Guest
home
beatmaps
rankings
community
store
help
forum
  1. Forums
  2. Beatmaps
  3. Mapping Discussion
  4. Ranking Criteria
  5. Finalized/Denied Amendments

[Proposal] Drastically large difficulty gaps between two difficulties

posted
Total Posts
14
Elcheer
423 posts
Joined May 2014
Topic Starter
Elcheer
Original
All game modes within a beatmapset must form a spread starting from the lowest difficulty level dictated by the song's drain time. For difficulties above the lowest required difficulty level, the spread cannot skip any difficulty levels and there cannot be any drastically large difficulty gaps between any two difficulties.

help/wiki/Ranking_Criteria#rules.1

This exists, but there are mapsets that are ranked where the difference between two difficulties can be over one star, an example being this mapset, the gap in question being 1.37 stars. In my opinion, from Extra (or maybe Insane) and beyond, any difference above at least 1.2 stars is a drastically large difference between difficulties. So here's my guideline (or rule) proposal:

yes
Should it be a guideline:
From Insane and beyond, avoid gaps between difficulties of over 1.2 stars.

Should it be a rule:
From Insane and beyond, gaps between difficulties should not exceed 1.2 stars.

I think I've worded it well enough


maybe this is just me being salty at seeing a set go from mid-5 stars to 7.0...
Shii
374 posts
Joined October 2016
Shii
SR (currently) isn't a reliable way of determining spread though.

Really don't think this'll help much :p
Serizawa Haruki
1,696 posts
Joined June 2015
Serizawa Haruki
While I agree that some maps have huge gaps between difficulties, I think Shii is right. The star rating is sometimes not representative of the actual difficulty, especially if you look at low diffs. The BPM also plays a major role for it. For example, if a Normal diff were 2.3* and a Hard diff 3.55*, it would already break your proposed rule although these are average numbers for the respective difficulties.

I think the problem stems more from the fact that mappers and BNs have a pretty lax interpretation of certain guidelines, leading to difficulty inflation over the years. For example, it is not uncommon for Insane diffs to have full-screen jumps and spaced 1/4 rhythms nowadays although the ranking criteria discourages it, just like some Hard diffs have really big spacing and lots of active 1/2 and 1/4 rhythm.

That being said, I don't think this issue can be solved with a RC proposal because it ultimately depends on whether BNs follow these guidelines or not. It's rather a quality assurance issue than a RC issue.
Elcheer
423 posts
Joined May 2014
Topic Starter
Elcheer

Shii wrote:

SR (currently) isn't a reliable way of determining spread though.

Serizawa Haruki wrote:

While I agree that some maps have huge gaps between difficulties, I think Shii is right. The star rating is sometimes not representative of the actual difficulty, especially if you look at low diffs.
my idea was geared more towards the higher diffs, since, at least from what i can remember, there's more difficulty between a 4 and 5 (or 5 and 6) star map than there is between a 2 and 3 star map

Serizawa Haruki wrote:

For example, if a Normal diff were 2.3* and a Hard diff 3.55*, it would already break your proposed rule although these are average numbers for the respective difficulties.
my proposals specify for at least insane and above


Serizawa Haruki wrote:

I think the problem stems more from the fact that mappers and BNs have a pretty lax interpretation of certain guidelines, leading to difficulty inflation over the years.

Serizawa Haruki wrote:

That being said, I don't think this issue can be solved with a RC proposal because it ultimately depends on whether BNs follow these guidelines or not. It's rather a quality assurance issue than a RC issue.
Fair points
Serizawa Haruki
1,696 posts
Joined June 2015
Serizawa Haruki
My bad, somehow I overlooked the "only for Insane and above" part. But my point still stands that star rating alone can't be used to determine how large a gap truly is. And I'd say that more than 1.2 stars is not necessarily too much, for example a 4.4* Insane and 5.7* Extra doesn't seem unreasonable, but again it depends on how they're actually mapped.

