I have a common hit error of about -10ms~10ms (something around 140 unstable rate). I don't play HR very much so I was wondering; am I accurate enough for OD10? I always feared HR for the OD, constant 100s would lead to death by drain (maybe I'm more afraid of the drain) so I never really play it. I was just curious as to how accurate you actually have to be in order to hit proper 300s instead of getting constant 100s when dealing with OD10 or Hard Rock. Also is it true that Double Time decreases the hit window for accuracy?
forum
Accuracy when dealing with HR/OD10/DT
posted
Total Posts
58
X means no-mod
Topic Starter
Oh wow, so I should pretty much be able to clear anything (within my speed range)
Well the 18ms is plus or minus 9ms from the beat. So you're slightly not accurate enough to consistently hit 300s on od10.
That thing doesn't show your max error made on a map. I just checked it on a map that has OD9.8 (20ms)
I got 3 100s here, but hit error info showed +6ms ~ -6ms with unstable rate 76
So, it's kinda not useful or I am misunderstanding something
I got 3 100s here, but hit error info showed +6ms ~ -6ms with unstable rate 76
So, it's kinda not useful or I am misunderstanding something
Actually OD10 is plus or minus 18ms (so the time window is 36ms).TRIYP wrote:
Well the 18ms is plus or minus 9ms from the beat. So you're slightly not accurate enough to consistently hit 300s on od10.
With Unstable Rate of 140 you should expect around 86.5% circle accuracy in OD10 though (song length doesn't affect accuracy much in low stability).
Unstable Rate 76 is enough to get about 99.7%acc in OD9.8 (in a short song). 99.27%acc in a long song.Kert wrote:
That thing doesn't show your max error made on a map. I just checked it on a map that has OD9.8 (20ms)
I got 3 100s here, but hit error info showed +6ms ~ -6ms with unstable rate 76
So, it's kinda not useful or I am misunderstanding something
How do you count that % ?
Assuming hit errors follow a Normal Distribution, then you math from there.Kert wrote:
How do you count that % ?
Does it count slider heads? (I have 120 unstable rate on OD10 900+ combo map with 99+ accuracy)
Because I don't have any other idea of why is this happening
Because I don't have any other idea of why is this happening
I think for something like accuracy in osu!, it won't really follow a normal distribution because making a mistake leads to an outlier. Even assuming perfect reading, humans will still screw up, whether that's due to nervousness, strain on fingers, or whatever else. Perhaps outliers are already accounted for, but I would guess that it will follow a roughly normal distribution centered closer to perfect timing with a fair amount of points that are severely mistimed. This is mostly coming from my experience, because my unstable rates on HDHR are usually around 85-90 (when I'm really comfortable with the map), and most of my 100s come from when I can feel myself making a mistake.
The calculations only considers accuracy for circles.Kert wrote:
Does it count slider heads? (I have 120 unstable rate on OD10 900+ combo map with 99+ accuracy)
Because I don't have any other idea of why is this happening
Sliders can screw the approximation. Because maps with a lot of sliders give you high accuracy even with high unstable rate (they are easier to time), and you can play the sliders with a more relaxed timing (increasing the Unstable Rate).
Unstable rate can certainly help to determine how accurate you are, but it's not an absolute number.Kert wrote:
Does it count slider heads? (I have 120 unstable rate on OD10 900+ combo map with 99+ accuracy)
Because I don't have any other idea of why is this happening
Honestly I have no idea how it works, but really accurate players like Cookiezi, thelewa and mugio3 tend to always be under 100 unstable rate.
OP: I think unstable rate 140 should give at least 93% though, as opposed to FullTablet's calculation of 86.5%. This is based off my own experience as a HR player but I realize it can differ from person to person.
I'm curious, have you tested in maps without sliders/spinners?koromo wrote:
OP: I think unstable rate 140 should give at least 93% though, as opposed to FullTablet's calculation of 86.5%. This is based off my own experience as a HR player but I realize it can differ from person to person.
According to the calculations 48.15% of the objects in the map have to be sliders/spinners to achieve an expected accuracy like that.
I can't really even think of any map without sliders tbh, so pretty sure I haven't.Full Tablet wrote:
I'm curious, have you tested in maps without sliders/spinners?
