osu!mania ScoreV2 live!

posted
Total Posts
476
show more
-Konner-
It's not drama. They want feedback and people are giving their feedback. Some people have been way too immature about how they've tried to make their position known but at least they are making people aware of their opinion as a whole. People should not join in with what those people have done but it still shows that people dislike the ideas.

If you don't give any feedback, nothing will change. If you don't give detailed feedback, no *specific* changes will be made. Just make sure to voice your opinions and ideas without filling the thread (or any other thread), with what is basically spam with a small point to make.
Hinpoppo
I don't really want to put the effort in right now, but I'm going to cast my vote in for not letting visual mods affect score for obvious reasons.
reyss

-Konner- wrote:

It's not drama. They want feedback and people are giving their feedback. Some people have been way too immature about how they've tried to make their position known but at least they are making people aware of their opinion as a whole. People should not join in with what those people have done but it still shows that people dislike the ideas.

If you don't give any feedback, nothing will change. If you don't give detailed feedback, no *specific* changes will be made. Just make sure to voice your opinions and ideas without filling the thread (or any other thread), with what is basically spam with a small point to make.
yeah i know, i'm not agree with this too except the LN.
Hotaru-
Visual Mod got multiplier score???
Oh man.. srsly visual mod just make player more confortable not make it hard
I said this coz I'm FL player and also HD player
f
Brownsville369
alright, I'm not an expert on any of this stuff, and I'm not THAT good at mania but here goes

Honestly visual mods for mania don't need their own score modifier, that's absurd they do nothing to increase the difficulty of the song and many people will use those mods to aid with reading so to unfairly add score just because of the way they read is ridiculous, though not nearly as much as this next point.

As for the combo addition to mania... that's absolutely asinine. I don't know where you got this idea for combo affecting score so greatly in the thing to presumably become the new default scoring system but you need to stop. Combo should not be a factor in score at all, let alone 20% the very idea of that is just ridiculous. It punishes missing spread out more than missing all at once, let's say you miss 4 times with everything else being Max 300, you have 5 short combos or 2 with one being long and the other being short. The 2 combos will be favored over the shorter one which is stupid because they both have the same accuracy which is the truest measure of skill of a player. If you believe that accuracy isn't the best measure of skill for a player, then just remove any semblance of "accuracy" from the game, and just make notes "hit" or "miss" without any variation and make the scoring purely combos.

I'm not trying to insult you or anything, but you're being incredibly stubborn on this issue when it's very very apparent that many people in the community don't like it and so I'm trying to make you see the flaws with it. I hope you reconsider this issue (mostly the 20% of score being combo) because if not and this becomes the new "norm" for scoring you're going to see a mass exodus of mania players to a game that isn't basing 20% of your score on combo alone. It's up to you what happens after this. No pressure.
PisRyder
First of all sorry for my bad English :3 Here are some of my opinions:
  1. LN: OMG YESS
  2. NF/EZ/HT: 0.5x is fine
  3. HR: 1.1-1.2x
  4. DT: "Make DT adjust to 100%/110%/.../150% with score bonus increments of 0.05x (or something like that)" -> Hell Yeah!!!
  5. HD,FL,FI: Like ALMOST people say, a big NO!
    People use these mods not to make the map harder, but easier to read. I used to play HD at some point and be like "Never gonna turn this sh*t off", but then a wild SV-heavy map appears and I be like "WTF is going on behind that covered space?" and switch back to nomod *the end*. These mods has it's advantages and disadvantages, the same applied to nomod so it's up to players preference to pick what they like best. If Standard has mouse/tablet/touchscreen players, then we Mania has HD,FL,FI,nomod players.
  6. Combo: I really hope that combo-based score nerver make it way to Mania *sigh*
    I quit Standard because combo-based score... When I look at the local ranking, I want to see the improvement, that means best play at top and worst play at bottom. But when I find my best play is at bottom and worst play at top, it's very anoying >.< I really like the current score system of Mania, 50% base score and 50% bonus score. The bonus score is some what "combo-based" but when I miss at half part of the song, it's not affect score as mush as scorev2. If it's necessary to make the upcoming MWC more interesting, I suggest when you get xcombo (ex:100) a portion of 20% combo-based score will be add to your current score, so the hardest part of the song (miss is somewhat unavoiable) still have impact on final score but not as heavy as current scorev2.

I really like Mania mode in osu!, but if scorev2 is necessary for the mode to move-on then I just have to adapt and find a way of dealing with it.
Thanks for reading!
Ciel
Time for my contribution to this.

Main sections organized by increasing order of importance. Read bolded sentences for tldr.

