This proposal aims to reduce the burden for ranking difficult (like 7*+) maps, by specifying number of required spread diffs as an alternative to the lowest difficult requirement.

(Edit: please also refer to uberfazz's post or my more recent post for alternative proposals to fix it, which relaxes expert level requirement while leaving lower diffs as is)
community/forums/topics/1880953?n=8
community/forums/topics/1880953?n=20

Currently, ranking high level difficulties requires a lot of spread diffs, making it a lot difficult to make and rank mapsets with them. I see a lot of mappers with capability to rank maps just graveyard difficult maps without attempting to rank them.

There are largely unsatisfied demands coming from improving player skill levels. For example, for nomod plays, merely 10k-ish players have mid-high 6* or even 7* full ver maps in best performance scores. Such diff already has very limited number of maps and it's difficult to find new maps that you'd specifically enjoy.

Mania already has a good solution to reduce burden for ranking difficult maps, specifying the number of required spread diffs as an alternative to having a specified lowest diff level. Their rule is like the following:

...
...between 3:15 and 4:00, each key mode and playstyle must either include a difficulty at Insane or lower, or provide a proper spread containing at least 2 difficulties.
...

It relaxes spread requirement for sets with high diff levels while also removing huge discontinuity of required spread at marathon length.

Similar approach can relax the burden for osu standard.
The current spread rule is the following:

If the drain time of each difficulty is...
...lower than 3:30, the lowest difficulty cannot be harder than a Normal.
...between 3:30 and 4:15, the lowest difficulty cannot be harder than a Hard.
...between 4:15 and 5:00, the lowest difficulty cannot be harder than an Insane.

The spread rule would be changed as follows, with this change:

If the drain time of each difficulty is...
...lower than 3:30, the lowest difficulty cannot be harder than Normal, or the set should provide a proper spread containing at least 4 difficulties.
...between 3:30 and 4:15, the lowest difficulty cannot be harder than a Hard, or the set should provide a proper spread containing at least 3 difficulties.
...between 4:15 and 5:00, the lowest difficulty cannot be harder than an Insane, or the set should provide a proper spread containing at least 2 difficulties.
...

One drawback of this proposal would be the reduced number of relatively lower diffs. However, the impact on lower diff players on the song variety would be minimal. The affected ones are sets that cannot have only one extra (top diff) for spread, typically mid 6*+, which is small portion of all ranked beatmaps. A lot of full ver mappers are also mapping lower diffs voluntarily despite RC not requiring it right now.

This should just be in all modes honestly.

I disagree with the proposal.

1. A material percentage of plays on higher difficulty maps come from the low difficulty spread.
2. The total amount of plays on any given map is already quite low nowadays.
3. People generally do not voluntarily make non required low difficulties for their difficult spreads.

Given these three points, I don't think it's a good idea to further reduce the amount of intermediate level maps coming into the game.

YyottaCat wrote:

This should just be in all modes honestly.

I'm not sure how much demand exists in other modes so I limited it to std, if it is needed for other modes separate proposal could handle it

Nao Tomori wrote:

I disagree with the proposal.

1. A material percentage of plays on higher difficulty maps come from the low difficulty spread.
2. The total amount of plays on any given map is already quite low nowadays.
3. People generally do not voluntarily make non required low difficulties for their difficult spreads.

Given these three points, I don't think it's a good idea to further reduce the amount of intermediate level maps coming into the game.

it is true that lower diffs have more ppl to play, but it doesn't mean they need even more maps. the problem is that higher diff players have few alternatives to play, while lower diff players can enjoy other maps

The part of the reason for 2 could rather be a lot of supplies and abundance of existing beatmaps. Providing more maps won't help adding more plays to new maps

In my personal experience, I could simply dl a lot of old maps to play with dt (which is like 5*ish) which I found enjoyable. I rarely play new maps for that just because there is no need to play new maps, unless it is mapped for a recent popular song I want to play.

