show more
Consider an FC on an OD 8 song with a perfect normal distribution and standard deviation of 20ms (i.e. 200 UR). Disregarding combo score, this would give a score 959.3k:
Consider the same play, with HR turned on. The score is now reduced to 910.3k before multipliers:
With a 1.06x multiplier, the HR score would've been 964944.5k, which is very close to the no-mod score. A 1.2x multiplier would make the score skyrocket to 1,092.4k. A mere 820k HR rank S score, including several 100s and 50s, would dominate a 980k rank SS score.
HR's 1.2x multiplier is so good, you're only better off playing no-mod when HR makes you fail a song or if you're trying to play a song that's way out of your leage. I would suggest keeping it roughly at its original multiplier of 1.06x. Emphasis added to roughly, because I'm sure 1.08x would work out just fine. I just can't be bothered to dig any deeper right now.
On a side note, I'd like to point out how incredibly high the aforementioned HR score is. Assuming a 1000 note song, that's 152 200s and 9 100s. Both the no-mod and HR scores would be even closer to both the 1 million mark and each other, because I ignored the combo aspect of the score. This just goes to show how severely unpunished bad judgements are in ScoreV2. Using a purely accuracy based scoring system similar to Stepmania's MIGS DP, does pretty well in this situation:
The current proposal has the combo cap fixed at 400, so now the scoring system is only flawed for songs with less than 400 notes. This may not sound like a big deal, but this means it's going to be easier on maps like Drop - Granat to get a higher score. Noodle maps with very little notes yet considerable difficulty also come to mind as songs that will be easier to score on with ScoreV2.
???
These are these score multiplier I would do
Reason why I think FL should give a slight score multiplier is because it does make vision a bit tighter so note will be lumped up also they would have to use about half their normal nomod scroll speed. It's lower than FI because if you get FI too the max, it's below the middle of the screen (middle being where most average/top players focus on looking while FL is always in the middle). The higher the bpm the harder it can get because of density. I use to play with FL only for months and I think FI is more challenging a bit.
I want to give HR to be 1.10x and not 1.20x is because are you trying to say HR = 140% rate of song?
HR is higher than visual mods is because accurate players should be rewarded more than visual mods since they doesn't even do much (except for FI/FL slightly)
DT max 150% should be 1.25x because it's makes the map way more challenging then playing with HR or a visual mod.
HT increments is different from DT because HT makes maps way too easy. Since it makes the map easier than intended, it should be more lower of course.
This is OD 10 nomod
This is OD 10 with HR
You could mitigate this by including 300gs into accuracy, but from what I've experimented it might create too much emphasis on MAX accuracy with charts that players have issues getting 96%+ on (and as a result would not be an accurate assessment of skill). You can try having a weightage of 310 instead of 320 for accuracy instead, but it's up to you.
But by including the rainbow component into accuracy, I don't really see much purpose in having a combo component in the first place, but I'm not quite sure how to justify it and put it in words.
Alternatively, you can avoid including MAXes in the accuracy component and just increase the importance of MAXes to like 360 to increase the emphasis of it by a noticeable but not overpowering amount in the combo component, but that requires a bit more experimentation.
Also, emphasis on underjoy's post. He brings up a pretty good point about LNs, and it wouldn't make sense for him to have so many misses with a distribution like that. Here's a HRv1 and HRv2 comparison as well for another chart. Given the 50/100 windows of v1, it wouldn't make much sense for the miss count in v2 to be that high in the first place, unless there's some bug that scorev1 has with LNs.
You can also confirm that by looking at ideally same combos on FCs which is nearly impossible to do on v1.
We all agreed to play that match on score v2 so everyone was prepared for it
EDIT: Also since the timing windows for MAX scales with OD on v2, could you maybe give us a small table with how wide are they for major OD and HR?
Please test again.
I don't know what to say.
Also reposting this I guess, because the overemphasis on 200s is still a major problem and this would fix the problem:
Also, The V2 scoring for osu!mania is available in the Stable version right now.
I want ScoreV2 to be the BestThingEver™ as much as everyone else here, but I'm pretty sure osu!mania is even better off with the old bonus score compared to ScoreV2's combo component. If the devs have decided to go with it anyway (have they even stated this?), then that's a pity.
/rant
Accuracy score should be the main value used for the rest of the performance and ranking system (as long there is only one ranking based on the one-dimensional pp value of each player), since the count of the judgments is a more reliable statistic to base inferences about the skill of a player, compared to the length of string of non-misses (mixing both by adding two different formulas together doesn't make it better than a pure accuracy-based formula either, the quality of the formula used for overall score increases the lower the weight of the combo portion of it is).
That said, basing the accuracy score on the current Accuracy Percentage formula is not good either. Accuracy Percentage underweights the importance of Rainbows over regular 300s, and overweights the importance of 200s over 100s/50s/Misses.
