Hello Everyone!
After a discussion with mania BNs while we encounter new issue that prevent player from playing the game effectively, we want to propose a new guideline regarding minimum amount of LN's length required for a beatmap to be ranked.


The proposal will be something like this:

"Avoid using extremely short Long Notes (with a length of shorter than 1/12 or the equivalent of 28 ms). Too short of Long Notes are unreasonable to play with full accuracy due to the game engine's limitations and the player's average pressing time for each keystroke"

This will fall under the guidelines section of the mania page, which is based on 180 BPM and will be scaled according to each song.


==================

FAQ:

Why is this guideline necessary?

Answer: Recently we encounter an issue where there's some charts that using an extremely short LNs that makes it almost impossible to actually hit the LN with 300 judgment if their PC is not on the high end of the spectrum (as some of the player feedback said exactly this). We believe that while we support mapping creativity, we don't want to restrict the players to simply have the best PC just to hit those LNs as we want to make all maps be as accessible as possible to all players, and the LN that is too short contradicts that point.



We want to emphasize that this proposal is NOT FIXED by any means. Anyone feel free to suggest an alternative and we would gladly take it into consideration!

yes

Can't agree with this more.

Will take a look through the replay data before suggesting a number but as of now, I'm actually inclined to think that the shortest rankable LNs could be 300-ed with 300g not being a hard requirement. Weird tech / flam / "manip" patterns can get by without massive pushback or acc concerns so short LNs could be treated in the same vein.

agree with lenpai. have seen many instances with 30ms LNs that have been deemed fine by players. 15ms not so much. so in my understanding if there were to be a hard minimum limit on LN length, it should be between there. 20ms lets say. although i will very often stray away from using long notes of such short length, 40 as a hard cap is excessive.

Guess I should express more about my opinion.

As what we can see from the data (if it's correct), 40ms already way too short for a *very very strong majority* of the players to hit perfectly (300g). Probably because of hardware limitation but not skill reason.

Yea, mappers can maps "weird tech / flam / 'manip' patterns" with shorter LNs, but I don't think such things should be pushed to ranked. (yea, you can say 300 already good enough, we don't have to force 300g, but I can't see the significance behind this...) Also I completely don't think it would be a huge problem or hurt the overall quality of the map by replacing such short LN with 40ms ones.

Just my personal opinion.

Thanks Maxus for doing this and shout out for you still sticking with mania, real quick to express my idea about that.



At first, I wanna say two concepts/premises about this:
-
#1. Judgment of LN's ending
As we know, the judgment of LN had many different processing methods within the different mania-like games. Some had ending judgment or others not. For those who had, that means your LN release still counts the accuracy
-
#2. how to define 1,000,000 score.
We can easily find RC said: "The Auto mod must be able to achieve the full score (1,000,000) in all difficulties.". But I think that means no notes got overlapped, cuz that osu topus can't hit it properly so that humans can't too.
-
And now there had a small problem about this (I don't know if I worry too much or not) does get 1,000,000 important?
-
If Yes, we should give a chance to hit very short LN to get max, and the possibility can be lower but can't be impossible.
-
If No, getting a high acc of LN doesn't important, notes won't happen thing like this cuz they just need to simple hit, but LN contained hold even hard to hold or make a combe break is fine.



ok now when I just saw datoujia's link and I feel it's interesting, so the deeper, I saw it, I recognize the player's control ability on hold had a limit. So I think this proposal is worth discussing based on "osu!mania had a judgment of LN's ending" and "get 1,000,000 score is important (maybe)"
-
that's my idea about this, if we can make a clear description about this, it will reduce the misunderstanding and misleading to everyone, I guess posting and posting in maps is really exhausting for each other, right. When we had a rule/formula, anything can be referred to, and more reasonable to discuss.

Thank you for the feedback so far!

After discussing again, we decided that we will change this rule to be a guideline and it will be based on snap rather than miliseconds.

The changes of the guidelines :

""Avoid using extremely short LNs (with a length of 1/16 and less). Too short LNs are unreasonable to play with full accuracy due to the game engine's limitations and the player's average pressing time for each keystroke"

this falls under guideline category, and it might be more lenient or stricter depends on the BPM. (it uses 180 BPM as the base).

Feel free to continue the discussion again!

