[Guideline Addition] Content bloat and score inflation

posted
Total Posts
54
Topic Starter
yaspo
Hi, this is a guideline proposal.

The proposal is in in a similar vein as https://osu.ppy.sh/community/forums/topics/844073 , but applies to a different case.

With a growing collection of ‘PP mappers’ banding together on mapsets, we’re starting to see quite the amount of sets that have a lot of difficulties with a similar star rating and very similar design ideas. This causes content bloat, and could be described as artificial score inflation.

To explain why it causes content bloat, let’s start with defining that in this context. Given the varied nature of ranked content, full of different songs, maps and mappers, the ranked content generates different play experiences by design. Content bloat then could be described as the opposite :
“The absence of different play experiences and/or the obstruction there-of”

By this definition, it’s clear that sets designed around players being rewarded for retrying multiple difficulties many times cause this issue. Spending a lot of time on 1 song on 1 mapset already cuts or obstructs the time spent playing maps on different songs by different mappers. If each difficulty is also designed around the same difficulty elements, emphasis methods and/or patterning style, then the player is also derived from different gameplay.
You could also say this causes a very disproportionate play distribution. Players at the lower levels of play get very very little content, while players at higher levels get too much content. While there is no need for equality, something more equalized would be better.

The reward, PP, is the main cause of this happening. If players see an opportunity to inflate their score and gain ranks, or even are afraid that they’ll fall behind, they have no choice but to take it. You could say that those players are to blame for it and not the system or that they are probably enjoying themselves, though I believe it’s important to protect players from themselves. Many of them are in fact not enjoying this, have accepted rank decay despite finding going for ranks fun otherwise, or disagree with the system even allowing this entirely.

Additionally, when several difficulties have the same design and star rating, it means they target the same audience. Players who can set a score on one difficulty can set it on the others as well. This essentially means players are given the chance to duplicate their score multiple times over. This artificially inflates score and is evidently unhealthy for the competitive side of the game.
Disproportionate play distribution causes issues here as well, giving players at lower levels much less opportunities to rank up than those at higher levels.

To fix these issues, I propose the following guideline :

  1. Avoid having many difficulties with similar rhythm, intensity and other emphasis methods on the same set. This is to avoid unnecessary bloating of the ranked section and artificially increased scores.”


Edit: Lots of discussion happened and the proposed guideline changed quite a bit, current version can be found on this post https://osu.ppy.sh/community/forums/topics/912089?n=40

What this is intended to prevent : Mapsets like https://osu.ppy.sh/beatmapsets/931452#osu/1944926 and https://osu.ppy.sh/beatmapsets/950289#osu/1985282

These should simply not exist under the reasoning provided; they are the exact reason for this proposal existing. Their highest difficulties use similar rhythms, especially in their most intense parts. Their main method of emphasis (big sharp angled jumps) makes minor differences across some difficulties negligible.

What this is intended to limit : Mapsets like https://osu.ppy.sh/beatmapsets/513590#osu/1091249 or https://osu.ppy.sh/beatmapsets/935111#osu/1952177

Monstrata’s Zen Zen Zense is an example of 2 difficulties being similar as a result of a concept. While not disallowed, notable differences should be made between the difficulties.

Sotarks’ Gold Dust has 3 difficulties at 5.55 stars with the similarities in the described aspects. This pushes the boundaries of the reasoning given and should be avoided.

What this should not limit or prevent :
  1. Mapsets that have a lot of difficulties with different designs. This still provides a unique gameplay experience. Megasets are a good example of this, as they often focus around individuals expressing themselves.
  2. Mapsets that have 2 difficulties with similar deisgn, given they are made by different people. These seem inevitable and aren’t nearly as harmful.
  3. Similar maps of the same songs on different sets, hosted by a different user. These require a full set to be made and more nominations, this evens out the amount of potential bloat causes.
  4. more .. ?

