Comparing mania to other VSRG's is a lost cause because it is a subjective matter. Players are going to cause backlash bacause it's aiming to be different than the standards those set. That's all can be stated and is not even a good arguement against scoreV2. I suggest you guys argue against scoreV2 objectively and explain what flaws it has within results and/or calculation.

abraker, here's something I came up with that could plug in to your diffscore[t], with t in my case being an actual timing/score object.

An individual hit difficulty would be based on two components, timing and physical difficulty.

1. Timing Difficulty:
   TD is a pseudo-reading difficulty calculation that looks at the note that came directly before the current note. If your current note is part of a double/triple/etc. the other notes are ignored and calculated separately. The shorter the time between the two notes, the higher the TD.
   TD only cares about timing points so if there was a double/triple/etc. the TD is not doubled/tripled/etc.
   
2. Physical Difficulty:
   PD compares the current note to the previous note on the same column. PD increases as the time between the two notes decreases. PD's purpose is to make a distinction between grace notes and minijacks, for example.

PD and TD would be combined at an appropriate ratio to give a final difficulty score. The ratio can easily be tweaked of course to find the best combination.

You might think that the PD calculation could be replaced with just checking columns on TD and multiplying accordingly, but PD has a few more subtleties which makes it better as a separate component. The following three examples show how the two components can vary:


As for how PD and TD scale with the time differences, I expect there to be a timing cuttoff where the PD or TD component defaults to just the minimum value. I'm thinking maybe a linear scaling for TD (as this is just reading based), and some form of normal distribution curve for PD.

EDIT: to nerf minijacks maybe PD could use the hit300 timing window in some way to accomodate rushing. IDK though, just a thought.

OppaiDefender wrote:

...

Interesting ideas, but describing how diffScore[t] works would deal more closely with star rating than score, so I am not so sure this is approprate for this thread.

abraker wrote:

Interesting ideas, but describing how diffScore[t] works would deal more closely with star rating than score, so I am not so sure this is approprate for this thread.

Yeah, I'm aware it's sliding into the SR domain a bit, but I believe it can still apply to scoring directly. I'm not sure how SR is calculated currently and grabbing the difficulty at a timing point t might not be accomodated for.

abraker wrote:

Comparing mania to other VSRG's is a lost cause because it is a subjective matter. Players are going to cause backlash bacause it's aiming to be different than the standards those set. That's all can be stated and is not even a good arguement against scoreV2. I suggest you guys argue against scoreV2 objectively and explain what flaws it has within results and/or calculation.

Score v2 = more combo based than score v1, yes?
Combo based = encourages spamming restart on the first miss instead of playing songs through, adds a lot of needless frustration, anguish and stress in a game that is supposed to be fun.
The people who want a combo based system exist most likely because of osu! standard, because they can't fathom the possibility of a game where a miss isn't a play ruiner.

Or perhaps you have people like smoogipooo who just want to make a system that's more entertaining to watch for MWC; to make scores more excitingly different from each other. In that case, why not instead add harder songs in the map pool? Is it because the same people would win consistently every time because they're better instead of randoms winning because they got lucky and didn't combo break that time? Yeah well, unfortunately that's the nature of rhythm games and better players are supposed to win. Maybe stop separating people per country for a change and together with adding harder maps in the pool you might get more exciting matches. But rhythm games just aren't as exciting to watch as other e-sports, and nothing will change that, specifically because skill differences in players are so well defined here, it's much more objective and straightforward.

Sure, I'll admit that a combo based scoring system makes tournament games more exciting to watch, but the trade-off which is making the game much more frustrating to play for the average user isn't worth it.

Sorry if this comes off as a rant, I'm just trying to give my opinion as constructively as I can. I'm not against anything in score v2 except making it more combo based.

robby250 wrote:

The people who want a combo based system exist most likely because of osu! standard, because they can't fathom the possibility of a game where a miss isn't a play ruiner.

