osu!standard is going through a substantial change right now, where sliders are going to require accuracy. I think now is the chance to discuss whether osu!mania also needs a change to its LN mechanic. What's your take on this everyone?

Right now LN gives a varying combo based on how long you can press it, this creates issues, such as:
1. When playing Hidden, the vision area keeps going up from simply holding an LN.
2. You're allowed to release and re-press an LN, something that is not allowed in any other scrolling rhythm game that I know of, and encourages mashing on LNs instead of trying to play them correctly when there are difficult LN patterns.
3. It's hard to keep track of your real combo when playing note+LN patterns, when the combo from one LN alone is worth a lot when it's really just one keypress.

Personally I think the 2 combo per LN style is the best, since those are the two points during an LN you actually have to time something and play to the rhythm, whether it's press or release.

I see no logic in gaining combo from simply holding a finger down... you're not exactly following any rhythm by doing so, the speed that the combo increments is entirely arbitrary and has nothing to do with the music. It's illogical both on a gameplay level and on a rhythm/musical level.

Also, it results in inconsistent combos on scoreboards. Two people can FC a map with different combos. It looks really stupid.

1. Personally think the flaw is with hidden anyways lmao, but agree that LNs definitely make it worse

2. I kinda doubt it's the issue with allowing the re-pressing of LNs, rather the less strict timing window. I don't really see many cases of people benefiting too heavily from repressing and getting a 50 instead of a miss

3. I don't really mind this, getting high combo is rather satisfactory personally but other than that it's nothing too important as of now.

+. Blazier has been working on the fixed lane cover, not sure when's that going to happen but that eliminates the hidden issue

Rather neutral on the 2-combo idea but I can see where you're coming from, I don't really have a strong stand on this :p

can we get mirror mod first though

Please fix the LN Mechanics: 3rd Style!!
(please do it tho)

By the way even if the start and the release of an LN are taken into account when calculating accuracy, an LN still affects it only once. I'm not exactly sure of how osu!mania handles this but the start and the release of an LN need to be treated as single notes when calculating accuracy (if I get a MAX300 on the start and a 200 on the release, I should get a MAX300 and a 200, not a single 300).

I agree about the 2 combo LN. it really annoying though

Can you make it so once you vote you can't change your vote?

Also some food for thought here.

2 combo per LN if and only if press and release both awards score. Else only 1 combo.

PyaKura wrote:

Also some food for thought here.

Looks like even after 2 years the vast majority still agrees that 2 combo per LN is the best style. not surprised. why would it even ever work otherwise

Implementation:



Each LN during gameplay has an integer state with 3 levels: 0, 1, 2
0 - was never triggered yet
1 - was triggered and is being held
2 - was triggered and released
When it spawns, it starts with state 0.
If you press a key on the LN's column within judgement timing range from the LN's start time, and its state is 0, then:
- the state will become 1.
- you get a judgement notification with your skin (just like pressing a regular note)
- you get combo incremented +1
- judgement gets calculated into the total accuracy like with regular notes
If you release a key on the LN's column within judgement timing range from the LN's end time, and its state is 1, then:
- the state becomes 2.
- you get a judgement notification with your skin (just like pressing a regular note, except now you're releasing)
- you get combo incremented +1
- judgement gets calculated into the total accuracy like with regular notes
If you release a key on the LN's column, and its state is 1, but you're not within timing range for LN's intended release time, then:
- state becomes 2.
- you get a judgement notification with your skin for a miss.
- combo resets to zero.
- miss gets calculated into the total accuracy.
(since the LN is state 2 now, it will be impossible to re-press it anymore)
This is pretty much all the event listeners you need. Now for just main loop things:
As the chart goes on, if LN reaches a point where its start time + judgement timing range are "below" the current time, and the state is 0, that means the player failed to press this LN at all (pressed too late), so:
- state becomes 2.
- you get a judgement notification with your skin for a miss (and you get 2 misses from this, since you failed the start you won't be able to release either)
- combo resets to zero.
- miss gets calculated into the total accuracy.
(since the LN is state 2 now, it will be impossible to press it anymore)
If the LN reaches a point where its end time + judgement timing range are "below" the current time, and the state is 1, that means the player successfully pressed the LN, but forgot to release it in time, so:
- state becomes 2.
- you get a judgement notification with your skin for a miss.
- combo resets to zero.
- miss gets calculated into the total accuracy.
(the LN is always outside of the playfield at this point)

This pretty much covers all situations. Should be easy to implement. Just need one basic type and some conditional statements.

