Agree. Songs that support low diffs should have low diffs, and songs that are sufficiently complex where making a low diff is actually hard should be exempt.

there might be a problem of song complexity being subjective (like, there exist people who say 15 red lines is complex timing)

but yeah we need this, id even say starting from insane would be fine enough - one of examples that comes to mind are high bpm deathstream songs, those are hard to make low diff spread work in general

completely agree with this, whether or not it is applicable to certain songs is up to the bns to decide, as with all other guidelines. if a bn makes wrong judgement, as usual, there can be a problem stamp or equivalent veto proposed.

sabaku de is a good example of where this could be applied in practice beatmapsets/1360892#osu/2815973

the solo is basically unmappable in a normal diff and the only reason the spread was able to get ranked was by mapping the intro on the top diff to get it over 3:30 drain

since this is my screenshot (lol) ofc i agree

will be left up to BNs to decide what is and isn't "complex" and even in grey area scenarios any possible negative impacts are hugely outweighed by the positives

also should consider the "lowest necessary diff" be left up to BNs as well, im afraid hards could be just as tricky to make a difficulty out of for songs like these (similar to what's mentioned here)

here's my rewritten proposal:

1. All game modes within a beatmap 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. For songs that are inherently complex (i. e. contain frequent timing signature changes or constantly employ uncommon rhythms), this rule does not apply.

OR, for a more strict rule,

1. For songs that are inherently complex (i. e. contain frequent timing signature changes or constantly employ uncommon rhythms), spread may start from an Insane.

i think hard diff is still doable on most of these complex songs. i would suggest starting from hard diff and going up as there are very few songs that truly cannot be simplified to an acceptable hard diff level. even stuff like sputnik the hard isnt that detached from the song that it not existing would be better. plus, the most played diffs tend to be light insanes so getting rid of those potentially has a worse impact than normals would.

that would work too, id just personally prefer to leave judgement to BNs on a case-by-case basis even if the use case for a min insane would be rare

would like to hear more opinions on this specifically

if its kept to normals i think i wouldn't be opposed to this

some normals get really stupid like beatmapsets/1403806#osu/2910203 where i just felt like a normal doesnt even make sense to have

i mean technically you could make a normal for any complex song if you ignored enough of it but that would be pretty bad right? in the interest of not doing that and ending up with a normal that follows nothing and feels arbitrary i think this could be good

I completely agree with this. I find mapping normals for variable BPM songs almost impossible. It's also the main thing stopping me from finishing my one mapset, as the song has a new red line every measure

+1

should define complex a little more objectively but i think its a good suggestion

agree w/wafer, needs a more concrete definition of "complex" baked in before this should move forward - if it did i could get behind it

There is no target audience for these difficulties

This is simply not true, new players can enjoy playing these maps even if they can't follow the rhythms perfectly or play the map well. If someone likes the song this already gives them a reason to play it. Also all the higher skilled players who play low diffs with mods typically enjoy more complex rhythms as it gives them a bit more of a challenge, so there's definitely a target audience for it.

Creating a coherent experience is exceptionally difficult

From a mapping perspective, I don't really think this is a valid concern as all the examples mentioned have either been mapped as Normals or could easily be mapped. Of course songs with complex rhythms/timing are much more difficult to make low diffs for than "regular" songs but any experienced mapper should be able to pull it off (and inexperienced mappers would probably not manage to map higher diffs for these songs either). In order to create a coherent experience, it's necessary to understand the concept of rhythm simplification very well and to know how to choose rhythms that are as intuitive as possible. A lack of experience in this area is probably the main obstacle for mappers but the logical solution is to either practice more and ask for advice or get someone else to make a guest difficulty.
And from a playing perspective, I think it just needs to be accepted that some maps are harder than others despite being on the same difficulty level. This is true regardless of complex rhythm or timing. For example, swing beat songs are generally more difficult than songs with simple 1/2 rhythms, but that's a weak reason not to have a Normal for such songs.