For higher difficulty nm maps, even if I dl all the old maps it's difficult to enjoy a lot of different maps that I'd specifically enjoy since overall number of maps are very small

I don't necessarily agree with the wording nor usage of "proper spread containing at least X difficulties", main issue being the definition of spread varies between maps and just may end up being cheesed. For example, 4 similar extra diffs that are gradually more difficult from one to another still counts as a spread with the way that the proposal's currently worded.

Nao Tomori wrote:

1. A material percentage of plays on higher difficulty maps come from the low difficulty spread.

This only applies to stuff that are Insane and above, as most of the playerbase's skill level averages around the 4*-6* area, anything below Insane doesn't neccessarily have lots of playcounts from what I find basing off N-X fullspreads like R3m's maps for example.

And thus, I would rather have an approach similar to ctb and mania in which the actual drain time is reduced instead, that way we can still make sure Insane/Extra lower difficulties can still be made instead of having the spread being cheeseable based on your wording above.

Riana wrote:

In my personal experience, I could simply dl a lot of old maps to play with dt (which is like 5*ish) which I found enjoyable. I rarely play new maps for that just because there is no need to play new maps, unless it is mapped for a recent popular song I want to play.

For higher difficulty nm maps, even if I dl all the old maps it's difficult to enjoy a lot of different maps that I'd specifically enjoy since overall number of maps are very small

full support for this reason. high difficulty but short length maps are prohibitively difficult to push for ranked for asinine reasons that are entirely unrelated to their quality.

there should be something done because the current system is extremely discouraging for anyone mapping these unless you have an excess of free time to deal with the bloated set requirements.

SupaV wrote:

I don't necessarily agree with the wording nor usage of "proper spread containing at least X difficulties", main issue being the definition of spread varies between maps and just may end up being cheesed. For example, 4 similar extra diffs that are gradually more difficult from one to another still counts as a spread with the way that the proposal's currently worded.

mania's RC defines "proper" spread as this:

A "proper" spread for difficulties Insane and harder is defined as a spread with gaps in difficulty similar to those between lower difficulty levels as specified in the difficulty-specific criteria.

the spread has always been a subjective matter. such definition will make it possible to prevent using multiple similar difficulties for spread with bns' own judgement, just like how too large gaps in the spread has been prevented.

SupaV wrote:

And thus, I would rather have an approach similar to ctb and mania in which the actual drain time is reduced instead, that way we can still make sure Insane/Extra lower difficulties can still be made instead of having the spread being cheeseable based on your wording above.

I don't disagree with reducing the drain requirement itself, but it impacts more mapsets that don't have difficult maps while very high diffs would still require a pretty large spread. thus I thought the proposed solution would be better for solving the problem

nao's argument assumes that the amount of plays on maps by people who comply and make massive spreads to fill in gaps outweighs the plays that currently don't exist at all from the maps that can't get ranked due to our unnecessarily restrictive spread rules, which i personally highly doubt

it also implies that any change to lessen ranking requirements should never happen if that change puts at risk any amount of plays that a given mapset sees, which is obviously ridiculous because that would mean any sacrifice shouldn't be done irregardless of gain

regardless, even if the original assumption was correct and ignoring the problems with the argument, there are so few high star maps in general compared to lower star maps that any plays on them are negligible. because of this, it only makes sense to focus on ease of ranking, making this an obvious and easy change to support (not to mention how arbitrary spread rules are to begin with and how much success other gamemodes have seen in loosening them...)

however, instead of the rather messy wording changes proposed in the OP, id simply change "...there cannot be any drastically large difficulty gaps between any two difficulties" in the general rc to "there cannot be any drastically large difficulty gaps between any two difficulties. This does not apply to any difficulty above the first Expert in a mapset." this rule could be moved to per-gamemode in order to comply with each gamemode's special naming system

Nao Tomori wrote:

3. People generally do not voluntarily make non required low difficulties for their difficult spreads.

and whyd u think thats the case

Because it's time consuming and uninteresting to make low diffs for 250 bpm boss songs? None of my arguments relate to mapper preferences, I recognize it's a burden, my arguments are that that burden is generally worth it for the game, not that they don't exist.