The only thing that can be done about this is defining what makes a gap too big, but I can't think of a proper way to do that right now.
Last edited by Serizawa Haruki , edited 1 time in total.
abraker
Global Moderator
8,327 posts
Joined July 2014
abraker
If you want to close difficulty gap, it has to be in terms of what is allowed and not in each difficulty type (Easy, Normal, Hard, etc). Star rating is too inaccurate and volatile for this sort of thing.
pishifat
osu! Paragon
3,812 posts
Joined August 2013
pishifat
is there a way to get this idea across without involving star rating? that's the only way i can see this proposal moving anywhere. atm people figure it out gaps between experts based on ~feeling~ which has worked okay so far, but i guess op's example is an exception to that, so having something clearer wouldnt be so bad
Dialect
1,535 posts
Joined May 2018
Dialect
^

tbh the only viable method is to do something similar to geometry dash's demon list system, where a handful of top players pick out which demon is harder than the other. obviously, this'll be really hard, because most top players aren't mappers, and unlike gd, which focuses on decoration (which is basically storyboarding in osu! terms), osu! focuses on representation of the music, so i feel like a situation would happen where a player would be biased based on map.

sotarks' map of flying out to the sky vs undeadcapulet's map of deal with the devil

despite flying out to the sky being technically more difficult (in sr), i think it's safe to say deal with the devil is actually more difficult, mainly due to the fast sliders at such high speeds. flying out to the sky IS hard, but i think players can handle jumps and streams easier than fast sliders. this is why maps like slow dancing in the dark are hard despite being almost a low 5*, due to their high sv sliders
chowch
43 posts
Joined February 2018
chowch

pishifat wrote:

is there a way to get this idea across without involving star rating?
It's unfortunate that SR isn't accurate in the job it was meant to do.

The most objective way would be through OD, but a cursory look at some modern sets tells me that there's no problem there. I think that determining a set's difficulty gaps will have to rely on other, less visibly objective metrics, in addition to the already established conventions on difficulty. The things I can think of are slider velocity, rhythm selection, and comfort of cursor movement -- basically, the aspects of a map not accurately represented by SR or PP.

These things would be nice to put in text (in addition to pre-established notions on difficulty!) as other ways to gauge difficulty across a set, as they are essentially the ~feeling~ of a map — just in a more explainable form.
Last edited by chowch , edited 2 times in total.
pishifat
osu! Paragon
3,812 posts
Joined August 2013
pishifat
based on replies i think this is a bit out of scope for the ranking criteria

Adam_S wrote:

The things I can think of are slider velocity, rhythm selection, and comfort of cursor movement -- basically, the aspects of a map not accurately represented by SR or PP.
defining these in a way that's suitable for the rc is pretty much impossible (and if they could be defined, they could be ported to sr calculations which solves our problems)

Li Syaoran wrote:

tbh the only viable method is to do something similar to geometry dash's demon list system, where a handful of top players pick out which demon is harder than the other.
this is similar to what we're doing now i think. bns determine if a gap is too wide for the spread based on their mapping experience. problems would come up if bns collectively stopped caring about wide diff gaps though...

im not really sure there's a better way to handle this :( star rating technically is an option, but removing star rating from the rc was a big goal of the diff-specific rc when those were created, so incorporating sr back as proposed seems like a weird step back

@BNs how do you think about expert->expert difficulty gaps?
Dialect
1,535 posts
Joined May 2018
Dialect
not a bn, but maybe if a new spread rule was in place, then maybe. but here's a bigger issue: how numbers may seem bigger than they are even tho they are the same gap.

for example:

5.1 > 6.8 vs 5.5 > 7.2

even tho they both have a gap of 1.7 (which is acceptable in some cases) i'd bet modders will point out the gap issue more with the 5.5 - 7.2 diffs instead of the 5.1 - 6.8 diffs. people won't really see the decimal and they'll only see 5 - 6 vs 5 - 7.

i think that we could find a better way to change it, but i dunno. i feel like it can be as is for now, but i wanna hear other's ops on this
Nao Tomori
osu! Alumni
3,107 posts
Joined December 2014
Nao Tomori
this is something that should be (and is) up to bn's discretion, as part of their skill set that is tested by apps. ideally bns should not qualify maps with large spread issues, and they should mod or pop maps with large spread issues. trying to stick everything into rc doesn't work well with how flexible the system has to be to function - this gives similar vibes to an amazing proposal once made about using industrial and natural sliders.
Elcheer
423 posts
Joined May 2014
Topic Starter
Elcheer
after reading everybody's replies

ok yea maybe this isn't something for rc, i agree with p much everything said
Noffy
Nomination Assessment Team
1,685 posts
Joined April 2012
Noffy
Archiving as a closed issue since there's basically agreement this isn't for rc to tackle in this format 🌟
Please sign in to reply.

New reply

0
1 / 14