The reason I supported this is because there are certain genres of music that consist entirely of strong sounds with no break for the listener. I'm not talking about popular music, but the kinds of compositions that no one maps.

Looking at SPUTNIK-3 and Biomes (Cut Ver.), I think these low diffs are cool and work because of the dynamic nature of the composition (tension and release, etc.). And having a Normal difficulty allows players to enjoy these sets when they otherwise wouldn't be able to.

I don't have any good examples for the music I'm talking about since I don't listen to it myself, and I'm unsure if mappers will start mapping this kind of music after this change (or if BNs would be willing to nominate them).

TL;DR: Experienced mappers can and should bring value to their sets with low diffs. I wanted to allow mappers to experiment with more advanced compositions that tend to be uncomfortable to listen to (aka not popular music), but the kinds of mappers that are capable of mapping this kind of music may also be able to create low diffs for them.

please understand the argument isn't whether or not mappers can make low diffs for these songs, it's whether or not they should be forced. we could sit here all day long debating whether or not the sputnik normal is a valuable addition to ranked, but i see no reason not to relax requirements for such a niche subset of maps

the arguments presented against spread removal in the related thread do not apply here

arguments near the level of that thread (such as the possibility of a huge change in what the ranked section could look like and/or a noticeable deficit of low diffs) would be the ones most relevant, since what we're talking about here is forcing volunteers to go through extra hoops just to rank a map when they're already inherently jumping through many just by mapping a complex song

just because there's "technically an audience" or "you could make them with enough experience" doesn't mean they should be mandatory

agree with nao/mir/wafer

The points brought up look valid to me too, even though I'm usually an advocate for making lower diffs.
Some songs are just way too complex to make a Normal out of them, unless you decide to do dumb rhythms which would disregard the song (but then, what are you even mapping?).

I'd still argue it depends on the mapper's skills, though. Even on complex rhythms, good mappers can make something out of it that isn't completely unrelated to the song and still playable/enjoyable.


Not sure how to define the word "complex". Perhaps if the main rhythm is shorter than, say, 4 measures for more than 15 seconds? This wording implies that the main rhythm keeps changing "too fast" (gives less than 4 measures to adjust your focus) for "too long" (it prolongs for more than 15 secs). In such a case you'd be obliged to keep mapping different rhythms, rhythms which may easily overlap in ways that would result in too many beats used for a Normal.
I bet pretty much everyone (or at least BNs) can hear when the main rhythm would require too many objects. (right? RIGHT?)


Regardless, this applies to Normal diffs, not Hard diffs. Those diffs have already enough room to accommodate complex rhythms in intuitive ways imo.

I agree that normal difficulties can be infeasible for some songs. I wouldn't be opposed to some allowances for there, but I think there should be some stipulations put in to avoid skirting the rules if possible.

Over everything else, the mapper should always at least attempt to map a Normal difficulty. Finish one, really. At that point they and BNs should be able to judge if the difficulty makes sense for ranked, or if it's not possible to make a decent Normal for the song. What should be avoided is "giving up" before trying to map a Normal.

I absolutely encourage the idea of not forcing people to make horrible normals. I just dunno what kind of precedence this will make in people trying to argue themselves out of a normal diff.

Until a somewhat solid threshold of complexity can be established, I am not sure if I can agree with this proposal.

As the mapper myself, I can feel it. Songs that aren't suitable to map the Normal difficulty may be omitted, but this can be rigorously difficult as the staffs should take a look on case-by-case of songs which aren't suitable to map with Normal difficulty.

The clear repeat offender of complex beat is, you know, II-L.

Sylvarus wrote:

Over everything else, the mapper should always at least attempt to map a Normal difficulty. Finish one, really. At that point they and BNs should be able to judge if the difficulty makes sense for ranked, or if it's not possible to make a decent Normal for the song. What should be avoided is "giving up" before trying to map a Normal.