Mods


Vision Mods

These are HD, FI, and FL.
As someone who play with a HD skin (not using the hidden mod), and is completely unable to play nomod otherwise, I think if the HD mod still stays the same (it shouldn't), then the 1.06x multiplier should be fine. This is due to the fact that playing with a lane cover which effectively moves up and down is not particularly easy to play, and results in those intentional misses in order to sometimes bring the cover down. On the other hand, if the HD mod gets fixed to become static, then I believe there should be no multiplier. This is due to the fact that (with the exception of slowjams (not something to discount)), most high level players don't even look below a certain point on the screen when playing. This is why people play with a cover, in order to hide the visual clutter that lies below a certain point where people pay attention. While there may be some merit to giving a score multiplier simply for the ability to deal with slowjams, this is such a rare occurrence overall, that I don't think it really should warrant any bonus.

Misc.

The remaining mods are not really worth commenting about yet, as the score mutlipliers they provide is more relevant when talking about the pp system, which is not the main topic of concern at the moment. As a side note though, playing a song with DT is close to playing an entirely different song, which no amount of score buff could fix as a problem (it would be too little or too much). In addition, with HR, the timing window for a 300g is reduced, which impacts the score by a large amount. Therefore, that mod would probably also require a significant buff. Of course, for tournament play, none of the non-vision mods actually matter that much, so lets just ignore this for now.

Long Notes


LN End Window

After playing around with this a bit, I believe that this timing window is currently a little bit too small. Considering how it's inherently harder to time releases over holds anyways, I think this should be increased to something around a 2.0x multiplier instead. (Obviously, this would require much testing. A more statistical approach to do this would be to analyze replays that we already have, and the spread of the LN release timings compared to LN holds, given easier maps (since LN heavy maps would give somewhat inaccurate results)).

HP

An unintended side effect of this LN change is the fact that people can now lose HP for failing to release a note on time, something which did not happen before. In addition, there are now double the chances for people to miss an lose HP, which would result in maps getting harder this way. There probably needs to be a look into the side effects of this, especially for higher level maps, but this is also probably not as relevant for tournament play, except for possibly the final few rounds.

Score


I won't be doing as much number crunching for this part as some other have, mostly since I don't have the energy for it.

Combo

As Shoegazer has already pointed out in his post, mania is not really a gamemode where people single-mindedly focus on FCs, and thus, having this quadratic (it's quadratic guys) scoring system will not work in this game mode, unlike std and ctb. Instead, we tend to focus on aiming for higher accuracy instead, with overall scores being decent.

Players in this mode will (unless they are attempting to FC) not care about a random miss somewhere in the middle of the song, especially in a slow part. It's just a single miss, and won't significantly contribute to the final accuracy of the play. Instead, we only really care when we either miss a large number of notes in a burst (in which case we just flailed the burst), or when we are constantly missing throughout some section (lets say a stream of some sort). In order to fix this, there should be diminishing returns on each note at some point, rather than an quadratic increase in returns instead. This could take the form of a logarithmic curve, or there could be a system similar to taiko instead, where the bonus increases up until a certain cap, where it no longer grows.

On a side note, one of the things I like about the old "combo" scoring mechanism was the fact that you still lost combo for not hitting a 300. This continue to penalize people who are mashing through some section of the song, but can retain combo nonetheless (for example, jumptrilling rolls or jumpstreams a la PLANET//SHIPPER). I'm not sure if this should stay or not, but it is worth keeping in mind to some extent.

Accuracy

I like the direction that this is taking, where scores are really calculated exponentially off of accuracy instead of linearly. The exponent is a bit too extreme right now however, as the score dropoff is quite extreme for even taking a small hit in accuracy. People have already ran numbers on this, and this would also obviously need testing with combo as well in order to determine what a good balance point is.

Which leads me to my final point.

Combo and Accuracy Balance

See this screenshot?
This is the extreme example of someone who has terrible accuracy, and yet can somehow FC the map. Where would we put him in the grand scheme of "how good is this score?" This is where the balance between combo and accuracy must be considered, and only after finalizing the individual scoring metrics within each section. As of right now, I think the balance of the 2 sections is extremely off, especially with the current combo scoring system in play. However, since that is subject to change, I will refrain for saying much more for now.
anim girl
people must really like combos.


this is just my opinion i'm not very professional to things like this ><

the LN changes are great !
mania rate changes will be interesting too

but combo game and score multiply for visual mods ?
no please

some people play better with nomod, some people play better with hd, some people play better with fl
preference

As a nomod player myself it's unfair to think that mods like these would give score bonus
from my experiences it messed with my reading and destroyed accuracy, more 200s, almost 1:1 300 ratios
sure it is possible to adapt to it but it will be pain in the ass to !