UberFazz wrote:

however, instead of the rather messy wording changes proposed in the OP, id simply change "...there cannot be any drastically large difficulty gaps between any two difficulties" in the general rc to "there cannot be any drastically large difficulty gaps between any two difficulties. This does not apply to any difficulty above the first Expert in a mapset." this rule could be moved to per-gamemode in order to comply with each gamemode's special naming system

I agree that the wording can be improved, but I'd like to mention that the suggested wording is different from my suggestion

beatmapsets/1939252#mania/4224583

For example, this map has 2 difficulties with "proper" spread from top diffs, 9.03* and 9.72*.

The suggested wording would require an insane diff, a low extra diff, and the top diff (if I understood it correctly)

I personally prefer progressive spread as mania does, as it would further fulfill needs for higher diff levels but it can be discussed further

Nao Tomori wrote:

Because it's time consuming and uninteresting to make low diffs for 250 bpm boss songs? None of my arguments relate to mapper preferences, I recognize it's a burden, my arguments are that that burden is generally worth it for the game, not that they don't exist.

I recognize the value lower diffs provide too, but reducing burden can encourage more high sr maps to be pushed, which would help solving the problem of the extremely small pool of high sr maps

The proportion of the loss from the same number of sets would be marginal for lower diff players, since they already have huge numher of alternatives to play.

For those who want to enjoy boss songs with low skill level, allowing "sparse" spread with large gaps (or something like uberfazz's suggestion), with the same lowest diff requirement can also be an option. like how many arcade games provide lower diffs with fixed number of diffs

agree that it's an issue that harder maps need way more total drain time across the spread and that just reducing the thresholds does nothing to address this

also agree with Nao that low diffs are important for the game

also, there is no conflict between these points; sets of harder maps are bloated by filler Experts, not low diffs

therefore +1 to UberFazz

I think there's an important reason to have a wide variety of difficulty levels covered. I think UberFazz proposal works for me. Covering low diffs of a variety of songs particularly new and popular ones is fairly important for the games growth.

I agree w/ uberfazz riana. There's too much effort in mapping/modding big high sr spreads hence the demand of 7+ star maps. No one wants to rank a 4:15 IXXX spread. Fwiw, I personally am not satisfied with the small selection of 7 star maps in ranked.

UberFazz wrote:

"there cannot be any drastically large difficulty gaps between any two difficulties. This does not apply to any difficulty above the first Expert in a mapset."

Uberfazz's suggestion sounds really good. Low diffs will still be required while making high sr maps wayyy less daunting to rank. Most high sr spreads are bloated with extra diffs that lack submitted scores anyways, so loosening spread requirements for extra+ diffs would barely affect a mapset's playcount.

I don't see how this wouldn't be a win-win for everyone.
- new players still get low diffs
- high level players will see more difficult maps being ranked
- mappers won't need to spend an insane amount of time/effort making high sr mapsets
- bns can check less diffs

After a quick peek into the ranked very difficult maps, I noticed something that makes me doubt if the filler extras in very difficult mapsets are actually negligible:
(I didn't check those maps with 7-8* top diff because 1-there are too many and 2-it will most possibly require only 1 more "filler extra" which I don't believe is an important addition to the modding effort)

1) For maps with 8*-9* top diff, nearly 90% of the extra diffs between the lowest expert and the top have more than 50% of the PCs of the top diff's, sometimes even higher, and I don't think this ratio can be called "negligible". IMO this shows that players, even the highly skilled ones, do need a proper & continuous spread to figure out which difficulty range fits them the most in a mapset with a very difficult top diff. Hence, I think this reason doesn't sound valid enough to relax the spread requirement for these maps.
2) For maps with 9*+ top diff, some of the filler extra diffs start to have sub 50% of the top diff's playcount, sometimes multiple of them. So I think starting from here, we enter the difficulty range of interest, in which for the current playerbase there aren't any noticeable differences. In this case, I think we can probably say some of the filler diffs aren't necessary. But then, following the rapid improvement of player skills, this boundary line can be changed over time, so I don't think this standard can last long. (SR can be tricky sometimes too.)