After doing some quick calculations, I found out that a length of 1/16 of a whole in 180 BPM is already at around 20ms (20.867ms precisely), and that is possibly beyond the limitations of many lower devices. It may also cause severe accuracy drops for players who are not physically capable of pressing and releasing in such a short time (because of hand injury, for example).

Therefore, my suggestion would be to extend the minimum snap to 1/12 (27.778ms) for 180 BPM, which is probably a more reasonable choice.

Also to correct _Stan: The AUTO mod can hit overlapping notes in some occasions. For example, AUTO can get max on a slider end directly inside another note. It's also possible to get a higher score than AUTO: AUTO completely ignores hit notes stacked together.

YyottaCat wrote:

After doing some quick calculations, I found out that a length of 1/16 of a whole in 180 BPM is already at around 20ms (20.867ms precisely), and that is possibly beyond the limitations of many lower devices. It may also cause severe accuracy drops for players who are not physically capable of pressing and releasing in such a short time (because of hand injury, for example).

Therefore, my suggestion would be to extend the minimum snap to 1/12 (27.778ms) for 180 BPM, which is probably a more reasonable choice.

Actually, the current ruling already implies exactly what you want there.

Because the guideline specifically saying "avoid using LN at 1/16 snap and below" which means 1/16 is also included in the snap that should be avoided for a LN. (therefore, saying that 1/12 will be at least the recommended snap).

Of course the wording probably can be changed in case it causes misunderstanding, but we will try to adjust it as more feedback comes!

Thought I'd put my opinions into the matter, because as that have been said, I do also feel like some things could be developed.

Muse Dash wrote:

As what we can see from the data (if it's correct), 40ms allready way too short for a *very very strong majority* of the players to hit perfectly (300g). Probably because of hardware limitation but not skill reason.

We can't necessarily say that the actual hardware is to blame; in this RC Guideline creation; Technology is the independent variable (changing). This means everyone has different types of technology, and that to be said, means technology is out of the matter (at least in my opinion). But as said, I believe that if there is some sort of player that can at least 300g this, then it's fine. If people have different pieces of technology, then that's not the games problem in my opinion.

Muse Dash wrote:

Yea, mappers can maps "weird tech / flam / 'manip' patterns" with shorter LNs, but I don't think such things should be pushed to ranked

I generally disagree with this statement. There are going to be mappers who have maps like this. For example, beatmapsets/1499627 (a map of mine)
Now, it might be a bit weird to use a map of mine, but hear me out.. this actually is useful to the discussion. There is a 1/16 ln in a 144bpm map (28ms) (Example: https://i.imgur.com/xeWQSRL.png). Which generally also gives a bit more information into how we can combat this guideline (because technically speaking, I don't think that this rule is in tip top shape yet.

I can agree with -mint- with the uses of LN's based on lengths.. but sometimes there are some maps that have some arbitrary expression that is actually a lot more better to express in those forms than in the forms of longer LN's, rice.. etc. Even then, we can't see more than a few maps with these types of expressions, and if then, they have some good enough expression to counteract THE amount of expression they're doing.

_Stan wrote:

And now there had a small problem about this (I don't know if I worry too much or not) does get 1,000,000 important?

But the overall key point I want to ask, along with Stan's quote is:
"If getting 1,000,000 on Human Experience count more than the actual osu!topus itself."

From a person who plays LN.. I don't think myself that a person SHOULD be able to hit a perfect 1,000,000.. but as it stands, it all depends on the opinions of many other people.. because when you look at it this way, most maps with a lot more LN perspective onto them are more aimed at those players. Likewise, patterns must be meaningful to song intensity, and due to some players who map stuff like so, generally to put more emphasis on patterns, I believe that there should be some sort of part in the proposal that:

"Avoid using extremely short LNs (with a length of 1/16 and less). Too short LNs are unreasonable to play with full accuracy due to the game engine's limitations and the player's average pressing time for each keystroke. This excludes maps that can provide proper context to how certain LNs have been used"

(with knowing Maxus's opinion, I am a lot more on board with the current proposal)

At least from my context (feel free to reply to anything I have said)

YyottaCat wrote:

Also to correct _Stan: The AUTO mod can hit overlapping notes in some occasions. For example, AUTO can get max on a slider end directly inside another note. It's also possible to get a higher score than AUTO: AUTO completely ignores hit notes stacked together.

thanks for reminding me of that, gracias.