Don't change the timing window of HR too please! it is really good to practice with the smaller timing window for people who want better accuracy
i agree on making the bonus much higher if ever, if you aim to value accuracy player

About combo game;

It seems to be too punishing as i saw it, scores such as 700k S, 570k S were seen, and even 100k-300k B score

i considered mania as chill game since combo doesn't as much, but with accuracy playing a bit more makes it better

for example someone plays a marathon map with 99% accuracy but shitmiss in different section of the map, it would end up not being a very good score i believe because he did not make good combo, i find it discouraging because he played very well accuracy but end up with score like 700-800k only

it has much to work on
W D Gaster
Score multipliers? Please no! Imo HD, FI and FL players dont want it! I am using HD only because I like it and its my preference. There is not actual reason to force it on other players.

Instead of adding multipliers you should rework Hidden. Static one, like that in Stepmania is way better and more comfortable to play. If someone wants to have bigger hidden - just give him option to make it bigger, but not smaller because it will be quite overpowered.

LN changes seems great, cool stuff for accuracy based players.
ikzune
vision mods
hd fl don't need a score multiplier i and many others have skins that have 50% lane cover the whole time which in turn helps us read better then without hd mod, some people are fl only players and read better this way, you would just be forcing your community to play with fl only if they want to compete.

Hold notes
Personally i am against more leniency, hold note patterns are just another technique players can choose to learn and play, there are people who start from stepmania and hate hold notes simply because their game didn't start with them and so they have to specifically improve in hold notes to reach a point where they can play ln as well as they can play fast patterns. Whilst people learning to acc hold notes may not be too much of an issue, hold note patterns have the potential to use more technique then single note patterns, i however lack the experience to say how much hold notes should be weighted compared to single notes in terms of a pp system as this would require judging how much technique is required for hold notes in any given map via a formula

1.1x rate at 0.05 value
Personally i would prefer a song to be rated based on how difficult the pp calculation judges a song at x speed however, there are certain patterns that can be inflated too much in value via rates, we have songs that do this already and for now this system is fine. My opinion is once, if we get a more accurate pp calculation formula in place then it would be a good time to have songs judged individually

Current issues in the pp formula
1. There are patterns at least in 4k that are used in several songs which are really simple for a player, essentially too little technique for the difficulty they are being judged at, the best example of this would be jumptrills being overrated, I don't like this type of pattern in the first place because they create hit and miss scenarios where players can often fluke scores and earn free pp.
2. index songs are underrated, index songs are streams of single note patterns, since they don't use 2 note patterns, they are seen as less dense by the formula and there for a lot easier then they deserve to be rated as. The best example in ranked 4k you can see of this is the empress scream off sc difficulty, there may be other difficulties from that map that are just as underrated however this is the main one i have experience with
3. Many people yell at mappers who get their maps ranked with this issue however, it is an issue with the pp system itself and not the mapper. There are several songs in 4k which will have 1 or 2 bursts in the song that inflate the difficulty and as long as someone can manage on that section, they can easily get a s rank and earn free pp. 2 recent examples of this are: triumph and regret, C18H27NO3 but there are many other songs that do the same thing and something needs to be done to the formula to at least reduce the effect a few bursts in a song can have.

tldr whilst it is good to see mania rewarding players who are more consistent, there are issues that are more significant within the current formula, i have only stated some issues that have been mentioned several times with 4k and i am sure there are plenty of examples with overrated and underrated patterns in other keymodes
Jinjin

aphixia wrote:

Hold notes
Personally i am against more leniency, hold note patterns are just another technique players can choose to learn and play, there are people who start from stepmania and hate hold notes simply because their game didn't start with them and so they have to specifically improve in hold notes to reach a point where they can play ln as well as they can play fast patterns. Whilst people learning to acc hold notes may not be too much of an issue, hold note patterns have the potential to use more technique then single note patterns, i however lack the experience to say how much hold notes should be weighted compared to single notes in terms of a pp system as this would require judging how much technique is required for hold notes in any given map via a formula
I can assure you that hold notes have become more difficult than before (especially in LN heavy charts). The new LN mechanic punishes players for not releasing the note at the end, and will cause a miss no matter what. Currently, if you hold down your key even when the long note ends, you get 200/100, but doing so with the new LN mechanic will cause a full miss. For example, I was getting close to 93% on sisters noise with the current LN mechanics, but I get less than 90% with the new LN changes. I think the 1.5x leniency is a good idea.
Shoegazer
Guess I'll post some formulas and some examples here - original post on my thoughts here. Ideas in this post take precedence - since I didn't really work with figures in the post before. The post is there just to give reasoning as to why the scoring system is not really ideal. Extremely long post ahead.

Ciel's post is also another post worth reading, and talks about the general ideas of visual mods and LNs as well.

Link to spreadsheet for reference of what the formulas I'm bringing up, and a more visual view: https://docs.google.com/spreadsheets/d/ ... sp=sharing
Feel free to make a copy of it if you want to fiddle around with it, too.

Accuracy
Use the score v2 (Accuracy) sheet for reference. The formula I'm suggesting is Accuracy^(1 + Accuracy * 4). The 3 columns on the right are meant to be a comparison - the formula on the right is Accuracy^5.