---

These being said I don't completely oppose this proposal since modding a full spread for, say a map with 12* top diff, IS exhausting. If our initial motivation is to give highly skilled players enough hard maps to play and make their ranking process less painful, then I think what Riana originally proposed suits this idea better. If we do need a compromise between new players, highly skilled players, and mappers, the best I can come up with for now is:

1) Removing the "there cannot be any drastically large difficulty gaps between any two difficulties" phrase in general RC and changing its following sentence to "Each difficulty and their spread must comply with its mode's difficulty-specific ranking criteria...". By doing that we can let every mode decide their proper spread rules (that can also avoid conflicts among modes since what Uberfazz said isn't compatible with the current mania spread rule);
2) In the standard specific RC, adding that deleted phrase back but with our standard, here "there cannot be any drastically large difficulty gaps between any two difficulties. This does not apply to any difficulty above 9* in a mapset."

In that case it should probably just be "for difficulties above the first extra difficulty in the spread" so then the target map spreads (stereotypical 9* top diff sdvx song) would basically look like nhi-x---x at a minimum, which is a bit more reasonable than the currently required nhixxxxx?

I prefer progressive spread from the top diff but many ppl seem to have concern with diminishing amount of lower diffs

limiting the change to 9*+ top diff would reduce the supply of 7*, 8*ish diffs, while those levels also don't have sufficient maps for target players

I think uberfazz's proposal is a better common ground

or how about uberfazz's proposal + allow the "first extra" level to be any extra diff that are reasonably distant from the top diff?
Like for 9* top diff, the extra can be whatever in the 5*~8* range

for the extremely high difficulty level, it's likely that very large portion of them were not pushed for ranked because of the spread requirement.

Supply of one extra level from much increased number of extremely difficult maps could sustain or even increase supply of the "filler" high extra levels for such songs imo, along with increased number of maps of the "filler"'s diff range as a top diff

Hi, spread rewrite opinion here and sorry for kinda hijacking the thread but i feel like it's important we be transparent what we worked on so far

We have basically this idea as one of the things for the brainstorming we did for spread (https://docs.google.com/document/d/16qaVUv5yfKUp6JLnuTMdjrK6XjgffEBsdcuJW2-QL10/edit Approach 3). The stuff to figure out there was just how much progression is reasonable and what is unreasonable

We seemed to generally be in agreement that something needs to change but not exactly on the nature of what because we recognize Aika's points regarding "how much target audience skipping is reasonable on boss maps"

Which is why we wanted to propose smaller changes in increments to first get maybe more modes on the same page about the draintime requirements before we go with something way more drastic like Approach 1 we had outlined

We were kinda stuck in the following:

1. we wanted player data to make more educated statements (not just playcount but how many unique people play which difficulty levels)
2. deciding on one of the approaches is difficult unless we just decide to go in steps(well maybe not the more out there ones, but the safer ones)
3. I wanted to have some sort of far-reaching survey because a change to how spread is handled usually has far-reaching consequences
   
   1. I do not know how to write such a survey without making it super biased lol
   2. A few people doubted such data would be useful either way

I'm just throwing this out here, in general i think the proposal itself is good so if this has chances of passing I'd instead urge people to try working out a solution for more than just osu! as the gamemode