Wouldn't your suggestion change the score system a bit, Aqo?

Because, currently, not releasing a LN counts as a 200, not as a miss (although I would prefer it that way, personally).

Of course it would change the score. And for the better.

From what I've heard, the slider changes in o!standard are only going to apply to new maps. I don't see why we can't do exactly the same in mania.

‎

Aqo wrote:

Of course it would change the score. And for the better.

From what I've heard, the slider changes in o!standard are only going to apply to new maps. I don't see why we can't do exactly the same in mania.

I thought the idea was just to make a quick change that wouldn't change too much. But I agree, it would be way better this way.

I'm surprised that so many people want LNs scored as 2 separate hits. An LN would award twice as many points as a normal note. Maybe that aspect of it needs some work.

But everything else about this system is promising. I especially like that you'll actually get feedback for hitting the start of an LN, instead of waiting until the end to get a 200. It'll fix the problem right now we have where you don't know if the 200 was because of your press or release.

As a side note, implementing "score v2" for osu!standard was a very small change. Pretty much a couple lines of code. Implementing a proper "v2" for mania will require a lot more changes in all areas--LNs, combo, accuracy, hp/drain, perhaps even penalties for mashing--almost a full rewrite.

Blaizer wrote:

An LN would award twice as many points as a normal note. Maybe that aspect of it needs some work.

This is completely normal because an LN has two "interaction points", where you have to be accurate, and a normal note only has one.

Blaizer wrote:

As a side note, implementing "score v2" for osu!standard was a very small change. Pretty much a couple lines of code. Implementing a proper "v2" for mania will require a lot more changes in all areas--LNs, combo, accuracy, hp/drain, perhaps even penalties for mashing--almost a full rewrite.

osu!mania finally becoming a real vsrg??? i'm hyped for this!

same as IIDX CN to calculate amount of combo and make 2000 combos ez

option 3 sounds like a great idea; just one question with regards to it: if a player releases a LN too early to get a judgment, will that be considered a combo break? if yes, then 100% support - if not, then I don't support this personally

2. You're allowed to release and re-press an LN, something that is not allowed in any other scrolling rhythm game that I know of, and encourages mashing on LNs instead of trying to play them correctly when there are difficult LN patterns.

lol stepmania (also you can do this on PIU if the tick multiplier is low enough despite it too giving combo for holding freezes: there's a mission that does it pretty hilariously)

but yeah, this is a good idea if executed properly

EDIT: also, weight of a LN release in terms of score should be reduced, imo - maybe 0.5x of a normal tap?

I also believe that it should be 2 events. It's 2 actions of the finger, so it should definitely count both separately.

Halogen, IMO that would complicate scoring more than necessary. I think I understand why you suggested that, but I don't think it's a terribly good way to achieve that goal.

Blaizer wrote:

As a side note, implementing "score v2" for osu!standard was a very small change. Pretty much a couple lines of code. Implementing a proper "v2" for mania will require a lot more changes in all areas--LNs, combo, accuracy, hp/drain, perhaps even penalties for mashing--almost a full rewrite.

Give me the source code and I'll write it myself. Just finished writing an assembler in C over the last week, doubt this would take more than a couple hours, unless the existing code is a complete terrible mess (and even then rewriting all the logic from scratch shouldn't be hard as long as the groundwork of music being played while sprites are displayed and key listening is in place exists).

Anything different from the current LN combo mechanics would be a huge improvement.

I don't like the third option, though. It's the combination of holding and releasing that makes LN's hard, so why reward players for failing to do exactly this? Even if you're going to reward players for the initial hold, it's often easier to time the hold right than to time the release right. Making the release weigh heavier than the hold in score/combo could be a solution to this, but imo this just adds an unnecessary amount of complexity, which mania already has enough of.

On a similar note: what about really short LN's? The kind of LN's you press and release immediately, or a stream of fairly short single LN's (like in The Limit Does Not Exist's Insane diff). Rewarding players twice for those LN's seems a bit over the top.

I prefer giving 1 combo/judge/score for the entire LN. It's a simple system, rewards getting LN's right, punishes mediocrity, and it just werks.

The only problem I have with timed releases is when it's "timed" to a sound that gradually fades out and doesn't have a strong beat or sound. I agree with having it worth two judgements with the same weight on a short LN because part of the difficulty of it being short is actually releasing fast enough to be in time.