I do want to avoid the proposal covering cases it shouldn’t, it’s made to target very specific mapsets and not to cause collateral damage. Another issue here is vagueness. What is ‘many’ and when are difficulties too ‘similar’, can the mapping community agree on these? Personally I find a case by case basis judgement to be best for this, so maybe precise wording isn’t necessary.

I’m also trying to avoid adding direct reference to PP in the guideline, since we still have to rely on 3rd party tools to check the rewards. It also has too little to do with actual mapping.

So, that's my thoughts. I'd like to hear yours as well
Basensorex
thanks for saving mapping yaspo
OliBomby
It must state very obviously in the rc how this guideline is only meant for a very specific type of abuse. You don't want someone to interpret it in a way that will make them post problems on discussions like "This spread is not creative enough".

Still I think it's a good guideline addition if it gets handled well.
Sylvarus
I really like the idea of this, seems like it will be very subjective though.

Does a set like pishi's Reign of Fear fall under this guideline? The difficulties are extremely similar but one could argue that difficulty setting and star rating differ enough between at least some of them to consider them rankable?

Also not entirely sure about the "different sets" part. Different users entirely sure, but what about user A ranking a set of song X and then making a GD for the ranked set of the same song hosted by user B? In my opinion this proposal should affect that case as well, if the GD is too similar to the ranked set it is bloated content.

Ailvs Saifu Kamiasobi no Uta is also a possible case for this, at what point does the ranked section become bloated when the same mapper maps the same song in a similar style x amount of times?
Topic Starter
yaspo
Right so
@Olibomby
I agree! "I do want to avoid the proposal covering cases it shouldn’t, it’s made to target very specific mapsets and not to cause collateral damage."

@sylvarus
Pishi's reign of fear would be debatable under these rules, but not entirely unrankable. The difficulty settings are definitely a plus, but it would be recommended to design more around them in a few instances.

The GD thing is already covered in another guideline, the one added through the proposal at the top.
Ailv would fall under this rule as well in case the map is too similar. Though if the maps are different enough there is essentially no reason to complain; it counts as unique content.
MaridiuS
Who remembers when megasets were used for gimmicky songs to showcase all kind mapping styles like Axion? https://osu.ppy.sh/beatmapsets/115193#osu/324516

Throughout the osu history even devs have tried imposing the idea of limited amounts of difficulties per spread range but were shotdown by community as they were way vague and didn't have a place back then.. Nowadays it is precisely what is needed to avoid content bloat and what yaspo proposes would fix that. https://osu.ppy.sh/community/forums/topics/420223 At that point it wasn't received greatly as it offered nothing but an inconvenience therefore the community's complaints were loud, so the proposal was not implemented.

Why it would be good to implement:

1.Ever since osu existed you couldn't gain points by doing the same thing, if I fced the same map twice with same acc I wouldn't get awarded for it which makes sense, I already got awarded. Nowadays like justadice or harumachi clover, mappers thought of adding GD's to avoid that limitation and use the most convenient methods in all the maps for performance gain.

2. It's as yaspo said, it just encourages people to play the map they already basically played to stay relevant and not feel like crap for dropping thousands of ranks for denying to do it. Implementing this rule can save players from themselves as playing the diffs to stay relevant can cause fatigue and tilt due to failure to come up with the terms. Having a more diverse section with more songs and styles to boost their rank makes the grind much more interesting. A single set of the same song with 10 diffs is a lot to take in.

3. Would inspire mappers to be more creative and add more to the ranked section.

What I would add as the guideline is that no more than two difficulties of same spread range may be mapped in a similar fashion. It is for members of BNG and NAT to decide whether the difficulties fall under that category.

Two would be fine because when you ask someone for GD you won't know how it will turn out, but having more than that is just too much to be looked over. This kind of guideline existing adds incentive to BNs being careful of nominating such maps and gives other BNs the possibility to veto such a blatant issue without mappers and other BNs using the famous "buT itS NOt In raNKInG CrITeRiA" argument.

