Can someone describe to me, in vivid detail, how a tier list is formed?

[Muhznit]

Please tell me what data gets collected from tournaments and details on how it's processed into the tier list we've all grown to love/hate.

I keep hearing that it's based on who most frequently wins tournaments, but I also hear that there's a group vote involved. In the case that silly (IMO) subjective methods like voting are used to determine something as important as a tier list, I think I miiight have stumbled into a reliable strictly data-driven algorithm that I'd like to mess with as a bit of a mental exercise and show to others if I think it's of merit.

 

[simpleglitch]

Please tell me what data gets collected from tournaments and details on how it's processed into the tier list we've all grown to love/hate.

I keep hearing that it's based on who most frequently wins tournaments, but I also hear that there's a group vote involved. In the case that silly (IMO) subjective methods like voting are used to determine something as important as a tier list, I think I miiight have stumbled into a reliable strictly data-driven algorithm that I'd like to mess with as a bit of a mental exercise and show to others if I think it's of merit.

It depends on who's tier list, but the ones people actually pay attention to around here are crate by compiling tournament results.

Because the game is new, and there is very little tournament data at the moment, there will probably be a lot of different tier lists that pop-up based on people's "Best guesses." But as more data is collected, those lists will die out.

That probably isn't as vivid detail as you want, but besides finding out what kind of formulas they run the collected data against I don't think there is too much more than that.

 

[Fafnir]

It's a combination of a bunch of things. Tournament results are a great way to get a general gist of things, as well as practical applications of match-up data and the extent of top play levels. That said, there's a lot more that goes into it. Hard game data and theoretical information also needs to be applied.

Since the goal of a tier list is to establish the hierarchy of characters at the highest known level of play, the tiers themselves might not entirely coincide with the tournament results themselves. There's also a balance between theoretical and practical applications that needs to be made, since what might be technically possible in the game might not actually be achievable (or achievable on a consistent enough basis to matter for the purpose of tier listing) by a human.

Additionally, tier lists are fluid. New techniques and strategies are discovered all the time, and tier lists themselves are really just an expression of the metagame. A single shift in the meta can cause quite a large change in how the tiers are represented, if that shift is big enough.

 

[Ehn Jolly]

How a tier list is made.

Well when a mama Fox and daddy Metaknight love each other very much...

...and 9 months later, a tier list is made!

 

[KuroganeHammer]

Additionally, tier lists are fluid. New techniques and strategies are discovered all the time, and tier lists themselves are really just an expression of the metagame. A single shift in the meta can cause quite a large change in how the tiers are represented, if that shift is big enough.

It's worth noting that tournament play can also shuffle tiers around. Sometimes people discover something about a low placed character on a tier list that other characters don't have an answer to.

An example in Brawl would be Olimar, when people discovered that they did in fact have a grab button.

 

[XDaDePsak]

People yell at eachother on the internet until people decide enough is enough and someone writes a final list and everyone shrugs and someone posts it on wikipedia.

 

[Raijinken]

Tournament appearances and results play a notable role, but theory goes a long way as well. Ice Climbers are too rare (despite their power) to be so high tier if you totally ignore the theory and discussion that goes alongside data collection.

 

[KaZe_DaRKWIND]

I think it has to do with matchups. The ones that win the most in matchups are higher on the list. Could be wrong, it's just how I understand it.

 

[Fafnir]

I think it has to do with matchups. The ones that win the most in matchups are higher on the list. Could be wrong, it's just how I understand it.

That's definitely a huge part of it. Although in a game like the previous Smash titles, where the list tends to be so heavily polarized, match-ups against characters already established as being higher tier within the meta tend to be worth more.
There's also the weight of the match-ups. A character with an obviously 20/80 match-up is going to suffer a lot more than a character with a bunch of 45/55 match-ups.

 

[Muhznit]

That's definitely a huge part of it. Although in a game like the previous Smash titles, where the list tends to be so heavily polarized, match-ups against characters already established as being higher tier within the meta tend to be worth more.
There's also the weight of the match-ups. A character with an obviously 20/80 match-up is going to suffer a lot more than a character with a bunch of 45/55 match-ups.

I'm more interested in the "hard game data" you mentioned. Specifically whatever functions, variables, etc go into it. If it takes in some kind of input, and spits out some kind of output, I want to know it. I'm not afraid of numbers, I mean I've poked around some of Brawl's files before P:M was a thing and I do programming for a living. I'm quite curious here...