The main reservations I've had is that the exponential magnitude is too steep and that lower accuracy scores might be punished far too severely. The steps to remedy this, is to use a lower exponent, and to gradually lower the exponent as the accuracy gets lower and lower.

So why essentially Accuracy^5, since at very high accuracies the exponent adds up to 5?
This is meant to replicate benchmarks in score v1 (this is also why there is a 1mil score equivalent for reference). This isn't necessarily a 1 to 1 comparison, only 98% scores are about the same - the rest are slightly lower. I've checked around with multiple people and they seem pretty content with the numbers came out for the combo scoring aspect. I don't think there should be that much of a drastic change for things like this for scorev2 - I'd say the only reservation that I had with scorev1 is that there's too little differentiation between a 99% and 100%, but there's too much differentiation with a 98% and 99%. This mitigates that.

So why Accuracy^(1 + Accuracy * 4), then? Or particularly, why the (1 + Accuracy * 4)?
The general idea is that the higher the exponential, the steeper the curve. By lowering the exponential with a lower accuracy, it makes lower accuracy scores noticeably less of a dead-weight. You can technically do something like Accuracy^(Accuracy * 5) for what it's worth, it initially slipped my mind when I did this. The main problem with this however, is that the differentiation in exponent between accuracy might be too insignificant. I'm not sure how you can mitigate this, but there's most likely a way to do it. However, this isn't top priority, this is more of a plus, more than anything else.

The reason why I initially suggested (6 - Accuracy) at first was because I'm retarded and I didn't know how exponentials work at first, so please ignore what I said in that area.

So what are the drawbacks of this?
The main drawback that I can think of is that I'm using scorev1 figures to calculate accuracy. While this doesn't seem like that bad of an idea on paper, I'm not sure how the new LN mechanics will affect accuracy, especially for LN-heavy charts. This is something that needs experimenting, but I think the easy solution for this would be to ask feedback from the general community as to how much lenience does a LN release have. Many people have said that it's considerably harder to release LNs, which makes sense, but considering that it goes hand in hand with accuracy - it's very much something that needs to be under scrutiny. If you want players to be more stringent with their LN releases, you can lighten up the exponential a tiny bit. If you want the accuracy exponent to stay as it is, make LN releases more lenient.

Minor drawbacks include the fact that there's also the fact that the situational exponent might be too insignificant (which is mentioned earlier), and that I only looked at accuracies all the way down to 93.20%, which is generally about what the worst players in tournament semi-finals/finals would be getting on average in the first place. I assume that it wouldn't be very harsh with accuracy lower anyway, but you can always modify the situational exponent if that's not the case. The base exponential (which is 5 at 100%) is most likely fine.

Combo
Use the score v2 (Combo) sheet for reference. The formula I'm suggesting is HitValue * min(logx(combo), logx(combo cap)). The 3 columns on the left are meant to be a comparison - the formula on the right is HitValue * logx (combo).

This is a lot longer to explain (and probably a lot more complicated), but I'll try to explain it to the best of my ability. The reason why a non-FC hurts so much is because of the fact that the combo score as combo is higher is quadratic (thanks Ciel for the clarification). The most intuitive way of solving this would be to use a logarithmic scaling down for combo. I know there's probably going to create a decent amount of difficulties programming-wise (particularly trying to find the perfect score so you can scale scores down appropriately), as the combo mechanic is designed to produce exact figures - but do read on.

Interestingly enough, it doesn't matter what logarithm you use to scale down the combo, the end result will be the same. I plugged in a log of 4, 10, 500 and decimals, they all work - except for the ones that don't work normally (e.g. 1, 0, negative numbers). I don't really recall the reason for this exactly, though. If someone knows, let me know because I'm actually pretty curious about this and I'm kinda overwhelmed by all of this information to really find out by myself.

The scenario I've tried to emulate is a hypothetical situation when there are 2000 notes in a chart. I made 3 scenarios: wa person with 1,000 max combo (1 miss), 500 max combo (3 misses), 250 max combo (7 misses). These are hypothetical situations that emulate the most extreme cases in terms of miss location. Under these scenarios, given that you've hit full 300gs aside from misses, you would lose 10.47%, 20.91% and 31.30% of your potential combo score - about 21K, 42K and 62K respectively. Given, this is not entirely correct, since my calculations assume that the combos are exactly the same, but this is incorrect in reality since the miss also counts as a note and as a result it reduces the combo by 1. (e.g. 1000-999, rather than 1000-1000) This is however, insignificant, considering that the difference is >0.06%.

There's a couple of problems with this, the main problem being the magnitude of penalty. The fact that a person could lose (up to) 20K over a single miss is most likely overkill, considering that 20K can potentially be a game breaker, and almost certainly will be a game breaker if you're in very early stages. It still encourages a no room for error attitude, which just hasn't been adopted in osu!mania just yet.