I'm against a combo based system for a mania type game but I feel it's appropriate for standard because I'm capable of separating these two games into categories that have nothing to do with each other.


2016-06-20 14:38 Khelly: Hey rate my combo based scoring idea
2016-06-20 14:38 Lampranthus: yeah
2016-06-20 14:38 Khelly: Combo is worth 100,000/1,000,000 points
2016-06-20 14:38 Khelly: 25% of a map's total combo
2016-06-20 14:38 Khelly: will give you all 100k points
2016-06-20 14:40 Khelly: Is that good or bad
2016-06-20 14:45 Lampranthus: I think it's pretty good
2016-06-20 14:45 Lampranthus: Again, if you back it up with making your game's content able to fit those standard there's nothing you can do really wrong
2016-06-20 14:45 Lampranthus: but speed players will leave, and MA players from Stepmaina and LR2 will flock in

@robby: Combobased scoring doesn't make tournaments more exciting to watch as the winner can already be decided halfway through the map which is lame. Don't give combo any pro-arguments it doesn't have.

Endaris wrote:

@robby: Combobased scoring doesn't make tournaments more exciting to watch as the winner can already be decided halfway through the map which is lame. Don't give combo any pro-arguments it doesn't have.

I remember in February 2015 when Rrtyui S'd image material about a third of the way through the map everyone just focused on him and the other team members and opponents were irrelevant.

But I mean then again it was the first S of a famous then un-fc'd map.

He would've got that spotlight in any other mode too. I think in TWC 2015 tasuke was FCing that dragonforce tiebreaker and he got all the spotlight too. And in both cases it was a "pseudo"-tiebreaker that was played after the match had already been decided.~

Khelly wrote:

I'm against a combo based system for a mania type game but I feel it's appropriate for standard because I'm capable of separating these two games into categories that have nothing to do with each other.

Sure, standard has the aiming aspect to it, you can't just mash through the hard parts and get lucky. You can sometimes emphasize on not dropping combo at the expense of losing a bit of accuracy, but it's nowhere near the same thing as mashing in mania. Even then you could argue that it's way too combo based, in no dimension should an 80% score be capable of being rated higher than a 99% score.

Endaris wrote:

@robby: Combobased scoring doesn't make tournaments more exciting to watch as the winner can already be decided halfway through the map which is lame. Don't give combo any pro-arguments it doesn't have.

It's more exciting in the sense that it adds an RNG aspect to the game, and it's not the same winners every single time. But that randomness is also what makes it bad.

robby250 wrote:

Or perhaps you have people like smoogipooo who just want to make a system that's more entertaining to watch for MWC; to make scores more excitingly different from each other. In that case, why not instead add harder songs in the map pool? Is it because the same people would win consistently every time because they're better instead of randoms winning because they got lucky and didn't combo break that time? Yeah well, unfortunately that's the nature of rhythm games and better players are supposed to win. Maybe stop separating people per country for a change and together with adding harder maps in the pool you might get more exciting matches. But rhythm games just aren't as exciting to watch as other e-sports, and nothing will change that, specifically because skill differences in players are so well defined here, it's much more objective and straightforward.

Oh so much this. smoogipoo want to add a bit of dice into the game by potentially magnifying any mistake the player makes. The maps they choose are pretty uniform in difficulty generally speaking, so it it shouldnt be surprising if the match was decided from the beginning. If you want more interesting matches, pick more interesting maps. Maps which are long and have difficulty spikes layed throughout like land mines should make it interesting.

robby250 wrote:

abraker wrote:

Comparing mania to other VSRG's is a lost cause because it is a subjective matter. Players are going to cause backlash bacause it's aiming to be different than the standards those set. That's all can be stated and is not even a good arguement against scoreV2. I suggest you guys argue against scoreV2 objectively and explain what flaws it has within results and/or calculation.

Score v2 = more combo based than score v1, yes?
Combo based = encourages spamming restart on the first miss instead of playing songs through, adds a lot of needless frustration, anguish and stress in a game that is supposed to be fun.
The people who want a combo based system exist most likely because of osu! standard, because they can't fathom the possibility of a game where a miss isn't a play ruiner.