[ Decku ] wrote:

_Stan wrote:

And now there had a small problem about this (I don't know if I worry too much or not) does get 1,000,000 important?

But the overall key point I want to ask, along with Stan's quote is:
"If getting 1,000,000 on Human Experience count more than the actual osu!topus itself."

From a person who plays LN.. I don't think myself that a person SHOULD be able to hit a perfect 1,000,000.. but as it stands, it all depends on the opinions of many other people.. because when you look at it this way, most maps with a lot more LN perspective onto them are more aimed at those players. Likewise, patterns must be meaningful to song intensity, and due to some players who map stuff like so, generally to put more emphasis on patterns, I believe that there should be some sort of part in the proposal that:

"Avoid using extremely short LNs (with a length of 1/16 and less). Too short LNs are unreasonable to play with full accuracy due to the game engine's limitations and the player's average pressing time for each keystroke. This excludes maps that can provide proper context to how certain LNs have been used"

At least from my context (feel free to reply to anything I have said)

well yeah, I agree with you, maybe I say something wrong and it's not proper to make an example with 1,000,000 things, but all I wanna say is "I think this proposal is worth discussing". No matter what consequence we got, I just wanna that we can express things like this with clear words written in RC.

For the record, my comment about hyper-technical / "manip" patterns are for rice oriented patterns and they go through just fine even if getting straight 300g is very unlikely.

Might be a controversial option but I do also like to mention that high end charts arent made to get very high ratios -- just as how high end charts are meant to test physical skill, tricky LNs are made to test reading and coordination. Though I'd argue that sub 20ms LNs aren't a great idea with an engine like osu's.

Going back on track, I've decided not to download the replay data due to the size which wouldnt fit in my laptop so I got some 2nd hand info that the standard pressing ranges around 35-40ms.

That being said, hitting a very short LN dead center at +-0 would grant 300 even at od 8 (+-40.5ms). If we would factor in potato PC polling rates (reference: https://drive.google.com/file/d/1wE_k3uODZHvC4Is_CW27dFFvp6f7dDac/view ), we're looking at about a 50-55ms pressing time. which would in turn at least allow LNs as short as 20ms when hit within a margin of error of +10ms. (30ms LN length for +20ms margin).

Very short LNs in a sense, would incentivize hitting earlier which is probably unintended behavior.

Feel free to correct or comment about the observations / assumptions I made.

For transparency sake, I'll move what was said on another server here for everyone to be able to see and discuss. While the update above does go over these points, it's important to echo the community's approval/disapproval with their unique reasoning.

- Personally yes, there should be a guideline, not a hard cap to the slider's length. It is unrealistic to say that in some cases, these are in fact not justified for what it's worth. As [REDACTED] said they should be contextual to the difficulty. Adding to that, I really do not see how it is bad on ""targeting"" a map to make changes in the RC that might help future cases if extrapolated correctly. Correctly as in not specific to that map and that map only.

- While I'd normally be reluctant to make specific wording that takes into account the average BPM in the RC (Mainly because most of it in mania ends up being... well, very nilly willy), something as sliders like this is a bit more "objective" so I'm certainly not against the current wording proposed.

However, the terminology used on the proposal should refer to LNs as Sliders, as that is the way they're mentioned all throughout the RC.

- At the end of the day, any change made should be made very carefully as to not dissuade certain viable mapping styles, while ensuring that adding such guidelines improves the QoL of maps to come. I'm certainly not against the proposal, but I'd really want to see how it develops out on the forum thread!

Small Edit: If there are inconsistencies with previous wording changes, do keep in mind this conversation was simply branched out. The person below me is correct here and thus should be kept as is, for example. This was just more of "for the record" than anything else.

RandomeLoL wrote:

However, the terminology used on the proposal should refer to LNs as Sliders, as that is the way they're mentioned all throughout the RC.

They're called "long notes" in the mania RC though.

---

After reading the above forum posts, I think the guidelines should be written as:

(For the insane and expert difficulty)

1. Avoid using 1/16 or higher snap long notes. These are unreasonable to play with full accuracy due to the high demand for the game engine's limitations, hardware capacity, and the player's press time for each keystroke. Use regular notes instead.

