Just like being good at fighting games or anything else in life, it's REALLY FUCKING HARD to be good at knowing what you're talking about. If nothing else, math just isn't on your side: there's only one truth, but there are infinite ways to be wrong. Even when you're dealing with a situation with a limited range of errors - say, a tier list - the odds aren't great.
Let's just take good ol' Street Fighter II as an example. The original SFII had all of eight playable characters. By today's standards, that's nothing - give the Marvel franchise another few versions and I guarantee we'll have eight fucking characters per team. But a random guess at the SFII tier list would only have a one-in-forty-thousand-three-hundred-and-twenty chance of being correct. A game with another twenty characters - say, vanilla SFV - makes this game practically impossible. And, okay, not even twitter trolls are dumb enough to make a tier list by just guessing at random. But how much do we KNOW about fighting game tiers, really? How big is the margin for error?
According to Maddy Myers, it's too big:
Well, actually, pretty fucking wrong! To test the accuracy of our tier lists, I've been keeping track of the performance of the characters in SFV relative to both Shoryuken's and Eventhubs' tier lists. Wanna guess how useful those tier lists were in predicting championship winners? NOT VERY! Even including Nash in year one, who was a no-brainer pick for top five, the average error for Evo and Capcom Cup winners is over SEVEN. In other words, that character in the top spot? They could be the best - or they could barely be in the top ten. And that mid-tier character that you love even though everyone tells you to stop playing them? They have the potential to be godlike.
Even the overall margin isn't great - it was about three in the first year and over four in the second year. Again: that's the AVERAGE. That means that, if you ask an "expert" about how any character might fare in real-life competition, you can roughly expect them to be off by a whole letter grade IN EVERY SINGLE CASE.
So, uh, what's the point of all this, exactly? Personally, I think Myers is on to something:
Let's just take good ol' Street Fighter II as an example. The original SFII had all of eight playable characters. By today's standards, that's nothing - give the Marvel franchise another few versions and I guarantee we'll have eight fucking characters per team. But a random guess at the SFII tier list would only have a one-in-forty-thousand-three-hundred-and-twenty chance of being correct. A game with another twenty characters - say, vanilla SFV - makes this game practically impossible. And, okay, not even twitter trolls are dumb enough to make a tier list by just guessing at random. But how much do we KNOW about fighting game tiers, really? How big is the margin for error?
According to Maddy Myers, it's too big:
"Lists are based on supposed data, like the raw power of characters' attacks and their movement speed, but they're also based on more subjective stuff, like which characters have become popular among pro players at the highest levels. Pros also create their own personalized tier lists, which then influence the rankings on larger fan-sourced lists. In short, there are too many variables that influence tier lists, in their current form, for them to have any real statistical use."And this kinda sounds like it might be bullshit, right? I mean, tier lists LOOK objective. Once you stick a rating on something, it might as well be the Gold-Plated Truth. Plus, also, people can give all sorts of reasons about their rankings, ranging from highly scientific (attack power, "movement speed") to moderately reasonable (matchup charts) to purely subjective (popularity). And people have used them forever! So how wrong could they be, really?
Well, actually, pretty fucking wrong! To test the accuracy of our tier lists, I've been keeping track of the performance of the characters in SFV relative to both Shoryuken's and Eventhubs' tier lists. Wanna guess how useful those tier lists were in predicting championship winners? NOT VERY! Even including Nash in year one, who was a no-brainer pick for top five, the average error for Evo and Capcom Cup winners is over SEVEN. In other words, that character in the top spot? They could be the best - or they could barely be in the top ten. And that mid-tier character that you love even though everyone tells you to stop playing them? They have the potential to be godlike.
Even the overall margin isn't great - it was about three in the first year and over four in the second year. Again: that's the AVERAGE. That means that, if you ask an "expert" about how any character might fare in real-life competition, you can roughly expect them to be off by a whole letter grade IN EVERY SINGLE CASE.
So, uh, what's the point of all this, exactly? Personally, I think Myers is on to something:
"Tier lists exist so that fighting game players can offer one another endless unhelpful advice about which character they should have chosen. It's a great way to justify why you lost—your opponent's character was overpowered, right? Or maybe you chose an underpowered character. Or both!"Now, I'm not gonna say that we should ditch tier lists altogether. Players should absolutely refuse to GIVE A SHIT about them, sure, but banter is a key part of the FGC and we can't have banter if we're not allowed to bullshit each other at least a little. But I will say this: if you want to actually know what the fuck you're talking about when it comes to tiers, stop trying to be fancy. In reality - like, the real matches that real humans really play - there are only two tiers: VIABLE and NOT VIABLE. Every other tier is just a personal preference hiding behind a capital letter, and I dunno about you, but I had more than enough of living my life according to someone else's grades by the time I got out of school.
Comments
Post a Comment