Or perhaps you have people like smoogipooo who just want to make a system that's more entertaining to watch for MWC; to make scores more excitingly different from each other. In that case, why not instead add harder songs in the map pool? Is it because the same people would win consistently every time because they're better instead of randoms winning because they got lucky and didn't combo break that time? Yeah well, unfortunately that's the nature of rhythm games and better players are supposed to win. Maybe stop separating people per country for a change and together with adding harder maps in the pool you might get more exciting matches. But rhythm games just aren't as exciting to watch as other e-sports, and nothing will change that, specifically because skill differences in players are so well defined here, it's much more objective and straightforward.

Sure, I'll admit that a combo based scoring system makes tournament games more exciting to watch, but the trade-off which is making the game much more frustrating to play for the average user isn't worth it.

Sorry if this comes off as a rant, I'm just trying to give my opinion as constructively as I can. I'm not against anything in score v2 except making it more combo based.

This guy.

I am having a lot of issue understanding this "RNG" concept that you guys are talking about. There's been plenty of instances of the game pulling RNG-like behavior on scores.

This one is beyond me. You'd need a PhD level dissertation to properly justify that one (hint: you can't).


Almost as bad, but still telling.


etc.


... it doesn't take long to see this. Anyone saying that the current score system is in any way or shape better than the proposed V2 without any actual data is out of their mind and is just spewing at this point.

@Halogen I'm not defending the current system. I'm simply against the combo component of score v2, which would enforce situations similar to the ones you've listed. For your examples, I'm guessing the distribution of misses/combo breaks is what made the difference in score.

Here's my proposed changes for score v2 specifically:

- Make regular 300s give 95% accuracy and make rainbow 300 scale with OD.
- Replace the fixed combo cap of 400 with 5%-10% of the max combo of the map, using the same algorithm as score v2.
- LN starts and ends should be separated like score v2 does, however LN ends shouldn't add combo (I don't mind if they do though) and they should be made more lenient instead of tighter (I don't really care either way, I just think that making them tighter would make even more people hate LNs and it would be harder to integrate LNs into a map without lowering OD).
- Increase HP drain and make a better anti-mashing system (discussing specifics is futile at this point)
- Add rates, give each mod/rate a multiplier that doesn't make it too difficult to beat your earlier score as long as you master the map relatively well.

The changes I've mentioned are very rough and could be interpreted for better or worse.

I always felt regular 300's should give 99% and rainbow's should give 100% because when it comes to multiple SS's or high acc scores you basically have to look at the score number or the actual # of rainbows and 300's each player got in order to compare instead of a quick % glance. Besides, rainbow's don't change from 16.5 ms no matter the OD.

gintoki147 wrote:

Khelly wrote:

I like the osu client and think a lot of the other games look and feel like complete shit to me

thank you so much
as someone who started playing VSRGs two years ago and tried many different games, those are exactly my thoughts of o2jam/stepmania lmao

Ever since stepmania introduced scripting as a part of skinning, theming, the diversity of stepmania themes skyrocketed.

Khelly wrote:

I'm against a combo based system for a mania type game but I feel it's appropriate for standard because I'm capable of separating these two games into categories that have nothing to do with each other.


2016-06-20 14:38 Khelly: Hey rate my combo based scoring idea
2016-06-20 14:38 Lampranthus: yeah
2016-06-20 14:38 Khelly: Combo is worth 100,000/1,000,000 points
2016-06-20 14:38 Khelly: 25% of a map's total combo
2016-06-20 14:38 Khelly: will give you all 100k points
2016-06-20 14:40 Khelly: Is that good or bad
2016-06-20 14:45 Lampranthus: I think it's pretty good
2016-06-20 14:45 Lampranthus: Again, if you back it up with making your game's content able to fit those standard there's nothing you can do really wrong
2016-06-20 14:45 Lampranthus: but speed players will leave, and MA players from Stepmaina and LR2 will flock in

cringe

Basically the point (indirectly but in relevancy to the topic at hand) of my post beforehand was that you can't really make a game based off of consistency and combo when your game mostly consists of charting styles based off of other scoring systems that don't include these systems.