---

Now that I think about it, maybe 1/9 should be the limit for the guideline according to press time data given by lenpai, which would make the preferred limit 37.037ms. That number falls within the range of 35ms to 40ms and should be acceptable.

I do think that more discussion would be needed to finalize the guideline.

YyottaCat wrote:

1. Avoid using 1/16 or higher snap long notes. These are unreasonable to play with full accuracy due to the high demand for the game engine's limitations, hardware capacity, and the player's press time for each keystroke. Use regular notes instead.

"Avoid long notes length 1/16 snap or smaller." is prob better wording

[ Decku ] wrote:

We can't necessarily say that the actual hardware is to blame; in this RC Guideline creation; Technology is the independent variable (changing). This means everyone has different types of technology...

I am a bit unsure what 'technology' actually meant in your argument.

If you mean 'input' or 'hardware'. Then I really don't know what to say. Game is for everyone, it's never fair to rank a map for 'some sort of player' to enjoy. And drive the rest mad because they could never acc such notes.

If you mean 'skill', again, unless I've misunderstood something, the data already shown clearly it's almost impossible to acc LN shorter than 40ms, and... I completely don't think this can be improved by practice.
https://twitter.com/transcendtoinf/status/1440331000127442956?t=tgElVwdRnwl_RPMbMDF6yw&s=19

[ Decku ] wrote:

I generally disagree with this statement. There are going to be mappers who have maps like this.... Now, it might be a bit weird to use a map of mine, but hear me out.. this actually is useful to the discussion. There is a 1/16 ln in a 144bpm map (28ms)

Let me apologize first, But... I really think use past maps as examples are invalid argument in this case,The propose why we discuss things here is because we want to add 'guidelines'. You know the guidelines are keeping update and almost all the maps ranked years ago are against the current guideline.

Also. I do feel it completely won't hurt the map if you redesign the LN length in the part you mentioned into something around 40ms.

Last but not the least, I think guideline should be set rigorous, because it could be broke. Use LNs shorter than 40ms... It's already an uncommon and controversial case, and it definitely worth to be confirm each time IMO.

--------------------------------

Thought I think few people hold the same opinion as me, but still... Thanks for reading my opinion and reply. ;w; Anyways, 1/12 at 180 BPM as a ***guideline*** is a bit loose IMO.

I would say to have the duration in ms into the wording as well, since the BPM is at the start of the RC guideline; this makes it a little bit confusing for mappers who just look at a particular part of it without noticing where it says everything is based on 180 bpm (of course it is assumed, but this would make it slightly more accessible to newcomers to ranked scene).

imo, unless the mapper make maps for ranked regularly, it is hard to comprehend/calculate the base BPM part of the guideline in relation to the map's BPM (this extends to some of the currently used guidelines as well, but I digress).

A bit off-topic but some stuffs in Catch mode RC uses ms currently and to me (as an extremely beginner mapper in that mode) it is much easier to comprehend than BPM + snapping in mania guidelines.

So, having both as an option here would be nice. In the current proposal, 180bpm at 1/16 snap would be around 21ms - so we can also have that in parentheses.

Apart from that, I support this change as getting good acc on the game (whether it be marv or perf) should not be a matter of a coin flip. I would say the current length proposed by Maxus is fine for me for this (personally I think around 25ms is where it gets too limited for mappers, so I would not agree with having a minimum LN length above that even as guidelines)

It would depend on each BPM but, it should be more likely a guideline that "LNs should be applied under the reasonable intention and release shouldn't be in the level that too demanding and requires concentration to the point of being impossible, considering the game engine's limitations"

"FAQ: Why is this rule necessary?

"... we don't want to restrict the players to simply have the best PC just to hit those LNs ..."

Mappers don't have to be restricted on their creation either, I think creativity should come first than each player's PC spec.

The current proposal is aiming at the LN that the length would cause an unnecessary concentration to hit with a perfect acc and it's good to have a guideline for it. But I'd like to disagree with that "all maps be as accessible as possible to all players" in FAQ.

