forum
osu!mania accuracy
posted
Total Posts
5
Topic Starter
Ok, let's break this down. Error first:
Error takes all your hits, removes the most extremely early or late 5% of them (I'm not COMPLETELY sure about that, but I think that's how it works) and ignores those, then slpits them into hits that were early (- x ms) or late (+ x ms) and averages the early notes together and the late notes together. This gives you an idea of how early and late you are hitting. If the early number is higher, that means any notes you hit that are too early are more early than the notes you hit too late. Same thing goes for if the late number is higher.
If you look at the difference you can also see an "offset" in your timing. For example, if you were to hit every note 100% accurate to the music, but the notes were 10ms late, all your hits would be -10 ms. If you didn't hit everything perfectly, but you still played pretty accurate to the music instead of the notes, you'd see a difference between the early and late numbers there that is 10ms.
In your example, there's only a difference of about 2ms between your early and late numbers, which means you're not playing the notes on average too early or late, but you can't hit very accurately. So when you get 200's and lower, you have a pretty equal amount of hits that are early vs hits that are late. Basically. It's a bit more complicated, but that's the basic idea.
Now, unstable rate is actually a calculation that is used in statistics, but in statistics it's actually called Variance. The basic way to calculate variance is to find the average (so, add up every note you hit, then divide that by the number of notes) and then subtracting the average from every note you hit. After that, you square the result of that subtraction, and average all the squared differences.
Unstable rate tells you if your hits were closely grouped, or very spread out. The bigger the unstable rate, the bigger the differences between your hits on notes were. For example, if you hit on time most of the time, but every time you got a bad timing, you got a REALLY bad timing, unstable rate would be big, but if you hit off time a lot, but you were accurate in hitting off time, it would be small.
In the example above where I talked about what would happen if you hit perfectly on time to the music, but there was a 10ms offset between the music and the notes, the unstable rate would actually be 0, even though you weren't hitting on time according to the notes.
I'm open to corrections because it's been a long time since I dug into how that stuff works, but that's my understanding of hit error and unstable rate.
Error takes all your hits, removes the most extremely early or late 5% of them (I'm not COMPLETELY sure about that, but I think that's how it works) and ignores those, then slpits them into hits that were early (- x ms) or late (+ x ms) and averages the early notes together and the late notes together. This gives you an idea of how early and late you are hitting. If the early number is higher, that means any notes you hit that are too early are more early than the notes you hit too late. Same thing goes for if the late number is higher.
If you look at the difference you can also see an "offset" in your timing. For example, if you were to hit every note 100% accurate to the music, but the notes were 10ms late, all your hits would be -10 ms. If you didn't hit everything perfectly, but you still played pretty accurate to the music instead of the notes, you'd see a difference between the early and late numbers there that is 10ms.
In your example, there's only a difference of about 2ms between your early and late numbers, which means you're not playing the notes on average too early or late, but you can't hit very accurately. So when you get 200's and lower, you have a pretty equal amount of hits that are early vs hits that are late. Basically. It's a bit more complicated, but that's the basic idea.
Now, unstable rate is actually a calculation that is used in statistics, but in statistics it's actually called Variance. The basic way to calculate variance is to find the average (so, add up every note you hit, then divide that by the number of notes) and then subtracting the average from every note you hit. After that, you square the result of that subtraction, and average all the squared differences.
Unstable rate tells you if your hits were closely grouped, or very spread out. The bigger the unstable rate, the bigger the differences between your hits on notes were. For example, if you hit on time most of the time, but every time you got a bad timing, you got a REALLY bad timing, unstable rate would be big, but if you hit off time a lot, but you were accurate in hitting off time, it would be small.
In the example above where I talked about what would happen if you hit perfectly on time to the music, but there was a 10ms offset between the music and the notes, the unstable rate would actually be 0, even though you weren't hitting on time according to the notes.
I'm open to corrections because it's been a long time since I dug into how that stuff works, but that's my understanding of hit error and unstable rate.
Not super relevant to the conversation, but i've always had the question, if anyone can answer that would be amazing. What is the time range for each hit, for a 300g 300, 100, 50 obviously, miss you dont hit the note at all, but what is the total time for each, can someone explain it?
if you hover over here you can see the time ranges
Also next time please make a new thread instead of resurecting old ones. Especially if it's something not relevant to the conversation.
Also next time please make a new thread instead of resurecting old ones. Especially if it's something not relevant to the conversation.