Against the argument off "this will limit the freedom of mappers" well sorry it may but, maps are being ranked for the sake of the community not just the individual mapper that wants to rank it, you're not disallowed to map the song if your heart wills it.
Mayflower
i love this so much thank you yaspo
Sameurai
I absolutely love this proposal, thanks for your hard work putting this together yaspo! I hope this gets the attention it deserves :)
abraker
I am against this because I specifically make my difficulties to build on top of one another. My style is progression across difficulties, I like establishing a core concept and building upon it across difficulties. I don't want this guideline to wreck havoc on such concept.

Instead of complaining people that mappers are creating similar maps, make different maps yourself and get them ranked. I hear everybody complaining that ranked is stale because there is little variation in mapping. What is preventing you from making maps with the variation that is lacking? Come on, people.
dennischan
I agree with Yaspo and Maridius that something should be done about the excess amount of pp maps in the same set.

I find Maridius' guideline especially viable since it clearly defines the maximum amount of difficulties you are allowed to have for each spread range. Which is why I'd like to change the wording a little bit to take in account Maridius' opinions.

  1. There should be no more than two difficulties with similar rhythm, intensity, difficulty spread and other emphasis methods on the same set. This is to avoid unnecessary bloating of the ranked section and artificially increased scores.”
abraker

dennischan wrote:

I agree with Yaspo and Maridius that something should be done about the excess amount of pp maps in the same set.
solution is simple: fix pp

I find it ridiculous that we need to limit map designs because of a broken calculator. This should not happen.
Duskyui
yaspo i love you and this proposal
Nao Tomori
progression across difficulties is the exact opposite of what this guideline is opposing lol.

however i think this guideline is ultimately pointless as it can be easily surpassed (for example, sotarks gets one of his clones to host a second set and then is able to even bubble it himself) and is simply a bandaid fix to the more overarching issue of bns purposely promoting horrifically bad maps as well as the pp system failing to accurately rate difficulty by ignoring several factors and overweighting others. unless you expand it to apply to copies of the same *song* being mapped in the exact same way, i don't see how this is useful in the slightest.

and even then, it won't do shit because you can stick amazing repeats on 1/4 and simple rhythm and a 2* diffspike with 1-2s on just about any song and get some amazing bns to bubble it and bob's your uncle. not like pp farmers care what song it's on, just whether they can get their new top play...
Tyistiana
I can see your good perspective for writing this proposal, but I personally against it.

I want to give the example based on my mapset which 7 mappers participate to create a GD for this mapset results in 3x(ENHI) spread. Petit Rabbit's - Happiness Encore

The pace of the music has pretty forced the mapper to make the creation of their difficulty become similar to each other (pattern structure is almost the same for my Oni difficulty and Blacken's Oni - some small difference has been shown, but in overall I'd say that it is still similar). Which this proposal gonna prohibited this kind of mapset too. But as you wrote here,

yaspo wrote:

What this should not limit or prevent :
  • Similar maps of the same songs on different sets, hosted by a different user. These require a full set to be made and more nominations, this evens out the amount of potential bloat causes.
So, it means that if I want to make this mapset become rankable according to your proposal, this mapset must be split into 3 mapsets instead. Which it is not match to my intention to host this map for "Making the player enjoy the song while no need to download too many mapset - All in one". To split this mapset into 3 mapsets will cause the song folder of the player becomes larger unnecessarily (since the player must download all of the maps to be able to play them all and thus causing the player to download redundant mp3 and bg). And for a nomination for this kind of map, even the map gets separate into 3 mapsets, each host can still manage to find the same nominator to nominate it.