Some players would feel uncomfortable on hard beatmaps because they might think it's unreasonable, some would feel uncomfortable with some patterns because they might don't agree with them. Same as those, some would feel uncomfortable with LNs because their PC spec(or keyboard, gear) isn't really great to get a full acc. But that wouldn't mean every LN should be comfortable to get a full acc, right?

And sometimes LN patterns get tricky when the mapper wants to emphasize the sounds and it'd be sad if it gets changed due to this proposed rule, not a guideline.

Providing an accessible ranked map is necessary but, this one should be more likely a guideline.

Mappers don't have to be restricted on their creation either, I think creativity should come first than each player's PC spec.

The current proposal is aiming at the LN that the length would cause an unnecessary concentration to hit with a perfect acc and it's good to have a guideline for it. But I'd like to disagree with that "all maps be as accessible as possible to all players" in FAQ.

I think i need to clarify something regarding this matters.

Your creativity does not rely on making an extremely small LNs way beyond what osu engine can do, and neither does you absolutely need to always have an LN way beyond what most people engine capable to, in order for you to be "creative", and i think we can agree on this point.

There are a tons other ways for you to be creative over certain songs , as the mapper we have the luxury to still pursue many many another alternatives for our mapping in order to preserve our creative point while still emphasize playability for the player. (making your LN from 1/16 to suddenly 1/12 in order to follow the guideline does not suddenly make you "not creative", i hope people can agree with this as well).

from my point of view, it's simply a selfishness at this point to not try amend a guideline that's simply being made for the betterness of the players. As the mapper we always have many many other leeways of alternative for other type of pattern that still preserve our 'creative integrity'. For players, they don't have such luxury over that. You are ranking your map for other people to play, so i hope you do understand that considering player's perspective also fall under mapper's responsibility as well.

and yes, i do think it's selfish when you force player buy a better PC just for them to be able to hit your 1/24 LN, when you can just change to 1/12 and it won't affect any of your creative ideas.



Also to avoid any other misunderstanding this time:
This is already a guideline, this is not a rule anymore. Please read the thread again. Thank you.

I went around to dig to see if I can find data for whether it's possible to find some sort of intersubjective LN length where a player could release a note and get a rainbow 300 even though they're hitting the note as if it's a normal rice note. Even when you play a map that's fully rice, you still hold the keys down to some degree, so finding that natural point I think would be a good start.

I think, while I was mapping short LNs while taking scorev2 into consideration, 45-50ms LNs were the sweet spot for me. The 300g release window for OD8 scorev2 is +- 24ms (16ms * 1.5), so this does align with with lenpai said with how long players naturally hold down their keys for (35-40ms) to some degree. I think players can release a bit earlier (25-30ms?) than that if the LNs appear visually short enough, but there comes a point where the LN end becomes completely invisible because of how most players have a cut off LN end. I think the sweet spot for forced staggers for scorev2 is about 35-40ms, but even then I think that's a bit much.

This is, of course, different from the RC suggestion here, for two reasons: We are not making maps under the consideration that every map's LNs can be SSS'd fairly cleanly, and that we're using scorev1.

I'm generally of the belief that a lot of LN maps, especially ones that are extremely technical (Kamah (Scythe) and Labyrinth come to mind), that they are not meant to be analysed with the intent of getting very strong accuracy in mind. I think, because of this, the appropriateness of an LN length highly depends on the map and the section where the short LNs are present. If the map is meant to be difficult to SS, then we shouldn't take the 300g window into consideration, but the 300 window in general. The 300 window is much larger, and if you wish to put a hard cap on the length of a LN, you have to keep this in mind.

The other thing is that scorev1's LN mechanic is really dysfunctional and, frankly, way more convoluted than it should be.

You can find a summary of how LNs in v1 are supposed to work, but there are a few weird things that happen in v1 that aren't well-documented:

LNs in score v1 are based on two criteria - the gap between last LN head hit and the LN head (let's say, LNStart), and the combined time of the difference between the last LN head hit and the LN head (LNStart) and the difference between the LN release and when you're supposed to release the LN (LNEnd).

Here are the timing windows for OD8, for the first criterion (LNStart):
300g: 19ms (16ms * 1.2 rounded down)
300: 44ms (40ms * 1.1)
200: 73ms
100: 103ms
50: 127ms

Basically, if the last LN head hit is 20ms from the LN head, you will never get a 300g. If your last LN head hit is 14ms, as long as it meets the second criterion, you will get a 300g.