Since you can't change the logarithm for this, you have to implement something else - a combo cap (which is done through something like min(logx(combo), logx(combo cap))). This means that at a certain point, the log(combo) component cannot go any higher, as log(combo cap) will be lower than log(combo), and the lower value will be taken. I used a combo cap of 400, which reduces the maximum losses to 3.42%, 10.26%, 21.70% (the format is 1 miss, 3 misses and 7 misses). Very noticeable drop for the 1 miss, and is arguably about right. Multiple misses are penalised quite appropriately as well, though it might arguably be too lenient, as this only highlights the extreme cases. You can increase the maximum penalty by increasing the combo cap - for example a combo cap of 500 increases the penalties to 4.16%, 12.47%, 23.97%. This is something that requires experimenting, and is probably one of the more important parts of the scoring formula to tinker around with.

You can also use a relative combo cap (and is probably better), rather than an absolute - I used 20% in my case, but 25% is probably fine as well. The relative combo cap can be truncated or rounded up, but the difference will be borderline negligible.

What are the drawbacks?
The figures used to show total losses are based on extreme cases - as a result, my suggestion for the combo cap might be a bit off. I don't know how the combo cap of 400 will be executed in practicality, but I don't expect the variance in performance to be that significant. Regardless of this, even if the variance does make the results look a bit unfavourable intersubjectively, this can be changed. This is something that requires community inquiry, more than anything else (perhaps it would be better to show combo score as if it's a 1mil equivalent?). You have to experiment around to see how the combo cap should be, but that's about it.

Location of misses still matter a decent amount, probably far more than what the community wants. This doesn't negate the problem entirely, but it does to some extent. Basically, the lower the combo cap, the less the location of misses matter, unless you're missing a lot in a concentrated area.

Another problem might be the fact that 300gs might not be weighted very heavily, since only the combo component of scorev2 looks at 300gs. The only two solutions that come to mind would to either increase the proportion of combo to emphasise more on 300gs, or to embed 300gs into accuracy (since as they are, they are weighted the same as normal 300s). The latter requires a lot more tinkering and probably creates more of a community uproar, so I think the former would be a better approach. Again, another thing that requires community response.

Lastly, there's the rounding problem, which I'm not sure how to do because I don't do programming aside from Python wankery. I assume that an extremely small logarithm should solve the problem - since it would eliminate the problem of rounding.

That should be all. Again, I'm not going to talk about visual mods and LNs, since Ciel (and many others) are more well-versed in those than I am. Hope this is a more tangible way of reshaping the scorev2 system - because I think it has massive potential to be a good scoring system that players widely agree on. If there's any questions/clarifications on what I mean or what each part of the spreadsheet does, I can respond to them.

For anyone who scrolled all the way down to read this, the main takeaways are that LN releases might be a touch too strict so it would be good for players to experiment and find out whether or not it's too stringent and to focus more on the formulas rather than blaming it on the name of the component (e.g. combo) and give constructive feedback as to how the scorev2 system can be improved.
Akary
rip mania
ikzune
i messed up and there were several factors i didn't consider especially when i was focused on ppv3 not scorev2 >.<

The one thing in terms of hidden,fl that i failed to consider was svs. I personally wouldn't give a 1.06 for this, my thoughts are that an experienced player with their vision mod will be able to manage svs and the drop will be much smaller then the bonus provides, from my experience i would be placing a heavy sv song at 1.02 or 1.03 at the most. One case this may differ is teleport notes in which there may be instances where they are too difficult for fl players to manage.

As for the other factors, i am not the best person to judge this, its just going to take a lot of playing around with to find a good balance, good luck ^^
Franc[e]sco
I completely suppport this and I feel like mania needed a scoring system like this for a LONG time.

>Score is made up of 20% combo and 80% accuracy.
The fully accuracy-based scoring was a huge problem with maps that had diff spikes, so this should fix the problem. On fully accuracy based scoring you could get free pp off maps that had a short difficulty spike even if you couldn't play the hard part at all.

>LN starts and ends are now judged separately.
>LN ends are given a 1.5x lenience to the hit windows.
>If an LN is broken but re-pressed, the LN end will not award more than 50 points.
>Works similar to the current system depending on when you release the hold, but is lenient enough to feel rewarding even for newer players (consider that ScoreV2 will be used as the normal ranking in the future).
Awww yes LNs were my major problem with mania coming from stepmania. LN releases were way too strict, to the point that it wasn't even fun. And it's not even about being accurate, since lots of maps don't even map LN releases to beats.