In conclusion, practically, this kind of proposal may can't help anything. As the mapper/GDer can host a new mapset to avoid this proposal's restriction by hosting another mapset while asking for the same nominator to nominate - which this causes the song folder to become unnecessarily larger.
Aiseca
Reading the proposal, it seems to me that the potential loopholes of this proposal is within the proposed guideline itself. You need to iron out the wording, defined and objectively leaning.
Taeyang
Hello, everyone~ My English abilities are quite poor so there may be mistakes here and there.

Originally, my definition of a Ma”pp”er was “PP(providing pleasure)”, not just “PP(Performance Points)”
If you look at my recent ranked or pending maps, you will see that they are not just maps for gaining PP

What I find enjoyable about maps when playing is the “movement of the cursor”, which is why I started mapping in the first place. I wanted people to enjoy my mapping style.
It just so turns out that my style of mapping was easier to farm PP than other maps, and eventually people started to misinterpret my definition of “PP (providing pleasure)” as “PP(performance points)”. I was fine with this since it still gave people a sense of accomplishment by gaining PP (performance points). This eventually lead me into thinking that “performance points” was a part of “providing pleasure”, and I just have been mapping like that since.

Enough with my own story. Let me explain from my point of view how this situation turned out like this.

  1. First, the PP system is focused too much on jumps and streams.
  1. Second, usually mapsets with a lot of difficulties tend to have shorter songs, which means that the play count will stack up quickly.

These two are the main reasons why maps become popular.

Most people want their maps to get attention, and if they do not people won’t receive motivation for people to continue mapping.
I feel like Yaspo’s view is more like a temporary solution and does not tackle the two fundamental problems stated above.
It's like you've taken away the fulfillment and the freedom of mapping from countless players by vetoing JUSTadICE(https://osu.ppy.sh/b/2044364)
I wonder if the adverse effects of the PP system can limit the mapping style of a mapper.
Of course, some people want more patterns, but some people will enjoy the current difficulties as they are.
Lots of Hype Trains, favourites and Play Counts provides evidence for that.
A player has their rights to choose and play a map, so it does not make sense to take away the freedom of mapping just because a map is similar to other ones.

For example, when https://osu.ppy.sh/beatmapsets/744772#osu/1570203 was ranked, it only had 10000 play counts after a whole week.
But it became a popular map in just a day, when a_blue HDDT FC’ed it.
More people had played the map after it went in the Popular Beatmaps section, and I was happy to see this since many people would gain a sense of accomplishment after gaining PP.
(You won’t know the feeling of numerous people sending you “thank you” messages…)

Also, the standard for determining the rank of Popular Beatmaps is the play count. This lead many mappers to map shorter songs, while a lot of good long songs received less attention.
For example, https://osu.ppy.sh/beatmapsets/890174#osu/1860808 is a very interesting and well-made map, but many players do not know about this and the play count does not even exceed 30000.
Do you think people would want to rank maps that are well-made but unnoticed?

I just want to say that the so called “PP maps” are worth being ranked as they give players a sense of fulfilment.
I think a faster way to fix this is to discuss ways to tackle the fundamental problems, rather than these guidelines.
One thing I would like to add is that my mapping style is definitely distinct from those of other mappers, but if they become more distinct than that then that would just be over-mapping or a mapping style that people do not want…
As you said, it is good to put variousity in mapping styles for a more diverse experience, but I want more people to enjoy my maps

I will stop here for now, and I will answer after further observing the situation.
MiGGo

abraker wrote:

dennischan wrote:

I agree with Yaspo and Maridius that something should be done about the excess amount of pp maps in the same set.
solution is simple: fix pp

I find it ridiculous that we need to limit map designs because of a broken calculator. This should not happen.
pp will never be perfect and it doesnt matter how many times pp is "fixed" there will always be someting to exploit and abuse. Stop blaming the pp system for mappers exploiting it, the pp doesnt exploit itself. Grow a backbone please and add this guideline so mapsets that deserve to be unranked jump map compilations stay in the graveyard.
Topic Starter
yaspo
Alright so, there seem to be 2 main counterpoints for the proposal
1) Limiting sets in themselves doesn't prevent the maps from getting ranked through different means
2) This doesn't come close to fixing or approaching the root causes, like the pp system, mapping for playcount, circlejerk, etc ..