 

[SevenYearItch]

I has a lot to do with results because you can have a character so low on the tier list, but have one player win a few notable tournies with them and jump the ranks so quickly.

 

[Fafnir]

I'm more interested in the "hard game data" you mentioned. Specifically whatever functions, variables, etc go into it. If it takes in some kind of input, and spits out some kind of output, I want to know it. I'm not afraid of numbers, I mean I've poked around some of Brawl's files before P:M was a thing and I do programming for a living. I'm quite curious here...

Frame data. As in, the data for the timing and hitboxes of characters and their moves.

There's no way to put in an objective algorithm for calculating tier placement. There are too many situational elements to the game, as well as the fluidity of the metagame and dependence on mind games for such a thing to really work.

 

[Amazing Ampharos]

Here's how it actually worked in Brawl at least when I was in the BBR.

At certain points in time decided by the BBR leadership, a vote would be held. There would be a lengthy backroom discussion, and then everyone would vote for the characters in an ordered list from best to worst. Something like...

1. Meta Knight
2. Snake
...
36. Zelda
37. Ganondorf

Among all BBR voters, the average placement of each character would be computed to form the tier list. There was some degree of BR discussion on where the tiers were divided, but it was typically not a contentious issue since usually the votes would have easy points where breaks could be formed.

In terms of the "why" of placing any given character above another, that was largely subjective based on the interpretation of the voters about what constituted a good or bad character. The BBR leadership provided some guidelines that operated as a suggestion for what to think about when voting as I recall, but there wasn't any actual policing of "your vote doesn't meet the criteria and therefore doesn't count" other than that you'd probably be called out if your vote was really stupid (it wasn't private). In general, it tended to do a very good job of exactly what you'd expect from this kind of system: tell you the average opinion of high level players, TOs, and community notables as per the game's character balance in terms of the ability of characters to see success at a tournament level.

I believe every tier list ever officially put out by Smashboards was definitively imperfect, but they were all considerably better information than most people would be able to make on their own so they were probably about as good as they could possibly be. Any tier list should always be taken with a grain of salt and full recognition that it doesn't actually affect the game (if your favorite character is given a really low tier position, it doesn't actually make them any less capable of winning than they were before the list was published...), but disregard tier lists completely at your own peril since as games develop it will get increasingly hard to win with bad characters. Also recognize that tier lists tend to be more accurate in representing the strength of higher tier characters since they are simply far more explored and the entire voting pool will have extensive experience dealing with high tier characters. This means that if a tier list says Meta Knight is #1 and Snake is #2 that it's probably making a strong statement that Meta Knight is better than Snake. If it says Captain Falcon is #33 and Link is #34, it's not saying much at all about whether Captain Falcon or Link is better. That being said the low tier placements aren't entirely useless; you're getting a good sense that the character is probably actually low tier just a poor sense of how that low tier character stacks up to the other low tiers.

 

[Muhznit]

Frame data. As in, the data for the timing and hitboxes of characters and their moves.

There's no way to put in an objective algorithm for calculating tier placement. There are too many situational elements to the game, as well as the fluidity of the metagame and dependence on mind games for such a thing to really work.

That specific data strikes me as a bit difficult to quantify. At least, for a human. But I wouldn't say that there is no way to make an objective algorithm just yet... or at least one that can be explained by purely numbers and probability.

Here's how it actually worked in Brawl at least when I was in the BBR.

At certain points in time decided by the BBR leadership, a vote would be held. There would be a lengthy backroom discussion, and then everyone would vote for the characters in an ordered list from best to worst. Something like...

1. Meta Knight
2. Snake
...
36. Zelda
37. Ganondorf

Among all BBR voters, the average placement of each character would be computed to form the tier list. There was some degree of BR discussion on where the tiers were divided, but it was typically not a contentious issue since usually the votes would have easy points where breaks could be formed.

In terms of the "why" of placing any given character above another, that was largely subjective based on the interpretation of the voters about what constituted a good or bad character. The BBR leadership provided some guidelines that operated as a suggestion for what to think about when voting as I recall, but there wasn't any actual policing of "your vote doesn't meet the criteria and therefore doesn't count" other than that you'd probably be called out if your vote was really stupid (it wasn't private). In general, it tended to do a very good job of exactly what you'd expect from this kind of system: tell you the average opinion of high level players, TOs, and community notables as per the game's character balance in terms of the ability of characters to see success at a tournament level.