>Mods are back! NF/EZ/HT give 0.5x score multipliers and DT/HR/HD/FI/FL give 1.06x score multipliers.
I think DT should at least give a higher multiplier than the other mods, but other than that it's all good!
Bubbler
So far, considering many suggestions on the score system (mere complaints do not count here guys...) Shoegazer's approach seems most attractive to me - I like the variation of accuracy^5 thing and log scaling for combos.

And more math ahead (Please skip this if you're not so math tolerant):
I found out that picking 1 as the minimum power on accuracy was actually a lucky choice, since too low minimum power actually makes the function slightly decreasing around accuracy = 0.2!

Accuracy ^ (1 + 4 * Accuracy)


Accuracy ^ (0.5 + 4.5 * Accuracy)


This happens when the minimum power falls below 0.596 = 5 / (e^2+1).
Maybe not too much of concern since playing any map with 20% accuracy will most likely result in a fail :D

Shoegazer wrote:

Interestingly enough, it doesn't matter what logarithm you use to scale down the combo, the end result will be the same. I plugged in a log of 4, 10, 500 and decimals, they all work - except for the ones that don't work normally (e.g. 1, 0, negative numbers). I don't really recall the reason for this exactly, though. If someone knows, let me know because I'm actually pretty curious about this and I'm kinda overwhelmed by all of this information to really find out by myself.
This is because log(a)x = log(a)b * log(b)x, in other words, picking a different base is equivalent to scaling all values by a constant. And when the maximum total value must be scaled into some constant (in this case, 200k), constant scaling obviously has no effect.
trexex55

KatayokuNoTori wrote:

nerf ht pp gain :(((
support
tkk
great moves loctav, keep it up, proud of you!
Shoegazer

Shoegazer wrote:

Guess I'll post some formulas and some examples here - original post on my thoughts here. Ideas in this post take precedence - since I didn't really work with figures in the post before. The post is there just to give reasoning as to why the scoring system is not really ideal. Extremely long post ahead.

Ciel's post is also another post worth reading, and talks about the general ideas of visual mods and LNs as well.

Link to spreadsheet for reference of what the formulas I'm bringing up, and a more visual view: https://docs.google.com/spreadsheets/d/ ... sp=sharing
Feel free to make a copy of it if you want to fiddle around with it, too.

Accuracy
Use the score v2 (Accuracy) sheet for reference. The formula I'm suggesting is Accuracy^(1 + Accuracy * 4). The 3 columns on the right are meant to be a comparison - the formula on the right is Accuracy^5.

The main reservations I've had is that the exponential magnitude is too steep and that lower accuracy scores might be punished far too severely. The steps to remedy this, is to use a lower exponent, and to gradually lower the exponent as the accuracy gets lower and lower.

So why essentially Accuracy^5, since at very high accuracies the exponent adds up to 5?
This is meant to replicate benchmarks in score v1 (this is also why there is a 1mil score equivalent for reference). This isn't necessarily a 1 to 1 comparison, only 98% scores are about the same - the rest are slightly lower. I've checked around with multiple people and they seem pretty content with the numbers came out for the combo scoring aspect. I don't think there should be that much of a drastic change for things like this for scorev2 - I'd say the only reservation that I had with scorev1 is that there's too little differentiation between a 99% and 100%, but there's too much differentiation with a 98% and 99%. This mitigates that.

So why Accuracy^(1 + Accuracy * 4), then? Or particularly, why the (1 + Accuracy * 4)?
The general idea is that the higher the exponential, the steeper the curve. By lowering the exponential with a lower accuracy, it makes lower accuracy scores noticeably less of a dead-weight. You can technically do something like Accuracy^(Accuracy * 5) for what it's worth, it initially slipped my mind when I did this. The main problem with this however, is that the differentiation in exponent between accuracy might be too insignificant. I'm not sure how you can mitigate this, but there's most likely a way to do it. However, this isn't top priority, this is more of a plus, more than anything else.

The reason why I initially suggested (6 - Accuracy) at first was because I'm retarded and I didn't know how exponentials work at first, so please ignore what I said in that area.

So what are the drawbacks of this?
The main drawback that I can think of is that I'm using scorev1 figures to calculate accuracy. While this doesn't seem like that bad of an idea on paper, I'm not sure how the new LN mechanics will affect accuracy, especially for LN-heavy charts. This is something that needs experimenting, but I think the easy solution for this would be to ask feedback from the general community as to how much lenience does a LN release have. Many people have said that it's considerably harder to release LNs, which makes sense, but considering that it goes hand in hand with accuracy - it's very much something that needs to be under scrutiny. If you want players to be more stringent with their LN releases, you can lighten up the exponential a tiny bit. If you want the accuracy exponent to stay as it is, make LN releases more lenient.

Minor drawbacks include the fact that there's also the fact that the situational exponent might be too insignificant (which is mentioned earlier), and that I only looked at accuracies all the way down to 93.20%, which is generally about what the worst players in tournament semi-finals/finals would be getting on average in the first place. I assume that it wouldn't be very harsh with accuracy lower anyway, but you can always modify the situational exponent if that's not the case. The base exponential (which is 5 at 100%) is most likely fine.