For measuring the ability of being able to hit hard notes without missing, miss count is a better statistic than a score based on combo.
If you want to isolate speed/dexterity skill from overall accuracy skill, ideally, you should be able to calculate 2 different scores for each aspect of skill, and have two different leaderboards and rankings for the same plays.
It sounds confusing because it's sort off tackling the same problem in the exact opposite way. You're not actually rewarding players for keeping high combos, you're just upping the punishment of misses. Fewer misses generally equal higher combo, with edge cases like FC'ing a song and then missing the last 5 notes being an exception to that rule. Unsurprisingly, these edge cases also produce funny scores in ScoreV1/V2, but not in MIGS scoring for example.
The main ways you can get a massive amount of misses and only misses is if you're very negligent in an easy section (misreads), minor misreads in hard sections (which are generally penalised less anyway) or if you're hitting extremely conservatively in hard sections. Being negligent in an easy section is usually minor (1-2 misses), it's very very rare to actually early miss a massive chord in the first place anyway. And even then, that's an extreme edge case - and an extremely rare scenario that is rarely replicated in normal gameplay and can be discarded as an anomaly in gameplay.
Combo multiplier has the issue of not being able to discriminate between 5 misses spread out in easy sections and 5 misses in a hard section, definitely, but so do misses. Having a combo multiplier is just increasing the magnitude of that source of error. This is ultimately why people don't think combo multiplier is a good idea - it's an unnecessary metric that adds nothing if the game itself does not promote the idea of "perfection". osu!mania does not promote perfection. Otherwise, the PP system would favour more towards 1mil scores rather than 700-850K scores, SR would look at only the hardest part of the chart (it technically does this already if the hard section is long enough, but it doesn't address super short spikes completely well), and non-300 judgements would be far far more punishing.
show more
But he really wants to shoehorn it in, people explaining why its bad just isn't enough.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.
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?
Increase osu!mania HR ScoreV2 mod multiplier to 1.20x.This seems like a really significant change, coming from a multiplier of 1.06x.
Consider an FC on an OD 8 song with a perfect normal distribution and standard deviation of 20ms (i.e. 200 UR). Disregarding combo score, this would give a score 959.3k:
SPOILER
R300: 57.63%
300: 37.75%
200: 4.50%
100: 0.20%
50: 0.00%
300: 37.75%
200: 4.50%
100: 0.20%
50: 0.00%
Consider the same play, with HR turned on. The score is now reduced to 910.3k before multipliers:
SPOILER
R300: 41.77%
300: 42.10%
200: 15.22%
100: 0.89%
50: 0.20%
300: 42.10%
200: 15.22%
100: 0.89%
50: 0.20%
With a 1.06x multiplier, the HR score would've been 964944.5k, which is very close to the no-mod score. A 1.2x multiplier would make the score skyrocket to 1,092.4k. A mere 820k HR rank S score, including several 100s and 50s, would dominate a 980k rank SS score.
HR's 1.2x multiplier is so good, you're only better off playing no-mod when HR makes you fail a song or if you're trying to play a song that's way out of your leage. I would suggest keeping it roughly at its original multiplier of 1.06x. Emphasis added to roughly, because I'm sure 1.08x would work out just fine. I just can't be bothered to dig any deeper right now.
On a side note, I'd like to point out how incredibly high the aforementioned HR score is. Assuming a 1000 note song, that's 152 200s and 9 100s. Both the no-mod and HR scores would be even closer to both the 1 million mark and each other, because I ignored the combo aspect of the score. This just goes to show how severely unpunished bad judgements are in ScoreV2. Using a purely accuracy based scoring system similar to Stepmania's MIGS DP, does pretty well in this situation:
SPOILER
MIGS-like scoring system, giving score like this:
R300: 3
300: 2
200: 1
100: -2
50: -4
Miss: -8
No-mod score: 841.6k
HR base score: 740.5k
HR * 1.20x: 888.6k
HR * 1.06x: 784.9k
Because of severe punishment when hitting 100s, 50s and misses, the 1.2x multiplier doesn't seem so bad anymore. If you can't play incredibly accurately, these bad judgements are going to lower your score to the point where you're better off playing with no-mod. That's how I think it should be.
R300: 3
300: 2
200: 1
100: -2
50: -4
Miss: -8
No-mod score: 841.6k
HR base score: 740.5k
HR * 1.20x: 888.6k
HR * 1.06x: 784.9k
Because of severe punishment when hitting 100s, 50s and misses, the 1.2x multiplier doesn't seem so bad anymore. If you can't play incredibly accurately, these bad judgements are going to lower your score to the point where you're better off playing with no-mod. That's how I think it should be.