I believe every tier list ever officially put out by Smashboards was definitively imperfect, but they were all considerably better information than most people would be able to make on their own so they were probably about as good as they could possibly be. Any tier list should always be taken with a grain of salt and full recognition that it doesn't actually affect the game (if your favorite character is given a really low tier position, it doesn't actually make them any less capable of winning than they were before the list was published...), but disregard tier lists completely at your own peril since as games develop it will get increasingly hard to win with bad characters. Also recognize that tier lists tend to be more accurate in representing the strength of higher tier characters since they are simply far more explored and the entire voting pool will have extensive experience dealing with high tier characters. This means that if a tier list says Meta Knight is #1 and Snake is #2 that it's probably making a strong statement that Meta Knight is better than Snake. If it says Captain Falcon is #33 and Link is #34, it's not saying much at all about whether Captain Falcon or Link is better. That being said the low tier placements aren't entirely useless; you're getting a good sense that the character is probably actually low tier just a poor sense of how that low tier character stacks up to the other low tiers.

Great, this is the type of answer I was looking for. Mostly. I'm not quite looking to disregard the tier list, but I always had a general dislike for even the slightest degree of opinion in attempts to quantify/rank/sort things. I'm weird like that. Matchup ratios are generally more useful to me, and through matchup ratios (based on tourney results), I think a better system can be made than the vote-based tier list.

It's based on a theory that those matchup ratios, if typically refreshed over a long enough period of time (and determined strictly from matches instead of discussion), contain the end results of all useful competitive knowledge for a character. Any ATs, special tricks, counters, all of them are compacted into those two little numbers. Through that abstraction, no one really needs to consider whatever combos or shenanigans that character can pull. If there's something significant that they have over another character, it shows up as a hard counter or something similar. That 90:10 matchup for Ice Climbers vs Ganondorf for instance, is proof enough of that Infinite Chain throw of theirs they have him. Through some careful merging of all of a character's matchup ratios (not a simple average), a probability of their general performance can be made to quantify their overall performance. I can't say I have much to show other than that summary, but I'll come back to elaborate more if I find some interesting correlations.

 

[Fathom778]

A bunch of "Pro Players" turn off all items, get on Final Destination, and play a character a bunch and say their thought on if their good or not. Then EVERYONE goes to the "Top Tiers".

 

[EgeDal]

I don't know really, but i think it's worthless, just play as the character you want and enjoy. Who cares about tiers?

 

[AttackstorM]

That specific data strikes me as a bit difficult to quantify. At least, for a human. But I wouldn't say that there is no way to make an objective algorithm just yet... or at least one that can be explained by purely numbers and probability.



Great, this is the type of answer I was looking for. Mostly. I'm not quite looking to disregard the tier list, but I always had a general dislike for even the slightest degree of opinion in attempts to quantify/rank/sort things. I'm weird like that. Matchup ratios are generally more useful to me, and through matchup ratios (based on tourney results), I think a better system can be made than the vote-based tier list.

It's based on a theory that those matchup ratios, if typically refreshed over a long enough period of time (and determined strictly from matches instead of discussion), contain the end results of all useful competitive knowledge for a character. Any ATs, special tricks, counters, all of them are compacted into those two little numbers. Through that abstraction, no one really needs to consider whatever combos or shenanigans that character can pull. If there's something significant that they have over another character, it shows up as a hard counter or something similar. That 90:10 matchup for Ice Climbers vs Ganondorf for instance, is proof enough of that Infinite Chain throw of theirs they have him. Through some careful merging of all of a character's matchup ratios (not a simple average), a probability of their general performance can be made to quantify their overall performance. I can't say I have much to show other than that summary, but I'll come back to elaborate more if I find some interesting correlations.

I think largely looking at tournament performance is a good way. This is probably the best environment for test data since the players have the best chance of playing at 100% of their ability and playing with their main characters. This will showcase how a character is played in general to the best of human ability with enough good tournament data because there is no other place where a character can be measured most efficiently and at the highest levels of human ability since in no other place you have a collection of players taking their matches seriously enough, also with the factors of adrenalin boosts and them largely playing their best characters.