Combo
Use the score v2 (Combo) sheet for reference. The formula I'm suggesting is HitValue * min(logx(combo), logx(combo cap)). The 3 columns on the left are meant to be a comparison - the formula on the right is HitValue * logx (combo).

This is a lot longer to explain (and probably a lot more complicated), but I'll try to explain it to the best of my ability. The reason why a non-FC hurts so much is because of the fact that the combo score as combo is higher is quadratic (thanks Ciel for the clarification). The most intuitive way of solving this would be to use a logarithmic scaling down for combo. I know there's probably going to create a decent amount of difficulties programming-wise (particularly trying to find the perfect score so you can scale scores down appropriately), as the combo mechanic is designed to produce exact figures - but do read on.

Interestingly enough, it doesn't matter what logarithm you use to scale down the combo, the end result will be the same. I plugged in a log of 4, 10, 500 and decimals, they all work - except for the ones that don't work normally (e.g. 1, 0, negative numbers). I don't really recall the reason for this exactly, though. If someone knows, let me know because I'm actually pretty curious about this and I'm kinda overwhelmed by all of this information to really find out by myself.

The scenario I've tried to emulate is a hypothetical situation when there are 2000 notes in a chart. I made 3 scenarios: wa person with 1,000 max combo (1 miss), 500 max combo (3 misses), 250 max combo (7 misses). These are hypothetical situations that emulate the most extreme cases in terms of miss location. Under these scenarios, given that you've hit full 300gs aside from misses, you would lose 10.47%, 20.91% and 31.30% of your potential combo score - about 21K, 42K and 62K respectively. Given, this is not entirely correct, since my calculations assume that the combos are exactly the same, but this is incorrect in reality since the miss also counts as a note and as a result it reduces the combo by 1. (e.g. 1000-999, rather than 1000-1000) This is however, insignificant, considering that the difference is >0.06%.

There's a couple of problems with this, the main problem being the magnitude of penalty. The fact that a person could lose (up to) 20K over a single miss is most likely overkill, considering that 20K can potentially be a game breaker, and almost certainly will be a game breaker if you're in very early stages. It still encourages a no room for error attitude, which just hasn't been adopted in osu!mania just yet.

Since you can't change the logarithm for this, you have to implement something else - a combo cap (which is done through something like min(logx(combo), logx(combo cap))). This means that at a certain point, the log(combo) component cannot go any higher, as log(combo cap) will be lower than log(combo), and the lower value will be taken. I used a combo cap of 400, which reduces the maximum losses to 3.42%, 10.26%, 21.70% (the format is 1 miss, 3 misses and 7 misses). Very noticeable drop for the 1 miss, and is arguably about right. Multiple misses are penalised quite appropriately as well, though it might arguably be too lenient, as this only highlights the extreme cases. You can increase the maximum penalty by increasing the combo cap - for example a combo cap of 500 increases the penalties to 4.16%, 12.47%, 23.97%. This is something that requires experimenting, and is probably one of the more important parts of the scoring formula to tinker around with.

You can also use a relative combo cap (and is probably better), rather than an absolute - I used 20% in my case, but 25% is probably fine as well. The relative combo cap can be truncated or rounded up, but the difference will be borderline negligible.

What are the drawbacks?
The figures used to show total losses are based on extreme cases - as a result, my suggestion for the combo cap might be a bit off. I don't know how the combo cap of 400 will be executed in practicality, but I don't expect the variance in performance to be that significant. Regardless of this, even if the variance does make the results look a bit unfavourable intersubjectively, this can be changed. This is something that requires community inquiry, more than anything else (perhaps it would be better to show combo score as if it's a 1mil equivalent?). You have to experiment around to see how the combo cap should be, but that's about it.

Location of misses still matter a decent amount, probably far more than what the community wants. This doesn't negate the problem entirely, but it does to some extent. Basically, the lower the combo cap, the less the location of misses matter, unless you're missing a lot in a concentrated area.

Another problem might be the fact that 300gs might not be weighted very heavily, since only the combo component of scorev2 looks at 300gs. The only two solutions that come to mind would to either increase the proportion of combo to emphasise more on 300gs, or to embed 300gs into accuracy (since as they are, they are weighted the same as normal 300s). The latter requires a lot more tinkering and probably creates more of a community uproar, so I think the former would be a better approach. Again, another thing that requires community response.

Lastly, there's the rounding problem, which I'm not sure how to do because I don't do programming aside from Python wankery. I assume that an extremely small logarithm should solve the problem - since it would eliminate the problem of rounding.

