Ingame Song Information

[Patashu]

Re: R1 - Actual Song Difficulties V2

that's going to be a problem no matter how many difficulty rankings there are. there are high 77s and low 77s just as there are high fmos and low fmos.

[Vanilla Mnm]

Re: R1 - Actual Song Difficulties V2

There is no way Silence is an FMO...
Southern Cross MAYBE, but probably not.
K8107 is an FMO, comparing it to other low FMOs.

[bmah]

Re: R1 - Actual Song Difficulties V2

Silence is persistent which is why it's mainly an FMO. Patashu basically said it.

[igotrhythm]

Re: R1 - Actual Song Difficulties V2

Silence has been moved down to VC in game last I checked. It's not blazing fast, but stamina and consistency is what's needed to get the FC there.

Southern Cross's jump bursts get me every time, but the streams aren't that bad.

K8107...pfft, I managed to PFC that song to 528 once. The ending is fairly tough to hit, but it's easily readable.

[qqwref]

Re: R1 - Actual Song Difficulties V2

I find silence hard to read, but not really enough to make it FMO anymore considering all the technical bursty tough songs that have been put into the VC category.

[Niala]

Re: R1 - Actual Song Difficulties V2

Silence>75
GoSW>77? Or am I the only person who finds this file kind of tricky? Mostly just the trills on 2-3, that section is really easy to mess up on. Plus, the chorus of the file plays kind of weird, IMO. Anybody with me on this?

[Superfreak04]

Re: R1 - Actual Song Difficulties V2

Silence>75
GoSW>77? Or am I the only person who finds this file kind of tricky? Mostly just the trills on 2-3, that section is really easy to mess up on. Plus, the chorus of the file plays kind of weird, IMO. Anybody with me on this?

I agree, it took me forever to AAA that file because of step patterns like that.

[Izzy]

Re: R1 - Actual Song Difficulties V2

Is there some standard evaluation of a songs difficulty that we go by? If not then it seems like we should make one so we can judge song difficult by the same factors every single time for consistancy.

Note density
Pattern difficulty
Length
Average bpm

etc...

[XCV]

Re: R1 - Actual Song Difficulties V2

I agree with Izzy, although the criteria he gives isn't completely objective without arbitrary numerical equation (pattern difficulty), and that really just makes my head spin thinking about it because there's way too much to accurately equate.

So basically, if this doesn't fall through we're after something like this:

hm(taps/second, pattern difficulty, length in seconds, bpm)

where hm() is the harmonic mean function, defined as the reciprocal of the arithmetic mean of the sum of the reciprocals of the values, which works out to:

4/(l/s + 1/d + l + b) = 4dt/(ds + t + dls + bds)

l = length in seconds
s = steps (notes)
d = pattern difficulty
b = average BPM

[Check my math, please.]

If there weren't four variables in this, it would be far less cumbersome to work with. As it is, algebraic simplification for d is beyond my immediate powers of thought and it's not worth spending time on, as it's a guess that I tossed out and very inaccurate.

I've spent a good 30 minutes typing this and/or doing math, so I'll leave off here. If anyone with better math education than me can get a better estimate of this function, then please do so.

[smartdude1212]

Re: R1 - Actual Song Difficulties V2

Or we just compare songs to each other.

Choprite is much easier than Southern Cross (a few fast bits versus jumpstream & some 32nds) yet Choprite is FMO and Southern Cross is not.

Dot Matrix Hero is easier than some of the 9s out there but it's still VC.

The list goes on and on.

[bmah]

Re: R1 - Actual Song Difficulties V2

the point was to add some objectivity into determining song difficulties

not all songs can be qualitatively evaluated and agreed upon unanimously; we've been comparing songs like that ever since, but there might be a way to make decisions more objectively by determining various parameters of a song

[TC_Halogen]

Re: R1 - Actual Song Difficulties V2

I agree with Izzy, although the criteria he gives isn't completely objective without arbitrary numerical equation (pattern difficulty), and that really just makes my head spin thinking about it because there's way too much to accurately equate.

So basically, if this doesn't fall through we're after something like this:

hm(taps/second, pattern difficulty, length in seconds, bpm)

where hm() is the harmonic mean function, defined as the reciprocal of the arithmetic mean of the sum of the reciprocals of the values, which works out to:

4/(l/s + 1/d + l + b) = 4dt/(ds + t + dls + bds)

l = length in seconds
s = steps (notes)
d = pattern difficulty
b = average BPM

[Check my math, please.]

If there weren't four variables in this, it would be far less cumbersome to work with. As it is, algebraic simplification for d is beyond my immediate powers of thought and it's not worth spending time on, as it's a guess that I tossed out and very inaccurate.

I've spent a good 30 minutes typing this and/or doing math, so I'll leave off here. If anyone with better math education than me can get a better estimate of this function, then please do so.

There's only one issue with this. Pattern difficulty is a highly subjective issue - some people are better at one thing and others are better at something else. No matter what equation you come up with, there's really no way to account for this. Pattern density isn't an effective equation either because some songs will spike to extremely high densities for short periods and it just makes mathematical equations that much harder to derive. I do like the setup you have as a foundation, though.

[bmah]

Re: R1 - Actual Song Difficulties V2

It is impossible to be 100% objective in the determination process, but there are a few things that can be accounted for objectively, such as song length. To what degree the length affects the song is another matter though. Overall, if we try to account for everything (as I've attempted to quantify before), it becomes overly complicated.

[One Winged Angel]

Re: R1 - Actual Song Difficulties V2

haha that's the first thing he mentions in his post, there's really no way to get around that subjectivity

a file that's one minute of 300 bpm random 16th stream is obviously going to be harder than a file of the same length that's 300 bpm 16th left to right rolls, but there's obviously no surefire way of determining the number that would separate the two in difficulty

bmah you ninja

[OneHandNow]

Re: R1 - Actual Song Difficulties V2

i would find the 16th left to right rolls harder actually..

subjectivity