According to the calculations 48.15% of the objects in the map have to be sliders/spinners to achieve an expected accuracy like that.
edit:
I'm pretty sure your unstable rate would be lower when playing AR10 since you have a reduced window of visibility of the circle so the OP's unstable rate might be in the range of 110-120 (just estimating) but there are also factors like playing the map based more on reaction time then with the music if AR10 is fast for you which might even out the unstable rate.Full Tablet wrote:
I'm curious, have you tested in maps without sliders/spinners?koromo wrote:
OP: I think unstable rate 140 should give at least 93% though, as opposed to FullTablet's calculation of 86.5%. This is based off my own experience as a HR player but I realize it can differ from person to person.
According to the calculations 48.15% of the objects in the map have to be sliders/spinners to achieve an expected accuracy like that.
For that specific map and unstable rate, according to the model you would have got 91.23%acc. (But that without considering if your sliders are less precise than your circles).koromo wrote:
I can't really even think of any map without sliders tbh, so pretty sure I haven't.Full Tablet wrote:
I'm curious, have you tested in maps without sliders/spinners?
According to the calculations 48.15% of the objects in the map have to be sliders/spinners to achieve an expected accuracy like that.
edit:
If Unstable Rate was separated between Circle Unstable rate and Slider Unstable rate it would be easier to check if the model is right (for example, checking if normal distribution is good or if it is better to use another kind of distribution).
If you want the best advice in this situation. Just play it and find out. I usually get a lot higher unstable rate when playing NOMOD but have gotten as low at 65 unstable rate playing Hard Rock. Don't try to base your skill off of a silly number in game just try it and find out.
Listening to all these people tell you what unstable rate you should play HR at or whether or not you can do it or not yet and what you can expect is just plain dumb.
Listening to all these people tell you what unstable rate you should play HR at or whether or not you can do it or not yet and what you can expect is just plain dumb.
^^^^^^^^JAKACHAN wrote:
If you want the best advice in this situation. Just play it and find out. I usually get a lot higher unstable rate when playing NOMOD but have gotten as low at 65 unstable rate playing Hard Rock. Don't try to base your skill off of a silly number in game just try it and find out.
Listening to all these people tell you what unstable rate you should play HR at or whether or not you can do it or not yet and what you can expect is just plain dumb.
everything that JAKA said is true, seconded by yours truly
unstable rate and hit error don't say jackshit about the accuracy you will get. you can have a super good unstable rate below 100 or even 80 and still get 100s on OD 8. vice versa, you can get really good accuracies upt to almost SS or SS on OD 10 with unstable rates just at 100 or slightly above it, if you hit many circles just at the border to a 100.
also you could theoretically get a SS on OD 9.8 but only 33.33% on OD 10 if you hit all the notes in the thin timing window inbetween those 2. and all that doesn't even include things like slidermisses or actual misses (fun fact: clearing a map with no fail and just hitting one circle will result in unstable rate 0,00).
in the end, you should not worry about unstable rate and hit error and such things all too much. what matters in the end is the actual accuracy, and if you get a SS, it will be a SS no matter what unstable rate you have or what others say which unstable rate results in which accuracy.
also you could theoretically get a SS on OD 9.8 but only 33.33% on OD 10 if you hit all the notes in the thin timing window inbetween those 2. and all that doesn't even include things like slidermisses or actual misses (fun fact: clearing a map with no fail and just hitting one circle will result in unstable rate 0,00).
in the end, you should not worry about unstable rate and hit error and such things all too much. what matters in the end is the actual accuracy, and if you get a SS, it will be a SS no matter what unstable rate you have or what others say which unstable rate results in which accuracy.
Theoretically, yes. In practice, no.Tanzklaue wrote:
unstable rate and hit error don't say jackshit about the accuracy you will get. you can have a super good unstable rate below 100 or even 80 and still get 100s on OD 8. vice versa, you can get really good accuracies upt to almost SS or SS on OD 10 with unstable rates just at 100 or slightly above it, if you hit many circles just at the border to a 100.
also you could theoretically get a SS on OD 9.8 but only 33.33% on OD 10 if you hit all the notes in the thin timing window inbetween those 2. and all that doesn't even include things like slidermisses or actual misses (fun fact: clearing a map with no fail and just hitting one circle will result in unstable rate 0,00).
in the end, you should not worry about unstable rate and hit error and such things all too much. what matters in the end is the actual accuracy, and if you get a SS, it will be a SS no matter what unstable rate you have or what others say which unstable rate results in which accuracy.