The second criterion works like this, keep in mind this accounts for the combined time of LNStart and LNEnd.
300g: 38ms (16ms * 2.4 rounded down)
300: 88ms (40ms * 2.2)
200: 146ms (73ms * 2)
100: 206ms (103ms * 2)
50s: 254ms (127ms * 2)

Assuming that your LN start input is 10ms from the LN head (which is pretty high but let's go with this), you have +- 28ms to release the note, 4-5ms larger than v2. You can probably afford to have 40-55ms LN lengths if you want players to treat the notes as normal notes while still keeping a visual accent. For staggers, I think the minimum length allowed is 30ms given what I mentioned about average stagger ability (40ms minus 10ms).

Of course, if you take the 300 window into consideration, the minimum LN length is much much lower if you take the release window into account (~78ms to release a short note). By then I think it stops becoming a matter of release length and more of how long the LN has to be for the LN tail to be completely invisible, because an invisible LN tail can cause people to accidentally overhold when they don't mean to. I think 20ms is a safe number for most skins (I remember people complaining about some of the mini LNs in Elekton's LOSHAXI to be invisible, and they're about 18ms), so that should be the theoretical lowest that any map should use in the ranked section. If you wish to apply a rule, 20ms is probably best. As a guideline, I think 30ms-35ms (accounting for framerate and polling issues) is fine, but the guideline should only be used for sections that are clearly meant to be on the easier side and intentionally SSSable.

I don't believe framerate and hardware are that much of a limitation here. It is true that framerate will cause the timing windows to be a fair bit smaller, but most players play osu! on at least 60 FPS and have a decent enough keyboard for timing. It would reduce the timing window (both release and input) by some amount, but I don't think that should be the main reason for limiting LN lengths. The main reasons have to do with LN tail visibility and the player's ability to naturally release notes and stagger notes in general.

I admittedly don't have much stake in the discussion because I don't use LNs anymore, but my suggestions are findings are based on my personal experience and general analysis while I was mapping with LNs in mind and my experience with fiddling with both scorev1 and scorev2 back then. Ultimately I think a 30-35ms minimum guideline is best, but it being a rule makes no sense given that there are many LN maps that are supposed to test short LN release at a very high level.

edit: didn't see maxus' post before he posted lol, ignore the rule/guideline part

I changed and tidied up some things to make it more obvious that it's guideline, i hope it's already been clear it's guideline now. (no more rule please.)

And i find that shoegazer suggestion does match up with Lenpai in a sense that 30ms seems to be reasonable amount for guidelines based on the explanation and data that they give.

Probably the only question left is that should the rule be written by using miliseconds or by snap terminology to be consistent with other mania RC wording?

[ping] did suggest to add the miliseconds, but it will be great to have more input for that.

I would also suggest to use miliseconds as it is easier to picture things out than using snap terminology by any mean.

Even if that's a must, there should be a formula somewhere to explain it, or make it clear for others to understand and follow it.

Or, like Ping said, why not both. People could easily overlooked the standardized BPM and snap in question and seemingly unaware of it when they look at that specific part alone. And even when they do, would they know how to apply that to their own map?

After further discussion while implementing community feedback that we got so far, we decides to implement both the snap terminology and the milisecond clause at once. (Credit to RandomeLol for suggesting the proper wording!)

So after lenpai and shoegazer assesstment, we come to a conclusion that the guidelines should be at 30ms (as the data that both of them show in the thread, suggest exactly that), and we implement that into the wording as well in the following statement:

"Avoid using extremely short Long Notes (with a length of shorter than 1/12 or the equivalent of 28 ms). Too short of Long Notes are unreasonable to play with full accuracy due to the game engine's limitations and the player's average pressing time for each keystroke"

^ once again, this will come under guidelines section, so i hope no one misunderstand about this.

Pretty much this gives people a rough snap to compare with BPM scaling, while also having miliseconds number to gives easier comparation for more mappers, so we find that this should be the most fair number and wording for everyone.


If you guys have any further concern, feel free to post in the thread and have a discussion.

1/12 is not the equivalent of 30ms and "LN" should be changed to "long note" to match with the rest of the RC.