Mini-LNs in stepmania are completely justified and are actually welcomed as a booster for your DP; the equivalent of raw score/%accuracy in O!M. In this game, it is the contrary; many conversions such as Brute Force have watered down the amount of mini-LNs used for what I assume is this reason (I don't want to put words into the "mouth" of the person who converted them).

I truly think that if Osu!Mania was a little more strict with creating their own original content while also mass producing it. You would also have to unrank stuff like haelequin's bullshit ending to fully justify these changes without making it a half-assed game like it currently is. (Unranked stuff doesn't really matter in scoring that much anyway aside from in tourneys which only affects the minority of players)

toolazytowriteanymore so

tl;dr Anything can work as a solid game as long as the work is put in to actually make it a fully-fleshed out thing. It may not appeal to all audiences of course; being a player who doesn't play for consistency I don't think it would much appeal for me, but it would definitely make for interesting content. Pretty much, if you're going to make a game that breaks the mold of current popular VSRG, you had better make the innovations elsewhere than in just the scoring system.








really lazy sry

*Also I should probably say that I understand that this is actually only going to be in MWC for now, just saying stuff

It's only sort of related to scorev2, but I wanted to make a small point about scoring and mapping style.

See, when osu!mania was released, I saw it as a possible way to create a place where every kind of chart/map/file/whatever is acceptable. Not just another game with it's own exclusive community that comes up with their own style and ignores everything every other game is doing. The fact that it supports a wide number of key modes just added to this sense of inclusivity. I'm all for osu!mania finding it's own style, but I'd rather not have a scoring system that is designed around that single style. Nor do I want to see ranked maps all conform to a singular style.

Here's a suggestion for accuracy score part. I tried to keep both essential score bonus for each additional percent above 95% and lenient enough loss for low acc players to give their teams a chance.


Also, i support Shoegazer's combo suggestion. In his system, one miss at 1000 combo map means loss of ~10k. In regular score system, this number is almost the same, lol. Shoegazer's suggestion would make score v2 not combo-based, but combo-adjusted. I didn't find any plots, so i've made one:


I hope, this will help people to understand advantages of Shoegazer's combo formula.

Kivicat wrote:

That means, if you miss in the middle of map and then get 300 combo, you'll lose ~18% of score for those 300 notes. If map has 1500 notes, that will be 18% of 1/5 of 200k, lol (that's about 7k).

If I understood your post correctly (correct me if I did not), this would mean that playing badly punishes you for notes played later on in the song. That's a big problem, consider this scenario:

Beatmap X, 2000 notes, hard 20 note burst @ 980th note and another equally hard 20 note burst @ 1980th note. Player A hits all notes, but misses 1 note in the first burst. Player B hits all notes, but misses 1 note in the last burst. Both players performed equally well, but player B gets a higher score.

The same problem exists in the current scoring system. I don't think I've seen any combo based scoring system proposed that doesn't have problems like these, which is why I think it's better to look for a scoring system that does not rely on combo.

Drojoke wrote:

If I understood your post correctly (correct me if I did not), this would mean that playing badly punishes you for notes played later on in the song.

See Shoegazer's post, there's no such a problem due to logarithmic function for combo and combo limit.

My Brain: *Math Error intensifies*

Kivicat wrote:

Drojoke wrote:

If I understood your post correctly (correct me if I did not), this would mean that playing badly punishes you for notes played later on in the song.

See Shoegazer's post, there's no such a problem due to logarithmic function for combo and combo limit.

The problem is still present in some extent.