You do have a point though. Stop caring about them statistics and start playing.
I use unstable rate just to scope which way I should adjust a local offset that comfortable to me for each song
other than that is just being cool when it's under 100 at least
other than that is just being cool when it's under 100 at least
i don't even look at unstable rate. when i see that my accuracy sucks i just rage and leave nonetheless
I have to look at it every time for my egowinber1 wrote:
i don't even look at unstable rate. when i see that my accuracy sucks i just rage and leave nonetheless
Outliers are accounted for in normal distributions as long as they occur both above and below the mean at reasonable (AKA low) rates, which is how it works for most people in osu!.pooptartsonas wrote:
I think for something like accuracy in osu!, it won't really follow a normal distribution because making a mistake leads to an outlier. Even assuming perfect reading, humans will still screw up, whether that's due to nervousness, strain on fingers, or whatever else. Perhaps outliers are already accounted for, but I would guess that it will follow a roughly normal distribution centered closer to perfect timing with a fair amount of points that are severely mistimed. This is mostly coming from my experience, because my unstable rates on HDHR are usually around 85-90 (when I'm really comfortable with the map), and most of my 100s come from when I can feel myself making a mistake.
Topic Starter
I didn't care for the unstable rate and numbers anyways, I was just curious as to how accurate you have to really be in order to hit stuff. I played and hard rock isn't nearly as hard as I thought it was. Playing a lot of DT helps and also having a correct UO helps in massive amounts. The test build has a great hit error rate bar that helped me adjust my UO to be perfect. I just hope it gets implemented soon.
Thanks guys for the information. This is all very helpful.
Thanks guys for the information. This is all very helpful.
Need to note that HalfTime and DoubleTime/Nightcore mods may decrease and increase your Unstable Rate along with improves and worsen your Hit Error respectively, because those speed-adjusting mods aren't considered toward the Hit Error and Unstable Rate.
Hit Error is showing average of early hit timings and late hit timings separately. The minus value averages your early hits, while the plus value averages your late hits. This doesn't show your peak hit error, and since it averages your timing instead, the more circles and sliders are involved, the approximation is less accurate. Take an extreme case example: there's a map with 2,000 circles, and you hit -101ms on just one circle while the remaining 1,999 circles are magically hit at -1ms all the time, and there's no late timings at all (HAAAAX! xD). The result of this Hit Error is -1.05ms ~ +0.00ms, yet you can still see that the 300/100/50 result is 1999/1/0.
Unstable Rate also averages your timing as well, but rather than based on the 0ms point, it's based on your overall hit window. Taking the same extreme case example again, you get -100ms all the time on the same map above. The result of 300/100/50 is 0/2000/0 with Unstable Rate of 0.00!
Of course those extreme cases are very far away from being likely to happen anyway, but this also proves that Hit Error and Unstable Rate should be taken as statistical purposes and not as an accurate measurement of skill.
Hit Error is showing average of early hit timings and late hit timings separately. The minus value averages your early hits, while the plus value averages your late hits. This doesn't show your peak hit error, and since it averages your timing instead, the more circles and sliders are involved, the approximation is less accurate. Take an extreme case example: there's a map with 2,000 circles, and you hit -101ms on just one circle while the remaining 1,999 circles are magically hit at -1ms all the time, and there's no late timings at all (HAAAAX! xD). The result of this Hit Error is -1.05ms ~ +0.00ms, yet you can still see that the 300/100/50 result is 1999/1/0.
Unstable Rate also averages your timing as well, but rather than based on the 0ms point, it's based on your overall hit window. Taking the same extreme case example again, you get -100ms all the time on the same map above. The result of 300/100/50 is 0/2000/0 with Unstable Rate of 0.00!
Of course those extreme cases are very far away from being likely to happen anyway, but this also proves that Hit Error and Unstable Rate should be taken as statistical purposes and not as an accurate measurement of skill.
JAKACHAN wrote:
Just play it and find out.
