An Alternative Tier System

[Kaladoga]

Introduction:
So we've all seen the tier lists for previous games in the Smash franchise and it seems like once the meta is established (after people get used to each of the characters) then a tier list can be formed. Of course in this day and age, devs can update their software and with that sort of thing comes rebalances that can throw off tier systems. That's not to say that the tier system isn't good or useful, far from it. I'm just saying that these days it is subject to change whereas Melee is stuck as it is unless it's hacked.

Now as time goes on there will be a tier list for Smash Ultimate no doubt. One based on experience, the meta, etc. But while that is developing, I have an idea for a parallel tier system that might help us as a community to track the power levels of each of the tiers of characters (all 76+ of them).

The Idea:
I'm suggesting that we data-mine Smash Ultimate and look up the variables and base stats of each character and then create a tier list based on how the stats come together. This would be a tier list to compliment the meta-tier list (when it comes out) to help provide raw data in the interest of informing the meta. So on the one hand you'd have the experiential tier list, and on the other you'd have the stat-based tier list. Then you could compare the two to see how much their movesets, and player skill come into play.

Tiers:
Since this game is absolutely brimming with characters, I think once we find a way to look up what stats they each have then we might want to expand the traditional S, A, B, C... tier list in favor of one using pluses and minuses. This way it can help us give the situation a higher resolution of variance and help us compare characters to a higher degree. Ultimately this would look like:

S+:
S:
S-:

A+:
A:
A-:

B+:
B:
B-:

C+:
C:
C-:

D+:
D:
D-:

E+:
E:
E-:

F+:
F:
F-:

By my calculations, with 76+ characters on the roster (while anticipating more later on), this should mean that each tier would have ~3.8 characters. Meaning that some would have 3 and others would have 4 since you can't have a fraction of a character obviously.

Again, the point of this kind of parallel tier system would be to help rank the characters based on their game stats and abilities. So as to help group the characters together more fairly. It WOULD NOT replace the traditional meta tier system we're used to. It'd be a supplement to it.

The Problem:
The issue I have is, we need to somehow get our hands on the data in order to properly create a spreadsheet and compare the characters with each other. I unfortunately am not a data-miner so I've no clue how to get my hands on the base character stats.

 

[ZephyrZ]

How do you cleanly "measure" how a characters stats come together? The only way to do that is to test it. What about stats such as weight which are both beneficial and detrimental? I can't imagine how you'd make an accurate scoring system out of this.

Tier lists can't be made into a hard science, even as a supplemental alternative. There's just far too many variables at play. Tournament results, as imperfect as they are, will always be the closet we can to doing that.

 

[Call_Me_Red]

Tier lists can't be made into a hard science, even as a supplemental alternative. There's just far too many variables at play. Tournament results, as imperfect as they are, will always be the closet we can to doing that.

This. Sometimes characters are better or worse on paper than how they actually play. Look at Chrom for example. Before release everyone thought he'd be trash, now he's in the runnings for best character in the game. The only number that matters for a tier list is tournament placing.

Although I do really like the plus and minus addition

 

[Kaladoga]

How do you cleanly "measure" how a characters stats come together? The only way to do that is to test it. What about stats such as weight which are both beneficial and detrimental? I can't imagine how you'd make an accurate scoring system out of this.

Tier lists can't be made into a hard science, even as a supplemental alternative. There's just far too many variables at play. Tournament results, as imperfect as they are, will always be the closet we can to doing that.

So then we'd need to reverse-engineer some kind of algorithm to model the game.

This. Sometimes characters are better or worse on paper than how they actually play. Look at Chrom for example. Before release everyone thought he'd be trash, now he's in the runnings for best character in the game. The only number that matters for a tier list is tournament placing.

Although I do really like the plus and minus addition

Your post gave me an idea. I think what I'll try doing is running a bunch of tournaments on my switch using the characters and seeing how each character does against each other. Maybe that data will be useful.

And thanks, I thought the plus/minus thing would be helpful too. :3

 

[Call_Me_Red]

Your post gave me an idea. I think what I'll try doing is running a bunch of tournaments on my switch using the characters and seeing how each character does against each other. Maybe that data will be useful.

Do you mean with CPUs? If so don't even bother. The game is new enough that you can find ample footage of any character in a tournament setting. It'd be better to review VODs than to host a tournament without any people.

 

[Kaladoga]

Do you mean with CPUs? If so don't even bother. The game is new enough that you can find ample footage of any character in a tournament setting. It'd be better to review VODs than to host a tournament without any people.

Well remember, what I'm after is the game's variables and stats here. This isn't like the normal tier list. I'm trying to eliminate the variable of player skill, if I can, so as to compare the characters based on their stats. And if they'll all CPUs then it's as similar of a player as I can probably hope for, that way we can make its factor have less variance compared to live players.

 

[Nah]

If data on characters' numbers is what you're looking for, that will be available in time. Kuroganehammer was the go-to source for character values in Smash 4, and while it is not remotely complete for Smash Ultimate at this time, it will be someday and will have a lot of what you're looking for. For the time being, you can look at this spreadsheet Zapp made to get start-up/total frames/landing lag values and combine that with the pages on Kuroganehammer for stuff like weight, air speed, intial dash values, etc (which do have values for the whole roster right now) as something to start with.

 

[Kaladoga]

If data on characters' numbers is what you're looking for, that will be available in time. Kuroganehammer was the go-to source for character values in Smash 4, and while it is not remotely complete for Smash Ultimate at this time, it will be someday and will have a lot of what you're looking for. For the time being, you can look at this spreadsheet Zapp made to get start-up/total frames/landing lag values and combine that with the pages on Kuroganehammer for stuff like weight, air speed, intial dash values, etc (which do have values for the whole roster right now) as something to start with.

Excellent! Thank you very much! It might not be quite what I'm looking for but knowing the frame values and more is definitely a start.

At present I'm still unlocking all the characters but suffice it to say that my scientific curiosity has been piqued.

What I'm now thinking of doing (once I have all the characters) is conducting a study where I have one single character in 2v2 3 stock battles go through a tournament with 9 CPU setting against all the other characters. And then compare how many wins/losses and such they undergo. It might give us something useful to compare to.