I don't think I have an opinion on the specifics discussed so far but will just add a +1 to either OP or UberFazz's, both are in the right direction for me. or any of 2, 3, 4 from Oko's google doc (I get the point behind #1 but feel like it would create awkward social situations, future mappers benefit from the past mappers' effort despite no explicit collaboration, idk maybe im overthinking)

small rant that basically amounts to "I don't think spreads matter"
I generally don't think it's the mapper's responsibility to create filler maps at all. if people want to make full spreads with shared motifs and all that jazz that's awesome but I also have no issue with someone ranking an Expert by itself or even some oddball EEXX spread provided the maps are of high quality on their own. from a player perspective I want mappers to create what they're passionate about and polish those things up for ranking; they're just playing the game too and it feels so corporate to me that we expect essentially unrelated contribution out of them to meet some kind of audience diversity requirement... I don't view that hypothetical EEXX spread as "missing NHI", rather that people bothered to create and rank 4 diffs in the first place that could be good fun for interested players.

clayton wrote:

I generally don't think it's the mapper's responsibility to create filler maps at all. if people want to make full spreads with shared motifs and all that jazz that's awesome but I also have no issue with someone ranking an Expert by itself or even some oddball EEXX spread provided the maps are of high quality on their own.

I partly agree with it, but I also agree with ppl's concern on the reduction of the new userbase, if new maps for new songs are lacking.

IMO separate category for pp/leaderboard with large amount of sets allowed (unlike loved), with less promotion than ranking maps (so that mappers would still have an advantage ranking maps) could be nice for mappers and already experienced players but not sure about good implementation

this one is out of the scope of this proposal tho

--

most people seem to agree that some changes to reduce the burden of high difficulty things would be beneficial

There are a lot of possible alternatives including ones from oko's google doc
Two main ones are:

1. Relaxed spread for extra diffs (uberfazz's proposal)
2. Specifying number of diffs (or specifying set's total drain)

Concerning difficulty levels are:

1. Mainly the loss of the lower diffs
2. possibly "filler" high difficulties (mentioned by aika)

IMO difficulties around lower extra (till around low 6*) would have sufficient supply regardless of rules
because a lot of frequently mapped songs have commonly acceptable maximum diff levels around that (for a lot of songs around bpm 180, with mainly jumpy w/ some burst kinda maps)

The uberfazz's proposal can be a basis of the change, since people are mostly concerned with lower diffs, and the proposal covers the problem while relaxing spread requirements for higher diffs.

UberFazz wrote:

"there cannot be any drastically large difficulty gaps between any two difficulties. This does not apply to any difficulty above the first Expert in a mapset."

Things I'm concerned with the og uberfazz's proposal is:

1. it does not impact high 6* or 7*-ish diffs much, one lower extra diff can cover the spread for such diffs already while it is also mainly concerning "high" diff levels. The "first expert" level would have large supply regardless of change, for the reason I mentioned earlier
2. it cannot cover the aika's concern on the reduction of the "filler" high diffs, but I think the gain would be higher for the

one way to solve them could be:
Allow larger gaps for extra diffs spread including insane-extra gap.

We could define spread more in a way that it'd explicitly allow large gaps like 1.5~2.5* for the spread between insane+ diffs.
7* can have high 4* or 5* insane as lower diff level, and boss lvl have halved extra diffs required. The reduction in the "filler" higher diffs would be smaller, and more of such diffs as a top diff from relaxed spread would surpass it.

using SR as a reference wouldn't be ideal, so a speed up ver of diff can be a reference for the spread gap definition, around 1.5x (DT) might be ok (typically 5* -> 7*ish, 7* -> 10*ish or so).


also it'd be really nice to have more statistics as oko mentioned
and changes can be made gradually after observing the actual impact, i'd like to see at least small changes implemented

The problem about spread proposals is that they rarely (if at all) include opinions of players (who don't map), especially novice or lower skilled players. It's clear that most mappers agree with making things easier for mappers, but since this type of changes have a large impact on new content for the playerbase, it's important to gather feedback from everyone affected by it.