That should be all. Again, I'm not going to talk about visual mods and LNs, since Ciel (and many others) are more well-versed in those than I am. Hope this is a more tangible way of reshaping the scorev2 system - because I think it has massive potential to be a good scoring system that players widely agree on. If there's any questions/clarifications on what I mean or what each part of the spreadsheet does, I can respond to them.

For anyone who scrolled all the way down to read this, the main takeaways are that LN releases might be a touch too strict so it would be good for players to experiment and find out whether or not it's too stringent and to focus more on the formulas rather than blaming it on the name of the component (e.g. combo) and give constructive feedback as to how the scorev2 system can be improved.
Bottom-paged, so reposting this again.

Bubbler wrote:

Accuracy ^ (0.5 + 4.5 * Accuracy)


This happens when the minimum power falls below 0.596 = 5 / (e^2+1).
Maybe not too much of concern since playing any map with 20% accuracy will most likely result in a fail :D

Shoegazer wrote:

Interestingly enough, it doesn't matter what logarithm you use to scale down the combo, the end result will be the same. I plugged in a log of 4, 10, 500 and decimals, they all work - except for the ones that don't work normally (e.g. 1, 0, negative numbers). I don't really recall the reason for this exactly, though. If someone knows, let me know because I'm actually pretty curious about this and I'm kinda overwhelmed by all of this information to really find out by myself.
This is because log(a)x = log(a)b * log(b)x, in other words, picking a different base is equivalent to scaling all values by a constant. And when the maximum total value must be scaled into some constant (in this case, 200k), constant scaling obviously has no effect.
Thanks for the information! The latter snippet makes a good amount of sense to me.
Sakura Kyoko
Just throwing my opinion in here;

1. Vision Mods
Now, I think vision mods shouldn't give any bonus at all. They are mostly personal preference, there's more than a bunch of players that couldn't even pass things WITHOUT Hidden, so I don't see why that should give bonus points. One might argue, that this applies to standard aswell, but I don't think Mania and Standard are comparable in that aspect.

2. DT and NC

1.06x is WAY too little of a score-boost. Example; a regular 4* song that you might have 950k score on will easily become a 5,5* song that you might barely pass or get like 700k on (especially because the starrating is broken and DT makes for some really stupid starratings, but thats something for a different thread). How is it fair that you then get like what, 1.06x the 700k but no other benefits? Except for the best of the best players, that can play songs on DT with almost the same scores, this benefits literally noone. Especially if it won't give you bonus PP for the increased starrating.

3. The new LN Mechanic

I really think this is quite nice. I like the dynamic of the new LNs, and I also think that giving 1 combo per LN is a good move. Overall, I'm really happy with the new LNs.

4. Score V2
Don't even get me started. So a song I previously had 992k on now gives me 900k, well okay then.
You LOSE points for poor accuracy. THAT is by far the worst idea I've seen in ages. Why would I lose points that I've gained for doing something well? Like, I'm already losing out on the points that I lose by having poor accuracy/misses. Why would I then lose points for stuff I had on 300S before? Dafuq is this?

It's like "yeah look, you had 10 of the 20 questions correct, but instead of giving you 50% of the points, we're giving you none, because fuck you!"

The new system ONLY benefits top-tier players, and even those will struggle to keep their scores where they currently are, except for accuracy-machines like Halogen-.
And the 200k for having a full combo, while only getting 25% of it for A SINGLE MISS OR COMBO BREAK is incredibly counter-intuitive aswell.
Imagine, 100% the entire song, great 300S-rate, boom, you miss the last note. Well good bye PP, because you just lost 150k for a single miss.

I think our current score system isn't even that bad and most people didn't complain about it either. Why would you change a running system? There are MUCH bigger problems that need to be fixed, like that incredibly inaccurate star-rating algorithm.


5. Conclusion
While I really appreciate the time, effort and the attention mania FINALLY gets, I really think we're heading the completely wrong way. Instead of looking what other games do well (SM with rates, o2JAM LN mechanics), we desperately try to do stuff "our own way", which ends up hurting Mania more than it helps. Give a fair PP bonus for DT or give none at all, don't give bonus points for visual preference and for the love of god, don't subtract points for poor accuracy.

Sometimes, it's better to look at other, long-standing rhythm games and copy what people like about them. They're not popular for no reason. If we keep heading that way (that way being Score V2, 1.06x score multiplier for DT and visuals, lol), mania will never be taken seriously.

P.S.: I really love osu!mania, I'm not hating on the game in any shape or form, but I'm loving it how it CURRENTLY is. And I still think we should just copy things that other games (that were created long before osu! even existed) do way better.

EDIT: After I've been corrected and many of my missconceptions have been clarified, I want to excuse myself if I sounded too harsh here. As mentioned, I appreciate the time and effort the staff is putting into osu!mania this year, I've just been wrong with some assumptions and misinformed by fellow players. Sorry!
show more
Please sign in to reply.

New reply