For example, in a map with 2000 notes, 3 20-note bursts at 500, 550, and 600.
Two players miss once during the first burst. Player A misses once during the 2nd bursts but doesn't miss during the 3rd. Player B doesn't miss during the 2nd burst but misses once during the 3rd. Player A gets a higher score overall than Player B.

The only case where situations like these don't happen would be when the current combo doesn't have an influence in the amount of score given by a note.

For increasing the influence of misses, it would be better to redefine how accuracy is determined. The current accuracy percentage formula (and any scaling defined by a monotonic function of it) underestimates the importance of misses and bad judgments compared to better judgments.

Kivicat wrote:

See Shoegazer's post, there's no such a problem due to logarithmic function for combo and combo limit.

By multiplying the hit value with the logarithm of your combo (up to combo limit), you're only making the problems Full Tablet and I suggested smaller. Differences in scores that really should've been the same are now smaller, but they're still there.

Full Tablet wrote:

The current accuracy percentage formula (and any scaling defined by a monotonic function of it) underestimates the importance of misses and bad judgments compared to better judgments.

Accuracy and score both underestimate the importance of misses and bad judgements, and they always will as long as misses give 0 score and 0% accuracy. Combo bases systems try to make misses matter more by (usually) reducing the score given by subsequent notes, but this creates all sorts of edge cases where such a scoring system produces very odd score differences.

I still think it's best to take inspiration from stepmania's way of punishing for misses and bad judgements; by reducing score. This punishes bad gameplay without edge cases like the ones Full Tablet and I mentioned.

lets be real the only reason they want combo scoring is because they're hoping the entire planet doesn't get blown out by an undefeated usa team that doesn't even play their game, again

NoSaucierMagic wrote:

lets be real the only reason they want combo scoring is because they're hoping the entire planet doesn't get blown out by an undefeated usa team that doesn't even play their game, again

This post made my day

Here is a formula for score based on the judgment counts (acc-based score).
It is based on linear approximations during the calculation of the normal curve that fits the distribution of the judgments (using exact formulas required using numerical methods that were somewhat computationally expensive, and maybe they were too complex); the linear approximation is better for low accuracy scores (but it still should gives sensible values for high accuracy scores).

Variables:
a: Hit window for a Rainbow (16.5ms with no-mod with old timing values).
b: Hit window for a 300 (OD and Mod dependent).
c: Hit window for a 200.
d: Hit window for a 100.
e: Hit window for a 50.
JMAX/J300/.../JMISS : Count of the judgments.

anm: Linear scale constant, to make the maximum score 1,000,000, then anm=a. To make scores between different OD or timing window mods (HR/EZ) in the same map directly comparable (a higher score represents a higher accuracy), then anm has to be constant regardless of the timing windows.

1. To balance EZ/HR/Nomod, it would be a good idea to always make anm equal to the value of "a" with Nomod. That way HR can go beyond 1 million score (as long as the player has very good accuracy during the play), EZ gets less than 1 million even with all rainbows, all-Rainbows with no-mod is always 1 million score.
2. If the timing window for a Rainbow is ever OD-dependent, then anm would need to be OD-dependent as well to make the max score with no-mod 1 million, but then, the difficulty of achieving certain score would vary depending on the OD. For that reason, I think it would be better to not change the timing window of a Rainbow depending on OD.

Score formula: https://dl.dropboxusercontent.com/s/4z6 ... a.png?dl=0

Edit: Here is a better approximation (but this formula might be too long): http://i.imgur.com/rcWR71N.png

NoSaucierMagic wrote:

lets be real the only reason they want combo scoring is because they're hoping the entire planet doesn't get blown out by an undefeated usa team that doesn't even play their game, again

Beautiful.

NinjaSM wrote:

NoSaucierMagic wrote:

lets be real the only reason they want combo scoring is because they're hoping the entire planet doesn't get blown out by an undefeated usa team that doesn't even play their game, again

Beautiful.

hi

i play this game

i'm advocating and have been helping with this score change

so do yourself and everyone else who has had nothing but a negative opinion without explanation a favor: kindly piss off.

as to others who are still helping out and giving suggestions: that's what this needs.