If any of the alternatives happen, could that lead to combo-based achievements (500/750/1000/2000 combo) being achievable through osu!mania? Because I like the sound of that...

Drojoke wrote:

On a similar note: what about really short LN's? The kind of LN's you press and release immediately, or a stream of fairly short single LN's (like in The Limit Does Not Exist's Insane diff). Rewarding players twice for those LN's seems a bit over the top.

Really short LN's still have some skill involved for when you release them, it's how we can differentiate short LN's from just normal notes. IMO anyway.

EDIT: I also want to point out that some osu!mania maps utilize LN releases, whether it's to represent a subtle sound that wouldn't be mapped to a normal note otherwise, or something else. So the release should still be accurate.

The poll says it all. Make it happen!

Aqo wrote:

As the chart goes on, if LN reaches a point where its start time + judgement timing range are "below" the current time, and the state is 0, that means the player failed to press this LN at all (pressed too late), so:
- state becomes 2.
- you get a judgement notification with your skin for a miss (and you get 2 misses from this, since you failed the start you won't be able to release either)
- combo resets to zero.
- miss gets calculated into the total accuracy.
(since the LN is state 2 now, it will be impossible to press it anymore)

I think that failing to press the LN start should give 1 miss, but still give a chance to play the release, that way, the judgments for the start and the release are as independent as possible (which is good for making a scoring system based on judgment counts).

So, what about this?

(ADD) If you press a key on the LN's column AFTER the judgment timing range from the LN's start time, and its state is 0, then:
- the state will become 1.

(MODIFY) As the chart goes on, if LN reaches a point where its start time + judgment timing range are "below" the current time, and the state is 0, that means the player failed to press this LN at all (pressed too late), so:
- state is still 0
- you get a judgement notification with your skin for a miss (you get 1 miss)
- combo resets to zero.
- miss gets calculated into the total accuracy.

(ADD) As the chart goes on, if LN reaches a point where its end time + judgement timing range are "below" the current time, and the state is 0:
- state becomes 2.
- you get a judgement notification with your skin for a miss (you get another miss)
- combo resets to zero.
- miss gets calculated into the total accuracy.
(since the LN is state 2 now, it will be impossible to press it anymore)


Also, currently, the game has a MISS timing window for early hits, how do you think those should be handled?
a) Removing that MISS timing window.
b) Handling that time window just like other timing windows, but instead of giving +1 combo, it resets combo.
c) Same as "b", but hitting a LN start within the MISS timing window sets the state to 2 instead of 1 (release becomes a MISS as well).
d) Same as "b", but hitting a LN start within the MISS timing window keeps the state as 0 (player has to re-press the key to release the LN)
e) Considering MISSES got this way the same as other judgments (giving +1 combo), combo breaks would only happen for notes that weren't pressed/released at all, or releases that are so early they are outside the MISS timing window.
f) etc...

Full Tablet wrote:

I think that failing to press the LN start should give 1 miss, but still give a chance to play the release, that way, the judgments for the start and the release are as independent as possible (which is good for making a scoring system based on judgment counts).

Most people would be against it, myself included.

Press and release and not entirely independent. Part of the difficulty is holding it all the way while playing the other patterns. It's a lot harder to play for example 12 or 67 trills while holding 3 or 5 compared to just playing 12/67 alone then tapping on 3/5 at the end to get the score for the release.

tl;dr Release shouldn't award points if not held all the way. You can only give half score for an LN by only-pressing, not by only-releasing.

Aqo wrote:

Full Tablet wrote:

I think that failing to press the LN start should give 1 miss, but still give a chance to play the release, that way, the judgments for the start and the release are as independent as possible (which is good for making a scoring system based on judgment counts).

Most people would be against it, myself included.

Press and release and not entirely independent. Part of the difficulty is holding it all the way while playing the other patterns. It's a lot harder to play for example 12 or 67 trills while holding 3 or 5 compared to just playing 12/67 alone then tapping on 3/5 at the end to get the score for the release.

tl;dr Release shouldn't award points if not held all the way. You can only give half score for an LN by only-pressing, not by only-releasing.

That method doesn't actually solve the issue of being able to cheese patterns like that completely.

With the requisite of pressing the start to being able to release the LN, you can just press the start, release the LN early and play the trills easily (you get 1 MISS as a sacrifice). With the system I proposed, additionally, you would be able to ignore the start, play the trills, and press at the last moment to be able to release (you also get 1 MISS as a sacrifice for the tactic).