I'm agree with this. Also, there should be a clause about discussing it with nominators whether the Normal difficulty is utterly bad to even include with such a complex songs.

i'd like to revive discussion on this as it seems most people agree that this should be implemented at least in some form! i think it would open the doors on a lot of things for the ranked section.

here are two examples of songs i mapped that have rhythms/timing complex enough that any normal (and arguably even a hard) difficulty made for them would be extremely low quality
beatmapsets/1689255
https://osu.ppy.sh/beatmapsets/1599997#osu/3267722

as the number of timing points goes up, essentially any rhythm in the song becomes a complex rhythm and there's not much that could be reasonably simplified for a new player to latch onto.

in the wobbly map, the entire song is end-to-end with constantly changing rhythms, essentially leaving no option but to put long sliders that couldn't represent anything or circles that will frustrate the player because it's impossible to tell which of the many sounds it is intended to represent.
in archangel, there is more of a loop that gives new players the chance to predict rhythms, but as all the rhythms have no true value, and there are multiple timing points each second, a difficulty in normal-hard range would still be frustrating as the rhythms are continually dense enough that much would have to be skipped, and the rhythm would be so repetitive and unexpressive that it would not be a good experience for a new player.

i think there is no question that a normal difficulty would be out of the question, but i would argue that for each of these maps, a hard difficulty would be unenjoyable for the players of the skill range even when mapped as well as possible.

so it leaves these maps in an awkward place in the community, where the only option for leaderboards is loved, though i personally believe the quality of the maps themselves is good enough for the ranked section.

Yrea i agree as well, especially for those 2 cases dsco posted as they are kind of crazy ass cases where such a removal/replacement should benefit, at the very least there should definitely be an option IMO

In the first place I don't really think it's ideal for new players to be playing songs like these when they are first introduced into the game considering resetting inner metronomes and shit is like another extra step thats added for more fucky songs like these, and starting out the game ur just trying to figure out how any form of rhythm works at all so its just naturally harder within the same realm of difficulty for normals

probably not just "dense timing point = no normal" tho since theres a lot of very easily mappable songs

I agree with what's suggested.

Another perspective I would like to add is the more difficulties being forced into a set, the more reluctant BNs/NATs are to even check/nominate it in the first place. (Complicated song + full spread + added length) means a BN would have to check all diffs throughout the set, which takes marginally more effort than 1 difficulty throughout a 4:15+ minute song, or 5 diffs over a simple 1:28 minute song, etc).

This essentially punishes the mapper for putting more effort than the average set for reasons outside of his control. Can't help but feel this discourages creativity, effort, and encourages mappers to take the easy way out for the sake of getting through the ranking process more easily.

pr has been made @ https://github.com/ppy/osu-wiki/pull/7084, if anyone has feedback on the wording posting it either there or here is much appreciated. wanted to just get something out there as we seem to agree on this being a good idea

we are looking for more input from other gamemodes because spread isn't exclusive to standard

+1

mania-wise, I feel like reading complex rhythm for lower levels is easier since the point of reference to hit the note is easier to follow than the somewhat "ambiguous" approach circle, but following those rhythms in the first place is very hard even with the ability to use 1/4 in normal. For most new players, it would be fair to assume that most new players would not have a musical background or get exposed to more complex songs on a frequent basis and therefore would only be comfortable hitting simple rhythm. This is y I agree that only normal and not hard should be optional for complex songs like what dsco provided as examples.

So far the only example I can somewhat get behind is the Wobbly song because the rhythm and timing changes are both very frequent and very drastic, but the conclusion to exonerate mappers from creating a Normal diff still doesn't seem logical (more on that below). The Archangel example is much different because the timing changes aren't as radical and the song actually has a quite steady rhythm, although not accurate.

dsco wrote:

...rhythms/timing complex enough that any normal (and arguably even a hard) difficulty made for them would be extremely low quality.

