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

New diff info - CPM, SSPM, ASPM, JPM, JSPM

posted
Total Posts
11
This is a feature request. Feature requests can be voted up by supporters.
Current Priority: +2
Blue Stig
378 posts
Joined November 2011
Topic Starter
Blue Stig
The star system has been a bit vague so I'm gonna throw out some stuff and see what sticks.

CPM - circles per minute
Basically this is the average amount of circles and slider starts you get per minute in a diff.
In the calculation you would not include breaks to keep an accurate reading.
This would give a general idea of how fast you have to click.

SSPM - slider speed per minute
This is the range of slider speeds you get in a diff.
It would be calculated by multiplying a base bpm of 60 to the actual speed of the sliders.
For example a 1.25 slider speed at 180 bpm would have a sspm of 3.75

ASPM - actual speed per minute
This would be the slider speed of streams.
It would be the same as sspm multiplied by the spacing modifier.

JPM - jumps per minute
Exactly what it says on the tin.
Again breaks will be excluded in calculation.

JSPM - average jump speed per minute
This will calculate the average speed of jumps in terms of sspm.
Kitsunemimi
2,188 posts
Joined April 2009
Kitsunemimi
I think this is a really good idea (reminds me of the stepmania difficulty gauges), although the jumps are going to be a bit tricky to do, because you don't always know what's considered a "jump"

There are also other factors in maps that determine the difficulty, but cannot be precisely measured, such as snappiness, anti-jumps (aka bad mapping), unusual rhythm, etc.

Regardless, I'm still supporting this because it's a massive step up from the completely broken star rating system we have now, and I've had this sort of thing in my mind for a while now.
jeffreydav
11 posts
Joined January 2012
jeffreydav
1) How do you propose calculating jumps per minute? How do you define a jump in an objective way, so that they can be tested for algorithmically?

2) As for average jump speed, you should also take into account circle size. Fitts' Law states that the amount of time you need to click on a target depends on the ratio between the distance to the target and the size of the target. If you double the distance of the jump but also double the circle radius then you don't change the difficulty of the jump (in theory, anyway) - of course, this is the same as just halving the mouse sensitivity. So you should measure jump speed in units of "circle radii per second". How you answer 1) also affects this stat, since if you included a bunch of "jumps" that aren't really jumps, then this stat will be too low.
deadbeat
osu! Alumni
9,218 posts
Joined June 2009
deadbeat
there is this -> t/98169
not much has happened there in a while
Stefan
Global Moderator
10,407 posts
Joined January 2011
Stefan
I find t/98169 also much better and more attractive. Not that I would not support this Thread here but if I need to choose I perfer RBRat's idea.
Blue Stig
378 posts
Joined November 2011
Topic Starter
Blue Stig

deadbeat wrote:

there is this -> t/98169
not much has happened there in a while
Ah.

Well although that thread is much much better this thread puts more of a focus on new information to better explain how the map plays.
Let's see where this thread goes before merging or locking.

As far as presentation I would put it somewhere on the map's information page first and then focus on t/98169.

I would calculate jumps as any note that exceeds the distance snap and anti-jumps (halts?) as notes that are less than the distance snap.
There also should be something for stacks as well.
Tshemmp
1,405 posts
Joined June 2011
Tshemmp
You could use average spacing instead of jumps, i.e. average distance you have to move your cursor per minute.
Blue Stig
378 posts
Joined November 2011
Topic Starter
Blue Stig

Tshemmp wrote:

You could use average spacing instead of jumps, i.e. average distance you have to move your cursor per minute.
That would be ok but there is no compensation for rapidly changing maps like this.
You would need a way to tell how dynamic a map is.

Also since jumps might be within the distance snap you could also use the timing between notes in part of the calculation.
MillhioreF
Pro Tester
5,004 posts
Joined July 2011
MillhioreF
Why not use the AiMod SDK to calculate these, then see how useful they end up being?
Blue Stig
378 posts
Joined November 2011
Topic Starter
Blue Stig

MillhioreF wrote:

Why not use the AiMod SDK to calculate these, then see how useful they end up being?
Well, I'm a bit of newb programmer who just sort of knows OOP loosely in a Java kind of way, but I'll take a look.
Who knows maybe I'll learn something.
Rei Hakurei
1,983 posts
Joined February 2011
Rei Hakurei
number of objects, speed rate, spacing rate...

sounds like what i'm calculating lol..

SPOILER
Shortest Spacing, Average Spacing, Largest Spacing, Slowest Interval, Average Interval, Shortest Interval, Maximum SV, Initial SV.... (having it's own weight for each mode)... and yeah.. it's how i'm calculating diff-stars for now)
Please sign in to reply.

New reply

0
1 / 11