1) For a while I felt like the pros would outweigh the cons, but now I more or less feel like both the current situation and the one that would result just bring different issues to the table (partly due to 2). In essence it's just a game of pros and cons
- 1 big content drop vs multiple smaller ones
- 1 normal/hard/insane vs multiple similar, generic ones
- downloading a song once vs seeing it appear multiple times
etc.
In that sense, yeah, the approach should be adjusted.

2) While I understand this, it's also the mentality I'm trying to counteract with the proposal. With fixes to the performance system currently happening, it's not unreasonable to also try fencing in the problem from the other side. It's just a matter of finding the best band-aids we can find, in a sense.

Other stuff
- Tyistiana : yeah, that's a case I didn't consider, combined with above it's clear that the current proposal isn't valid
- Taeyang : I do get this sense of accomplishment, and I'm fine with PP maps existing. But, in the case of JUSTadICE the mapset originally gave 8 times the same sense of accomplishment, as the maps are about as difficult and provide the same rewards. Players being spoonfed 'accomplishment' quickly turns into something meaningless, bloated. That's mainly why I think limiting it per set is a good thing.
There's also a lot of players that just dislike this constant mapping for attention. You might not get their messages, but there's plenty of them. So I'd like to look for a better balance.
If the fundamental issues you mentioned were 'quick fixes' then they'd have already happened, or aren't happening for some reason (eg. https://github.com/ppy/osu-web/issues/2730 this thing that never happened)
- Maridius : imo I'd prefer an allowing guideline over a super strict one. Binding things to spread gets kinda pesky. Generally I'd rather tell mappers to "c'mon don't do this shit" rather than "you must work within this small space of options"

Anyways, I'm kinda brainstorm mode right now. People seem to agree with the premise of the proposal, but the execution has issues. So time to try a different angle on it.
Things that come to mind are

  1. Avoid oversimplifying Extra difficulties. Players at this level require a suitable challenge.
    [counteracts more blatant design choices done for pp mapping, also disconnected from the set transferring issue]
  2. Avoid having many similar sections across difficulties in the same set. This is to ensure that every difficulty gives a different experience.
    [Loosens up the proposed guideline by not overfocusing smaller similarities. Gives room for mappers who have different approaches to the same thing, like Taeyang refers to. Also encourages actually acting on it, making sections different seems easier than ranking another set. Still conflicts with Tystiana I think, but maybe just not fit for other gamemodes .. ?]


I'll add more ideas when I think of them. Thanks for all the feedback so far, though open for suggestions and even more feedback!
Mun
I fail to see how this prevents the same content bloat occurring when a mapper can just take those similar difficulties, make more sets with exactly the same everything else and some new lower diffs, and rank that without running afoul of this guideline. I'd recommend that if this is going to have any effect at all, it should come with an additional restriction not to make redundant sets of the same song, as that is a clear attempt to dodge the guideline.
abraker

MiGGo wrote:

abraker wrote:

dennischan wrote:

I agree with Yaspo and Maridius that something should be done about the excess amount of pp maps in the same set.
solution is simple: fix pp

I find it ridiculous that we need to limit map designs because of a broken calculator. This should not happen.
pp will never be perfect and it doesnt matter how many times pp is "fixed" there will always be someting to exploit and abuse. Stop blaming the pp system for mappers exploiting it, the pp doesnt exploit itself. Grow a backbone please and add this guideline so mapsets that deserve to be unranked jump map compilations stay in the graveyard.
Because thus far it has been one bandaid fix after another. A real solution to fixing pp has yet to be proposed. What you are failing to understand is that when pp is fixed this proposal will look archaic and as a desperate attempt. Its a short sighted solution.
show more
Please sign in to reply.

New reply