GuamoKun
I Break Hearts, Not Combos
This specific community is kinda obsessed with numbers and assigning them to players.
"X is one of the top 20 players"
"Y is definitely bottom 5"
This is wrong.
We are assigning interval levels of measurements to different categories.
We can only assign ordinal levels of categories to things.
Ordinal levels of measurements assumes there is rank order (X>Y>Z) but does not assume there is an equal interval of difference between variables. When you assign interval levels of measurements it assumes that there is a difference of a known size between each category in addition to a rank order (Y>X by 5, so Z>Y by 5)
Apply this to Injustice:
Superman is better than Shazam, and we can quantify why that is so, but the "skill gap" between those is not the same as how Killer Frost is better than Shazam, so you can't really say specifically why they are better.
You can include things such as damage, frame data, mobility, use of interactables, and even hit/hutbox sizes but the differences between every character is so varied you can't come up with number tiers very easily.
This is why letter tiers work a lot better and why the "Top 10 conversation" phrase is more easily used compared to "top 10". Top 10 conversation doesn't necessarily mean "top10" but assigns a somewhat understood tier. I think of top 10 conversation as A tier.
tl;dr
If you mess up the level of measurement in your statistics, then you're gonna have a bad time mkay
shoutouts to irl applications of what I learned in school
"X is one of the top 20 players"
"Y is definitely bottom 5"
This is wrong.
We are assigning interval levels of measurements to different categories.
We can only assign ordinal levels of categories to things.
Ordinal levels of measurements assumes there is rank order (X>Y>Z) but does not assume there is an equal interval of difference between variables. When you assign interval levels of measurements it assumes that there is a difference of a known size between each category in addition to a rank order (Y>X by 5, so Z>Y by 5)
Apply this to Injustice:
Superman is better than Shazam, and we can quantify why that is so, but the "skill gap" between those is not the same as how Killer Frost is better than Shazam, so you can't really say specifically why they are better.
You can include things such as damage, frame data, mobility, use of interactables, and even hit/hutbox sizes but the differences between every character is so varied you can't come up with number tiers very easily.
This is why letter tiers work a lot better and why the "Top 10 conversation" phrase is more easily used compared to "top 10". Top 10 conversation doesn't necessarily mean "top10" but assigns a somewhat understood tier. I think of top 10 conversation as A tier.
tl;dr
If you mess up the level of measurement in your statistics, then you're gonna have a bad time mkay
shoutouts to irl applications of what I learned in school