Some other alternative is considering 3 independent judgments for the LN, the start, the hold, and the release (to not make LNs be worth so much compared to regular notes, each judgment could have a lower weight, for example, each one worth 2/3 of a 300g/300/etc...). The holding judgment would be based on the percentage of time the LN was held (out of the possible holding time) while the LN is in state 1.

With that system, using the first tactic (only pressing start) would give 4/3 Misses (and 2/3 of the pressing judgment), the second tactic (pressing at last moment to release) would also give 4/3 Misses (and 2/3 of the release judgment).

let me word it differently
no rhythm game at the moment has the release independent method, no one would support this, and imo it's just awkward and pointless

also it would imply you need to press again in order to release, meaning it literally rewards you for intentionally playing against the map. that's a big no-no

‎

Yes, this is a chance to look at it from a new angle, but you NEED to consider how other rhythm games work because we can clearly see the results in gameplay from the mechanics.

For example, in stepmania, there are a few elements that together cause LNs to be used in ways that only work because of the mechanics. For example, since you don't need to release them properly, you see cases where short LNs are used as visual accents, or cases where a 4 note chord has short LNs with the ends arranged in a staircase pattern, instead of ending at the same time. And because LNs are harder to read with larger skins, many players dislike LN heavy patterns in that game. On top of that LNs are seen as not being a particularly important element to have, at least in part because they don't penalize you for accidentally releasing them as long as you repress them soon after. Both the lack of requiring proper releases as well as the ability to repress LNs reduce the importance of LNs in stepmania.

If you look at o2jam instead, you see a different picture. You cannot repress an LN if you miss it, you must release at the correct time, and the penalty of missing an LN is very steep (I do believe that you lose more health for missing an LN, but if that's not the case I'd appreciate someone correcting me here). These mechanics lead to LNs having more emphasis due to the added difficulty of the stricter mechanics.

Of course, that's not the whole story, since in o2jam, the timing windows for accurately hitting notes are much larger, and that serves to reduce the overall difficulty of hitting a note and releasing it at the right time.

Osu!mania is currently somewhere in between, thanks to the mechanics of health gain while holding LNs combined with the relatively strict timing windows compared to o2jam.

Something I want to point out here is that game design is incredibly difficult, and it's often very hard to understand just what effect a mechanic will have when you are designing it. Just talking about things is unlikely to come up with a solution that will truly fix the problems people have with the current mechanics, since some of the people doing the talking don't seem to understand the possible effects of the mechanics they are suggesting be used.

Instead, What I'd like to see is some sort of testing where several options are rolled out on separate test versions of osu and the community can actually test out the different options and come back to discussions with a better understanding of the actual effects that the proposed mechanics have on gameplay overall. Of course, I highly doubt that will happen, but I think that would be a far better way to move things forward than how things have worked so far.

Bobbias wrote:

Instead, What I'd like to see is some sort of testing where several options are rolled out on separate test versions of osu and the community can actually test out the different options and come back to discussions with a better understanding of the actual effects that the proposed mechanics have on gameplay overall. Of course, I highly doubt that will happen, but I think that would be a far better way to move things forward than how things have worked so far.

This can be done when peppy publishes the game client's source code (this year maybe?). I'm totally willing to give this a shot, along with other idea's I've had like proper audio/visual calibration and disabling autoconverts. We'll just have to wait on peppy (peppy pls).

otwo!mania
2 years waiting for this and still counting..

How did the current system com into play anyway?
It seems so half-assed...

kiyoemon wrote:

How did the current system com into play anyway?
It seems so half-assed...

Weird quirks found in mania's LN's are also present in standard's sliders, so I'm guessing mania inherited these traits from osu!standard.

2 combo per LN makes most sense to me since 2 inputs = 2 combo.. It also could be 1 combo but with average from both pressing and releasing points

Drojoke wrote:

kiyoemon wrote:

How did the current system com into play anyway?
It seems so half-assed...

Weird quirks found in mania's LN's are also present in standard's sliders, so I'm guessing mania inherited these traits from osu!standard.

That's most likely where it came from. I'd like to remind everyone that Peppy was not the creator of osu!mania's code. Woc, a chinese guy who ended up leaving after some drama with Peppy was the actual creator, and as far as I understand, a lot of the code for mania borrowed bits and pieces from standard, and is a bit of a mess that needs to be cleaned up sometime in the future (and I do believe Peppy plans on reworking the code to mania sometime after Osu!Next happens).