Just sharing my two cents on FL mod, the visual mod I am usually playing with.

Certain people with visual mods ON may find it easier to play, and perform better, why? They are only required to focus a small portion (example: upper half ) of the screen, and process in their brains n' hit the keys accordingly. Notes on the lower half is deemed unnecessary to the players; they may treat anything appears at the lower half as a distraction which may make them lose focus.

However, personally, FL mod makes things harder mostly, only when a song contains lots of SVs. Take these songs for example:
xi - Happy End of the World by Blocko
FamiRockP - Oni Kanojo by Entozer
Risk Junk-G - Candy Galy by Staiain

Recall the portions of these songs which has slowjams.

As player score is judged heavily based on accuracy in this game, heavy SV of a song (especially unpredictable "slowjams", or too many notes to be processed during slowjams) will affect accuracy of a player, this brings the difficulty for the player to get a better score.

For some people, they may think FL is difficult and very impressive....as only minority group of people plays with FL. However, IMO, it is not that hard but it takes time for someone to get oneself familiar with the mod.


It's pretty layman as I am not someone good in describing things, but I hope the community gets my point.

smoogipooo wrote:

1. Make DT adjust to 100%/110%/.../150% with score bonus increments of 0.05x (or something like that).

Please do.

Can you apply that to HT of 0.8 and 0.9?

Khelly wrote:

Can you apply that to HT of 0.8 and 0.9?

This would be amazing. I play a lot of 0.8x-0.9x in stepmania because I usually suck at 180-200 bpm jumpstreams and 150-180 bpm jumpstream maps are harder to find. I definitely miss being able to practice harder songs without going full snail mode with HT @0.75x speed.

There's lots of fun maps for me that are too easy ht and way too hard nomod; that's why I widh osu mania had those speeds

Are you think LN (lonte noooooooodel) is easy?

And you give 1 combo for 1 ln?

Ok try ENTODZER map like dis
1, https://osu.ppy.sh/s/130464

Dis

2. https://osu.ppy.sh/s/146623

And dia

3. https://osu.ppy.sh/s/138430


I waiting you result score and kombos.

And say ln is easy like nornal notes.

masdafugh wrote:

Are you think LN (lonte noooooooodel) is easy?

And you give 1 combo for 1 ln?

Ok try ENTODZER map like dis
1, https://osu.ppy.sh/s/130464

Dis

2. https://osu.ppy.sh/s/146623

And dia

3. https://osu.ppy.sh/s/138430


I waiting you result score and kombos.

And say ln is easy like nornal notes.

Considering that the actual combo number didn't affect score at all in the V1 scoring, LNs are worth more than they used to be, since they have two judgements now.

I had a multiplayer session with a friend to test out v2 a bit. Wanted to see it in action after hearing complaints. Here's a play we had that stood out to me, and which I thought could make for some good input.
These spoilerboxes will contain an image each.


Sv2 score, first play:

Sv1 score, second play:

On my first impression of these, I thought this was fair, even scores and all for fairly even plays. What with the 200 count being the same.
However, what stands out is the MAX/300 ratio. It's a fair bit higher in the v2 play (thank you LNs), while managing to receive a lower score.

V2: 980/198/5/0/0/0 = 991944 score
V1: 921/210/5/0/0/0 = 991995 score

This is giving me the impression that MAX count impacts score less in v2 than in v1.
What am I missing here? Is this how it's supposed to be?

‎

Redon wrote:

Gameplay: [smoogipooo] Increase osu!mania FL ScoreV2 mod multiplier to 1.10x.

Please stop. I thought ten pages of people explaining why this is a bad idea was enough? Just get rid of it, it's nothing but a visual aid.

But he really wants to shoehorn it in, people explaining why its bad just isn't enough.

oi... I was pretty sure that a bunch of people, including those who mained Flashlight... were (rightfully) against a scoreboost... why are we adding it now?