Just on a side note just based on all the pre-north american release tournaments I personally have ever heard of I will tell you my personal opinion of the top 5 as of now based off of Grand Finals so far:
1. Ness
2. ZSS
3. Bowser
4. Little Mac
5. Greninja

Others on my personal radar based on Grand Finals are: Kirby, Zelda, Pac-Man, Robin

 

[Hitzel]

In most fighting games, a tier list is formed by looking at each 1v1 character matchup in the game, and figuring out if a character has an advantage and if so how much. This is determined by asking the question "If two equally skilled high-level players played ten games as characters A and B, character A would win X times and character B would win Y times. Examples are things like "Cammy vs Rufus is 6-4, Cammy," "Sagat vs Zangief is 8-2, Sagat," etc.

In the end, the matchups for each character are added up and produce something that looks like this:
(although the decimal places aren't that common)

If a fighting game is something like Marvel where there are teams of characters and therefore too many matchups to make a chart, this system is usually abandoned and replaced with something even more subjective. I've never seen any kind of official chart like this for Smash's tier lists though, which is kinda puzzling since people always say matchups are the most important factor. I'd be interested to see what a Smash matchup chart would look like and how closely it would resemble the established tiers.

 

[Jack Kieser]

Here's how it actually worked in Brawl at least when I was in the BBR.

At certain points in time decided by the BBR leadership, a vote would be held. There would be a lengthy backroom discussion, and then everyone would vote for the characters in an ordered list from best to worst. Something like...

1. Meta Knight
2. Snake
...
36. Zelda
37. Ganondorf

Among all BBR voters, the average placement of each character would be computed to form the tier list. There was some degree of BR discussion on where the tiers were divided, but it was typically not a contentious issue since usually the votes would have easy points where breaks could be formed.

In terms of the "why" of placing any given character above another, that was largely subjective based on the interpretation of the voters about what constituted a good or bad character. The BBR leadership provided some guidelines that operated as a suggestion for what to think about when voting as I recall, but there wasn't any actual policing of "your vote doesn't meet the criteria and therefore doesn't count" other than that you'd probably be called out if your vote was really stupid (it wasn't private). In general, it tended to do a very good job of exactly what you'd expect from this kind of system: tell you the average opinion of high level players, TOs, and community notables as per the game's character balance in terms of the ability of characters to see success at a tournament level.

I believe every tier list ever officially put out by Smashboards was definitively imperfect, but they were all considerably better information than most people would be able to make on their own so they were probably about as good as they could possibly be. Any tier list should always be taken with a grain of salt and full recognition that it doesn't actually affect the game (if your favorite character is given a really low tier position, it doesn't actually make them any less capable of winning than they were before the list was published...), but disregard tier lists completely at your own peril since as games develop it will get increasingly hard to win with bad characters. Also recognize that tier lists tend to be more accurate in representing the strength of higher tier characters since they are simply far more explored and the entire voting pool will have extensive experience dealing with high tier characters. This means that if a tier list says Meta Knight is #1 and Snake is #2 that it's probably making a strong statement that Meta Knight is better than Snake. If it says Captain Falcon is #33 and Link is #34, it's not saying much at all about whether Captain Falcon or Link is better. That being said the low tier placements aren't entirely useless; you're getting a good sense that the character is probably actually low tier just a poor sense of how that low tier character stacks up to the other low tiers.

I just want to add on to and reiterate a bit of this.

In pretty much every other fighting game known to man, tier lists are designed to be and try to be as objective as possible; the entire point of the list is to figure out, on a theoretical level, what tools Character X has against Character Y (and the reverse) and how those tools affect the options that each character has. In most games, the ONLY things that matter are hurtbox size (so, how easy is it for you to get hit: a perfect character would have either no hurtbox ever or a hurtbox of only one pixel on one frame), hitbox size, placement, and speed (so, how effective are your attacks at controlling space: a perfect attacker would have normal attacks that come out in one frame and have hitboxes that cover their character's entire hurtbox plus a circle around them radius the longest disjointed attack in the game), and movement options (how easy and quickly you can move around: a perfect character would have the ability to invincibly teleport to any position on the screen in one frame).

In Smash, most of those things still matter, but because the name of the game is KOs, and not stamina damage, and because movement and positioning is so freeform and not limited to a small subsection of screen (there are no corners, like in Street Fighter, for instance), it's much, MUCH harder to pin these things down. Even adding one platform to the stage screws all these calculations up, because one move may be trash on an even field but godly when the opponent is above you. This positioning aspect makes EVERYTHING less calculable and much more subjective.