Serizawa Haruki wrote:

The problem about spread proposals is that they rarely (if at all) include opinions of players (who don't map), especially novice or lower skilled players. It's clear that most mappers agree with making things easier for mappers, but since this type of changes have a large impact on new content for the playerbase, it's important to gather feedback from everyone affected by it.

my original proposal would definitely need more feedbacks from lower-end players, but uberfazz's proposal and alternatives of it do not impact the lower diffs requirement, so I think the low skilled players are not affected parts of the playerbase

and for higher level, bunch of mappers are also players of such skill level so mappers' opinions in player's perspective are also relevant. the main motivation for the initial proposal was also the lack of the contents i felt while playing nomod maps, and complaints from some players as well

some surveys on necessity or the perceived quantity of maps of different level of expert diffs would be cool tho

Regarding surveys/feedback, I think any such effort should be focused on gauging opinions on the current state of Ranked as well as data about what people are actually playing. I don't see a survey of the actual options presented here telling us more than which option would benefit the largest group of people, which is pretty useless when we want a solution that'll balance the needs of the entire playerbase.

Riana wrote:

for the extremely high difficulty level, it's likely that very large portion of them were not pushed for ranked because of the spread requirement.

From my POV there are some good maps that do feature difficulties of all levels, but do not seem to get BN attention, for instance Ov Sacrament And Sincest. So I guess the issue is more on nominators not wanting to mod a set with a billion difficulties rather than mappers not willing to conform to spread requirements.

Castagne wrote:

Riana wrote:

for the extremely high difficulty level, it's likely that very large portion of them were not pushed for ranked because of the spread requirement.

From my POV there are some good maps that do feature difficulties of all levels, but do not seem to get BN attention, for instance Ov Sacrament And Sincest. So I guess the issue is more on nominators not wanting to mod a set with a billion difficulties rather than mappers not willing to conform to spread requirements.

I agree with this too, but it's inevitable imo
checking tons of spread diffs is just painful even if you want to push the set from the bn side as well

while I like to push some difficult maps to play, it's really difficult to decide to mod sets with many spread diffs when I know it could cost like hours for merely a few minutes of the top diff I want to push

edit: it could be a bit better with activity check accounting for drain time etc but i still don't enjoy spending most of the time checking fillers

/

about stats or surveys, there are too many options, and idk which are really needed
just doing everything possible would just end up in infinite data acquisition without a conclusion

imo stats on the "drain time vs ranked diffs of certain SR range" could be useful, and can be used to show the impact of the proposal
I'd like to see more opinions on what's needed, and why it is needed

also with minimal stats/survey, spread proposals can be experimented with some grace period and rollback option imo
without experiments nothing would move forward, since the actual impact is unknown till it's tested

edit: community/forums/topics/1893597?n=1 this post contains some relevant statistics

I can probably provide other statistics if requested. However, I am limited by my terrible programming skills and osu api call limit of 50.

fully agree with what uberfazz suggested, this already kinda is a thing in taiko as seen by occasional 4* -> 7* spreads getting ranked and there doesnt to be any issue there, but given that osu has bigger playerbase i think its beneficial to include lower extra as well. its a compromise that both makes spreads more lenient and doesnt impact any player audience much

HE HE HELLL NAHH
mappers must be forced to make normal difficulties so my 10-year-old brother can play osu! too

on a serious note, this proposal sounds like an excuse to neglect lower difficulties as most people, prefer to map in the insane+ range.

In all honesty, these 10*+ maps you guys rank nowadays push the limits of what is properly expressing the music, it's like mapping a breakcore song without rhythm simplification.

The majority of osu! ranked mappers are at least somewhat experienced players (rank 500k or below) and most osu! players ... are rank 500k+ with how fast this game is growing nowadays. Why should these ranked osu! mappers cater to such a small audience?

As much as I and many experienced mappers would LOVE to not have to account for lower difficulties when making a set, it's unfair to those who are not skilled enough to play insane+ difficulties.

why did u not bother to read any of the suggestions that would avoid that specific issue

The topic needs to be bought into the beatmap management forum since that's a thing now, it cant just die like this

it's on the right forum. and not dead either, there has been a lot of related spread discussion as part of community/forums/topics/1836226