Do you mind elaborating on how exactly the song's complexity affects the quality of the map? I don't see any correlation in this regard because it depends entirely on how it's mapped. I disagree with the argument that simplifying the song's rhythms a lot results in low quality, otherwise any low diff for songs with high rhythm density would be awful which is certainly not the case. Of course any Normal diff for a track like this is going to be more challenging from a rhythmic perspective but that has nothing to do with quality, some songs are just naturally harder than others.

dsco wrote:

...that will frustrate the player because it's impossible to tell which of the many sounds it is intended to represent.

This is true but I don't see how it only applies to Normal and/or Hard. In fact, any difficulty is very likely to be perceived as frustrating - but because of how the song is structured, not because the map fails to represent it.

Xilver15 wrote:

Another perspective I would like to add is the more difficulties being forced into a set, the more reluctant BNs/NATs are to even check/nominate it in the first place. (Complicated song + full spread + added length) means a BN would have to check all diffs throughout the set, which takes marginally more effort than 1 difficulty throughout a 4:15+ minute song, or 5 diffs over a simple 1:28 minute song, etc).

This essentially punishes the mapper for putting more effort than the average set for reasons outside of his control. Can't help but feel this discourages creativity, effort, and encourages mappers to take the easy way out for the sake of getting through the ranking process more easily.

This is definitely the case but it's not directly related to this topic because the same can be said about any map that has more than the bare minimum number of difficulties etc. It's not uncommon for bigger mapsets to be rejected by nominators purely because of the amount of time and energy needed to mod them (which is understandable though), regardless of song complexity. This is an issue that has to be addressed in a different way.



Another thing worth mentioning is the fact that there didn't seem to be an attempt to map a Normal (or even Hard) for several examples of tracks used as arguments to corroborate this proposal, and that there was seemingly very little input collected from actual players of this difficulty range.
Also, the phrasing doesn't clearly define what kind of songs would fall into this category.

+1

i agree. your points make sense completely

Serizawa Haruki wrote:

This is definitely the case but it's not directly related to this topic because the same can be said about any map that has more than the bare minimum number of difficulties etc. It's not uncommon for bigger mapsets to be rejected by nominators purely because of the amount of time and energy needed to mod them (which is understandable though), regardless of song complexity. This is an issue that has to be addressed in a different way

This is a false equivalence; more diffs on a set than the amount required are a conscious choice by the mapper and are not required for ranking. This post is referring to instances of BNs being reluctant to check sets that only have the bare minimum of diffs due to the complexity of the song.

Xilver15 wrote:

This is a false equivalence; more diffs on a set than the amount required are a conscious choice by the mapper and are not required for ranking. This post is referring to instances of BNs being reluctant to check sets that only have the bare minimum of diffs due to the complexity of the song.

Okay but my point still stands because following this logic any song that has timing changes or variable bpm should be exempt from creating a full mapset, even if there are many songs of this kind where mapping a Normal is not problematic at all (as proven by the large number of said maps in the ranked category). From my understanding this proposal is not aimed towards every "complex" song, only those that are exceptionally difficult to map low diffs for.
I'd also argue that mapping a song that is extremely complex regarding rhythm or timing is a conscious choice by the mapper, too. If someone is not willing or capable to face the challenge that comes with it they can choose to map a different song. It's similar to how mapping a Normal diff for songs with 1/3 rhythms or high bpm "tech" or metal songs is typically more difficult compared to a simple anime opening - it's the mapper's responsibility to choose a song they can reasonably map.

bump

Adding this is something I mostly agree with and would allow for these niche songs to be better accommodated for in the ranking criteria, but the wording of a change like this should be really clear to avoid people abusing it.

Would it be possible to provide examples linked under the rule in the RC to demonstrate what would be considered 'sufficiently complex' enough to omit a normal difficulty from a spread?

Doing so would make it more clear as to what is meant by 'sufficiently complex'. Maybe an example of a song with lots of timing such as the one dsco mapped beatmapsets/1689255#osu/3452047, and one with complex snapping changes such as stuff from II-L beatmaps/artists/172, or whatever else is suitable.

going to close this thread as the proposal was essentially rejected on the github discussion
see https://github.com/ppy/osu-wiki/pull/7084