So, at the end of the day, we, as humans and not supercomputers, cannot possibly calculate every single scenario in the same way we can calculate every single scenario in a Cammy vs Rufus match. This means that Smash tier lists, by the very nature of the game we are playing, are inherently WAY more subjective than any other game's lists. The best we can do is nail down a voting procedure, but past that, there is NO set in stone, consistent procedure for calculating even ONE matchup in this game. Such a procedure, in fact, cannot exist.

 

[Muhznit]

In most fighting games, a tier list is formed by looking at each 1v1 character matchup in the game, and figuring out if a character has an advantage and if so how much. This is determined by asking the question "If two equally skilled high-level players played ten games as characters A and B, character A would win X times and character B would win Y times. Examples are things like "Cammy vs Rufus is 6-4, Cammy," "Sagat vs Zangief is 8-2, Sagat," etc.

In the end, the matchups for each character are added up and produce something that looks like this:
(although the decimal places aren't that common)

If a fighting game is something like Marvel where there are teams of characters and therefore too many matchups to make a chart, this system is usually abandoned and replaced with something even more subjective. I've never seen any kind of official chart like this for Smash's tier lists though, which is kinda puzzling since people always say matchups are the most important factor. I'd be interested to see what a Smash matchup chart would look like and how closely it would resemble the established tiers.

I had high hopes that chart would be for smash characters at first. But that's pretty much exactly the type of data I'm seeking to use. Though now I think about it, My algorithm's kind of just counting how many cases are there in which the character is the best matchup against another. May need improvement.

I just want to add on to and reiterate a bit of this.

In pretty much every other fighting game known to man, tier lists are designed to be and try to be as objective as possible; the entire point of the list is to figure out, on a theoretical level, what tools Character X has against Character Y (and the reverse) and how those tools affect the options that each character has. In most games, the ONLY things that matter are hurtbox size (so, how easy is it for you to get hit: a perfect character would have either no hurtbox ever or a hurtbox of only one pixel on one frame), hitbox size, placement, and speed (so, how effective are your attacks at controlling space: a perfect attacker would have normal attacks that come out in one frame and have hitboxes that cover their character's entire hurtbox plus a circle around them radius the longest disjointed attack in the game), and movement options (how easy and quickly you can move around: a perfect character would have the ability to invincibly teleport to any position on the screen in one frame).

In Smash, most of those things still matter, but because the name of the game is KOs, and not stamina damage, and because movement and positioning is so freeform and not limited to a small subsection of screen (there are no corners, like in Street Fighter, for instance), it's much, MUCH harder to pin these things down. Even adding one platform to the stage screws all these calculations up, because one move may be trash on an even field but godly when the opponent is above you. This positioning aspect makes EVERYTHING less calculable and much more subjective.

So, at the end of the day, we, as humans and not supercomputers, cannot possibly calculate every single scenario in the same way we can calculate every single scenario in a Cammy vs Rufus match. This means that Smash tier lists, by the very nature of the game we are playing, are inherently WAY more subjective than any other game's lists. The best we can do is nail down a voting procedure, but past that, there is NO set in stone, consistent procedure for calculating even ONE matchup in this game. Such a procedure, in fact, cannot exist.

I'm not seeking to calculate every scenario for a given match between two characters. That's something entirely different in that while I won't say an algorithm to solve it doesn't exist, it IS computationally infeasible. But generating a sorted list of characters based strictly on quantifyable, previously-collected data? Much easier.

 

[Jack Kieser]

I'm not seeking to calculate every scenario for a given match between two characters. That's something entirely different in that while I won't say an algorithm to solve it doesn't exist, it IS computationally infeasible. But generating a sorted list of characters based strictly on quantifyable, previously-collected data? Much easier.

That's the problem though; as other people have also said, any tier list that only compiles collectable information, like tournament wins or number of players used or something like that, isn't telling us what the *true matchup* for any given pair is... only, what we THINK the matchup is based off the information we have now. And even so, that data usually doesn't even sync up with THAT; if a character has high strengths, like Brawl Diddy, but a high learning curve or high skill barrier, like Brawl Diddy, then maybe only one or two players nationwide, or even globally, can come anywhere near using the character "properly", and if those players have bad player matchups with people they see in bracket often (which can happen; one player just reads the other well, and it doesn't matter what characters are being played, one is just better than the other), the data we have from events may not accurately reflect how good a character is as well as if it was widely used and had many representatives.

It's certainly useful to have that kind of info, don't get me wrong, but only as statistical information; that kind of info should STAY AWAY from Tier List formats because it's misleading and incomplete.

