Guest
home
beatmaps
rankings
community
store
help
forum
  1. Forums
  2. osu!
  3. Development

Current accuracy

posted
Total Posts
10
CeilingPanda
148 posts
Joined June 2012
Topic Starter
CeilingPanda
Hello,
As I was playing some easy maps with a friend, as well as playing some hard/insane maps alone, I noticed my accuracy going first up then down, even if I did not really improve much.
As the current accuracy is quite "misleading" or just off, would it be hard to implement:
An accuracy for each difficulty/star level or something similar?
An accuracy for the latest, lets say, 5000 maps you've played?

What does this do?
It gives you a more(far from perfect though) idea of how good accuracy you have.
It doesn't encourage easier maps for keeping a high accuracy.

Remove nofail from ranked and put it onto the unranked mods(or a practice mod) which lets you try the map without losing accuracy as you do now when you play a hard map with no fail to see if you can make it.

While this might seem unnecessary I know some people cares, including me, so please hold off the "who cares" messages :)
Comments and thoughts are always welcome!
Aqo
2,710 posts
Joined April 2010
Aqo
I recommend that you just stop looking at profile stats. The accuracy stat on profile will never tell anything since different maps have different OD values. Until a change to the entire pool of osu ranked maps that forces same OD on all maps as well as adding accuracy for sliders happens, there will be no way to make this stat represent anything remotely consistent.

So basically just play more, and don't worry about what your account says.

If you want to know what OD is, try opening some maps you've played in Edit and changing their Overall Difficulty value in Song Setup > Difficulty and then seeing the different results that you get on different ODs. This will show you why accuracy stat on profile says nothing :b
CeilingPanda
148 posts
Joined June 2012
Topic Starter
CeilingPanda

Aqo wrote:

I recommend that you just stop looking at profile stats. The accuracy stat on profile will never tell anything since different maps have different OD values. Until a change to the entire pool of osu ranked maps that forces same OD on all maps as well as adding accuracy for sliders happens, there will be no way to make this stat represent anything remotely consistent.

So basically just play more, and don't worry about what your account says.

If you want to know what OD is, try opening some maps you've played in Edit and changing their Overall Difficulty value in Song Setup > Difficulty and then seeing the different results that you get on different ODs. This will show you why accuracy stat on profile says nothing :b
Still, it would show a better accuracy than it does now as the same accuracy is based over all difficulties/OD's. Farming easy maps to show off a good score on hard is possible, which would be prevented this way.
Sure it would be possible, but it would require a bit more accuracy than now. If it does not say anything they could remove it all together as its purpose is not filled.
Forcing the same OD's for the same difficulties(ish) could be a good fix? Still sliders remains a problem but well, their influence is not the biggest problem with the accuracy if you ask me :).
D33d
1,697 posts
Joined April 2011
D33d
Solution: Do your best and you'll have good accuracy as a matter of course.
CeilingPanda
148 posts
Joined June 2012
Topic Starter
CeilingPanda

D33d wrote:

Solution: Do your best and you'll have good accuracy as a matter of course.
While I do understand your point, do you even understand mine? The problem is not having low accuracy, the problem is that it doesn't reflect your accuracy.
If you for example, implement 4 different accuracy stats, one for easy maps, one for medium, one for hard and one for expert it will reflect it much better than it does now.

If I do my best on easy maps I would probably have between 98%-100% without any mods. On expert on a hard beatmap I would probably end up somewhere between 50-80% depending on how hard it is.(And therefore fail), but I hope you understood my point.
Having an accuracy on 75% is not my accuracy on easy songs as it is almost 100% and 75% is too high for the hard beatmaps.
But by doing this I would have an accuracy on 100% on easy and 65% on expert which would be much more accurate!
Aqo
2,710 posts
Joined April 2010
Aqo

Madlollipop wrote:

If you for example, implement 4 different accuracy stats, one for easy maps, one for medium, one for hard and one for expert it will reflect it much better than it does now.
It might be a tiny bit better, but once you play this game for long enough and understand how variable the different maps are you'll see why this would still be zero progress towards anything that resembles accuracy.

In the end, this stat is useless. Just ignore its existence.
The most you can make out of it is "the higher a profile's accuracy is, the more that person plays easy maps".
CeilingPanda
148 posts
Joined June 2012
Topic Starter
CeilingPanda

Aqo wrote:

Madlollipop wrote:

If you for example, implement 4 different accuracy stats, one for easy maps, one for medium, one for hard and one for expert it will reflect it much better than it does now.
It might be a tiny bit better, but once you play this game for long enough and understand how variable the different maps are you'll see why this would still be zero progress towards anything that resembles accuracy.

In the end, this stat is useless. Just ignore its existence.
The most you can make out of it is "the higher a profile's accuracy is, the more that person plays easy maps".
While I do understand what you're saying, and I completely agree! That's why I made this post, to make it a not useless stat :). I do not know how to make it useful but I hope that thoughts could maybe help that part to be solved.
D33d
1,697 posts
Joined April 2011
D33d
Yes, I understand your point. As Aqo said, maps are variable. All that matters is that you improve over time. This will eventually reflect in your stats.
Full Tablet
2,542 posts
Joined September 2011
Full Tablet
A solution would be basing accuracy on average absolute timing offset instead of the current system.
For example, if a person has "10ms" accuracy, it means that in average that person presses circles/sliders 10ms early or late.

A problem with this is that some people (at least me) tend to intentionally click sliders slightly earlier to make keeping the combo easier. (So in that case there can be separate accuracy stats for circles and sliders).
boat
2,369 posts
Joined February 2010
boat
Just like your rank isn't actually a measurement of "skill". Compare map specific scores and accuracies instead.

Madlollipop wrote:

Remove nofail from ranked and put it onto the unranked mods(or a practice mod) which lets you try the map without losing accuracy as you do now when you play a hard map with no fail to see if you can make it.
Unranked play for practice got removed about a year ago.
Please sign in to reply.

New reply

0
1 / 10