glory insinuates a smaller skill level, as anyone at any skill level can hop onto it. By very definition, competitive play means that it only includes competition-level players. This is what they mean by competitive tier and matchup lists, and this is where people get confused. Tier lists do not mean much anything at lower-level play, and thus beating 3 level 9 CPUs with a Ganondorf in Brawl doesn't change the definition of these things because the computers are fairly bad.

If a villager went against a spam duck hunt and lost and never uses an attack, assuming the skill levels of both players were appropriate, then yes, because it would show that the villager's pocket is inadequate and that it doesn't have enough tools to deal with said duck hunt. It is perfectly fair, but it does not mean that all villager players are like that. This is the basics of scientific experiments- if a pattern persists, then the mean should by all means trump outliers to reflect the pattern properly. It is too early yet of course, because there haven't really been much tournaments to cull data from yet, and this is why it is said that it's too early to make competitive matchup charts. (Among other reasons) There is really nothing that can be said to defend the inclusion of CPU ability in competitive consideration.

I think I will do it based off of common results in for glory with specific matchups. Such as saying, I can't beat little mac with villager, so little mac is good towards villager. I've decided this can be the only way that is fair, as a common consensus is based off of experience.

So submit your main and your bad/good matchups instead of cpu lvl 9 results

Ultimately conviced by forsage, shaya, danGR, indigo jeans, and gheb_1 (looked at his chart and tried to find a new way to do it)

What you are describing is the layman's/wifi matchup chart. The level of play here is below that of competitive, meaning you intend to say "Most players tend to do better with (blank) against (blank)". You are asking for any random person to weigh in on matchups. I don't mean to sound elitist here, but that's not how it works.

If you want matchups/tier lists anyone can edit, you might wanna try EventHubs or a different forum section.

Otherwise, for a tier list or matchup chart to have legitimacy, it must be made with the authority of players better than the average player, usually having proved it in a tournament setting. If CPUs on 9.0 Difficulty give you problems, then this is likely not the case. This is what gives matchup charts and tier lists authority; because they came from a reliable source who has experience playing against the best players of such characters.

Take a look at this.

Now, you see, when someone says that something isn't good enough for competitive play, they (usually) aren't trying to be elitist. That up there is competitive play, and as you can see, they play on a very different level. Seeing it, how can you trust that you know as much as they do at that level? I certainly wouldn't say that I'm as good as them.

Wow, you are just so elite. How can we be as competitive and cool as you? Get over yourself. Yes this is absolutely the time to stand up for the "little guy", who for all you know could be a huge contributor one day (and he seems to have the intention). Your attitude is garbage, and you are wrong for posting what you did. Go ahead and mod me if you want. Show the kid how we do things at smashboards.

This is a competitive subforum. A competitive subforum filled with people (including both of us) who know that it's far too early to make tier lists, complete matchup charts, etc. So what are we going to do, not post anything that you can't nerd out over with numbers? You do remember Super Smash Bros Melee, right? The game where we thought Jigglypuff was bottom tier for about five years until oops turns out all our "analysis" completely missed? Or how about Young Link, who rose even later?

What is your argument? The CPU doesn't know enough advanced techs? How long did it take to find wavedashing again? The CPU sucks? How good are our tourney players in their first year vs their fifth? What argument can be made against the suggestions he's made that can't be applied to our current lack of skill?

It doesn't matter how long you wait to start gathering data for this kind of list. No matter what logic you try to apply, whatever you come up with is going to be wrong. The sooner you have the discussion, however, the sooner you can start figuring out what is wrong and why, and before long you have something that actually works a little bit. Even if this guy's list turns out to be completely upside-down, if it spurs us on to learn something new, then it is a worthwhile contribution.

Instead of tearing this guy down and threatening me, why don't you set a positive example for competitive players, and use your knowledge to help him refine it as much as possible?

Huge contributor or not, this is not a proper competitive matchup chart in any way. It's not so much "You're wrong" or "You can't do this" as it is miscategorization. What zackattack is describing is simply not a competitive matchup chart. There is no need to be insulted.

Also CPUs are just plain bad. Take their difficulty and halve it. At best, I'd call 9.0 difficulty CPUs a 4.5/10. That's not competitive play.

 

[The Slayer]

All I'm going to say (like others has mentioned) is you can make a CPU matchup chart. However, not sure how that will fit in the competitive scene since they're fighting humans than CPUs. And even then, it's a lot of work with little to no importance. I don't see what's the point for it really.

 

[Deluxe0111]

This might be overcomplicating things and its just an idea but..

You could use the spectator mode, ignoring all games but 1v1s. However this would have to include the stats the mode gives you and then using a bit of good old maths you could turn it into a meaningful chart of some sort
(Given some time I could work out the relative percentages etc)

Granted this would take ages because you never know what characters your going to watch

 

[6thGodBillTrinnen]

As many others have explained, I just don't think this will give you any useful data for real world matchups. Matchups aren't something that you can really know too well this early in the meta and with so little tournament results to base it off of.

Also keep in mind that excluding dittos, there are 1275 matchups in this game

So say you do something like 5 games per matchup, then you're looking at a ridiculous 6375 games to get all of the data! I would say don't even bother.....a meaningful matchup chart isn't a task for one person, it takes the entire community's input and many, many months worth of tournament results. Simply pitting AI against each other would take an incredible amount of work for something that will inevitably be inaccurate.

 

[BBC7]

Don't waste your time making a Level 9 Match-up Chart. Luigi was one of the best Level 9's in Brawl and Meta Knight was one of the worst. That is pretty much the opposite when it comes to human players, because human players actually have minds and knowledge.

 

[zackattack_056]

UGH read ALL THE POSTS not just the 1ST

BTW ok I accept input, as this is my 1st time doing this, which is purely experimental.

The for glory thing is what still stands, as people who focus on one character can definetly give opinions on what won't work against other characters (villager v little mac, trust me on this, pocket won't even work on any attack and has too many high % attacks and ability to launch) but this input is not purely for glory, just for glory battle standards. Also, tell me reasoning along with opinions about it. This is shown like in the parenthesis earlier.

Thanks to sleek media for sticking up for me most of the people I actually physically have met that would never do that

Bbc7 I agree, but his ai didn't involve spamming tornados. Read all posts and know this bad ai has been realised by me, get over it.early? It been a month, so I wanted to try (rosalina is NOT top tier)

Don't say I suck at villager, it wasn't my example and I beat master core at 8.5 (working on 9.0, as villager is light)

Ryik, you really care a lot about proving me wrong, and I can't prove you that wrong, so congrats