 

[Muhznit]

That's the problem though; as other people have also said, any tier list that only compiles collectable information, like tournament wins or number of players used or something like that, isn't telling us what the *true matchup* for any given pair is... only, what we THINK the matchup is based off the information we have now.

No, any tier list that compiles observed data from tournements tells you the closest thing you'll get to an accurate matchup ratio. Actually, it may be more accurate to call it a win:loss ratio.

And even so, that data usually doesn't even sync up with THAT; if a character has high strengths, like Brawl Diddy, but a high learning curve or high skill barrier, like Brawl Diddy, then maybe only one or two players nationwide, or even globally, can come anywhere near using the character "properly", and if those players have bad player matchups with people they see in bracket often (which can happen; one player just reads the other well, and it doesn't matter what characters are being played, one is just better than the other), the data we have from events may not accurately reflect how good a character is as well as if it was widely used and had many representatives.

Humans can only come so close to using a character perfectly. I don't think there's been a single high-level tournament match where someone takes 0 damage. But the people that come closest serve as a lower boundary for the actual skill ceiling of the character. Since those people would be the best currently-existing players of the character, there's no need to care about using the character at his absolute best; there's no one that can. We may as well treat the lower boundary of the skill ceiling as the "known skill ceiling" and work with that. I don't think there is any Ganondorf player in Brawl that can nail every guaranteed Side-B followup, so until it happens, you may as well assume that it doesn't exist. The data will support stuff that DOES exist, and will serve as a general metric of what's to be expected when two characters fight.

Also, Player-specific matchups can be dealt with by collecting data from a large enough sample size to eliminate outliers as much as possible.

It's certainly useful to have that kind of info, don't get me wrong, but only as statistical information; that kind of info should STAY AWAY from Tier List formats because it's misleading and incomplete.

It's not misleading. If the data is actually there, out in the open, free to be verified by any data analysts and to be copied to any person interested in organizing it. There is no being mislead. If the community works together to keep track of these wins and losses at tournaments, and play at least the required (numberOfCharacters * (numberOfCharacters - 1)) / 2 matches to fill a matchup chart (Ditto matches will naturally be 50:50), the data won't be incomplete.

I won't say that it should replace the classic tier list, but something based on purely statistical data instead of human opinion (as educated as that opinion may be) should AT LEAST coexist alongside it at all times. Sort of a "Potential Performance vs. Observed Performance" kind of deal. It doesn't make sense to predict a character's tournament performance based on theory and potential alone. Execution matters too.

 

[Cactusblah]

That SF chart with top players testing the matchups is the best way to do it.

 

[Hitzel]

I just want to add on to and reiterate a bit of this.

In pretty much every other fighting game known to man, tier lists are designed to be and try to be as objective as possible; the entire point of the list is to figure out, on a theoretical level, what tools Character X has against Character Y (and the reverse) and how those tools affect the options that each character has. In most games, the ONLY things that matter are hurtbox size (so, how easy is it for you to get hit: a perfect character would have either no hurtbox ever or a hurtbox of only one pixel on one frame), hitbox size, placement, and speed (so, how effective are your attacks at controlling space: a perfect attacker would have normal attacks that come out in one frame and have hitboxes that cover their character's entire hurtbox plus a circle around them radius the longest disjointed attack in the game), and movement options (how easy and quickly you can move around: a perfect character would have the ability to invincibly teleport to any position on the screen in one frame).

In Smash, most of those things still matter, but because the name of the game is KOs, and not stamina damage, and because movement and positioning is so freeform and not limited to a small subsection of screen (there are no corners, like in Street Fighter, for instance), it's much, MUCH harder to pin these things down. Even adding one platform to the stage screws all these calculations up, because one move may be trash on an even field but godly when the opponent is above you. This positioning aspect makes EVERYTHING less calculable and much more subjective.

So, at the end of the day, we, as humans and not supercomputers, cannot possibly calculate every single scenario in the same way we can calculate every single scenario in a Cammy vs Rufus match. This means that Smash tier lists, by the very nature of the game we are playing, are inherently WAY more subjective than any other game's lists. The best we can do is nail down a voting procedure, but past that, there is NO set in stone, consistent procedure for calculating even ONE matchup in this game. Such a procedure, in fact, cannot exist.

There's no "procedure" for figuring out a Street Fighter matchup other than experienced players showing what tools characters have in a matchup and how it plays out. Not much different than Smash to be honest.

It's odd to me that people can support the idea of a tier list, but not individual matchups.