Yeah but in osu!, pretty much all the outliers will lead to 100s on a high OD because outliers will often be the result of nervousness where the player completely mistimes a note in the heat of the moment. I just don't think unstable rate can properly predict accuracy because someone who has perfect accuracy under normal conditions could still mistime notes due to stress.Ekaru wrote:
Outliers are accounted for in normal distributions as long as they occur both above and below the mean at reasonable (AKA low) rates, which is how it works for most people in osu!.pooptartsonas wrote:
I think for something like accuracy in osu!, it won't really follow a normal distribution because making a mistake leads to an outlier. Even assuming perfect reading, humans will still screw up, whether that's due to nervousness, strain on fingers, or whatever else. Perhaps outliers are already accounted for, but I would guess that it will follow a roughly normal distribution centered closer to perfect timing with a fair amount of points that are severely mistimed. This is mostly coming from my experience, because my unstable rates on HDHR are usually around 85-90 (when I'm really comfortable with the map), and most of my 100s come from when I can feel myself making a mistake.
Mistimed notes due to stress increase the Unstable Rate too.pooptartsonas wrote:
Yeah but in osu!, pretty much all the outliers will lead to 100s on a high OD because outliers will often be the result of nervousness where the player completely mistimes a note in the heat of the moment. I just don't think unstable rate can properly predict accuracy because someone who has perfect accuracy under normal conditions could still mistime notes due to stress.Ekaru wrote:
Outliers are accounted for in normal distributions as long as they occur both above and below the mean at reasonable (AKA low) rates, which is how it works for most people in osu!.
Unstable rate only tell how unstable you are. You can, for example, constantly drift from -17 to +17, thus SSing OD10 but also having ridiculously high unstable rate
You can't say which accuracy you should get with xxx.xx unstable rate
You can't say which accuracy you should get with xxx.xx unstable rate
Unstable Rate can help to tell what accuracy you should expect, with a certain probability.GoldenWolf wrote:
Unstable rate only tell how unstable you are. You can, for example, constantly drift from -17 to +17, thus SSing OD10 but also having ridiculously high unstable rate
You can't say which accuracy you should get with xxx.xx unstable rate
It's analogous to something like this: You have a die with 10 sides (each side marked A,B,C,etc...), so with 1000 throws, the expected number of times you get an A is 100 times (but of course there will be some variation, and the probability of getting EXACTLY 100 times an A is very low).
There are some things you can know for sure, though. For example, if your average hit error is exactly 0 (this isn't shown in the results screen), you pressed each object (managed to at least click each object while on the MISS time frame), and your Unstable Rate is 100, then the worst possible accuracy you can get in OD11 (in the worst case scenario) is 30.55%; and if the Unstable Rate is 60 then the worst possible accuracy you can get is 75% (knowing this doesn't actually help in anything though).
Similarly, if your average hit error is 0, and your unstable rate is below 12.0605, you will always get a SS in any map under 100 circles/sliders (if you also complete the slider paths correctly, and the spinners). With unstable rate under 1.69723, you will always get a SS in maps with up to 5000 circles/sliders (if you also complete the sliders paths and spinners correctly).
I've SS'd with 140 unstable rate I've SS'd with 65 unstable rate.Full Tablet wrote:
GoldenWolf wrote:
Unstable Rate can help to tell what accuracy you should expect, with a certain probability.
Once again do not judge what you should have through numbers. Just play and find out yourself.
Trying to put numbers on everything starts to complicates things and just adds more stress to something that should be a fun game to begin with people.
Well, some people like numbers.JAKACHAN wrote:
Trying to put numbers on everything starts to complicates things and just adds more stress to something that should be a fun game to begin with people.
You quoted me wrong JAKACHAN ;_;
so basically, adding hr would be the same between 8~10? it doesn't seem at all. ex in Gensou no satellite (od9) is way harder keeping a good acc rate rather than airman (od8).Kert wrote:
X means no-mod
HR increases OD and AR by 40% and it caps at 10. So both OD8 and OD9 become OD10 after adding HR, OD7 becomes OD9.8.Genshin wrote:
so basically, adding hr would be the same between 8~10? it doesn't seem at all. ex in Gensou no satellite (od9) is way harder keeping a good acc rate rather than airman (od8).
let's see how many people freak about the necrobump instead of realizing what I'm trying to say