With the score system I made, HR/EZ get a balanced score by design. In previous posts I have posts I posted some approximate formulas for it, but I think that using a calculation method that is more accurate would better, I wrote some code that can calculate score using double precision floating numbers quite accurately.
The algorithm basically does a maximum likelihood fit of the timing errors in the play, using a normal curve with zero mean, using as input the distribution of the judgments in the play (taking into account their timing windows); the parameter to fit is the standard deviation of the curve (the lower the standard deviation, the more accurate is the play). The timing window of LN releases would be considered be same as of presses, to account for the increased difficulty of releasing accurately (assuming using the current timing windows already balance the difficulty of releases compared to presses)
Then, a penalty is applied based on the amount of judgments in the play, to take into account if the results were just by chance. This is analogous to the case of estimating how loaded a coin is by throwing it 5 times, if you get 5 heads in a row, a safer estimation of the probability of getting a head would be 87.055% (probability when getting at least 5 heads when throwing 5 times happens half of the time) instead of 100% (the probability that makes the result of the throws most likely).
Then, the standard deviation (std) found is mapped to a score value (standard deviation of 0 = 1 million score, standard deviation-> Infinity = 0 score). The exact function used to map standard deviation to score doesn't really matter (as long as it strictly monotonically decreasing), but a good option would be using something in the form of: (A*Erf(a/std) + B*Erf(b/std) + ...)/(A+B+...)*1,000,000, since that way score scales in the same way as a accuracy%/DP/MIGS/Exscore system (while still avoiding the balancing problems those systems have).
Note that since a penalty is applied based on the amount of judgments, getting 1 million score is not possible with a map with a finite amount of notes (you can only get very close, when the amount of notes is high and the timing windows are tight), this might be seen as a problem, but it actually this solves another bigger problem. Without that penalty, getting all rainbows with EZ or No-mod would both give 1 million score, so there would be no reason to ever use HR if you want to maximize your score; with the penalty, using more strict timing windows makes it possible to get closer to 1 million score, giving reason to players to use HR if they can handle it.
Examples of scores (using a standard deviation -> score function that scales similarly to Exscore): Note that those plays were done with Scorev1 mechanics (only 1 judgment per LN), but the idea remains.
https://osu.ppy.sh/b/647965 Drop - Granat [5K EZ] (a very short map)
All Rainbows No-Mod: 993,738 score.
All Rainbows EZ Mod : 966,397 score.
All Rainbows HR Mod: 999,830 score.
https://osu.ppy.sh/b/552745 Soleily - Renatus [Hard], all plays done by me in the same day: Rainbows/300/200/100/50/Miss
1040/641/145/16/4/7 No-Mod: 650,298 score.
1343/446/52/7/0/5 EZ Mod: 646,645 score.
710/745/316/63/3/16 HR Mod: 657,208 score.
https://osu.ppy.sh/b/866194 deadmau5 - Orange File [Emptiness]
235/137/31/1/0/0 No-Mod: 729,438 score.
293/101/10/0/0/0 EZ-Mod: 723,009 score.
170/138/86/9/0/1 HR-Mod: 704,914 score.
All Rainbows No-mod: 999,143 score.
https://osu.ppy.sh/b/746951 UNDEAD CORPORATION - The Empress scream off ver [MX]
677/530/162/23/6/30 No-Mod: 519,410 score.
432/564/295/82/8/47 HR Mod: 544,990 score.
930/408/55/14/1/20 EZ-Mod: 501,934 score
All Rainbows No-mod: 999,757 score.
The algorithm basically does a maximum likelihood fit of the timing errors in the play, using a normal curve with zero mean, using as input the distribution of the judgments in the play (taking into account their timing windows); the parameter to fit is the standard deviation of the curve (the lower the standard deviation, the more accurate is the play). The timing window of LN releases would be considered be same as of presses, to account for the increased difficulty of releasing accurately (assuming using the current timing windows already balance the difficulty of releases compared to presses)
Then, a penalty is applied based on the amount of judgments in the play, to take into account if the results were just by chance. This is analogous to the case of estimating how loaded a coin is by throwing it 5 times, if you get 5 heads in a row, a safer estimation of the probability of getting a head would be 87.055% (probability when getting at least 5 heads when throwing 5 times happens half of the time) instead of 100% (the probability that makes the result of the throws most likely).
Then, the standard deviation (std) found is mapped to a score value (standard deviation of 0 = 1 million score, standard deviation-> Infinity = 0 score). The exact function used to map standard deviation to score doesn't really matter (as long as it strictly monotonically decreasing), but a good option would be using something in the form of: (A*Erf(a/std) + B*Erf(b/std) + ...)/(A+B+...)*1,000,000, since that way score scales in the same way as a accuracy%/DP/MIGS/Exscore system (while still avoiding the balancing problems those systems have).
Note that since a penalty is applied based on the amount of judgments, getting 1 million score is not possible with a map with a finite amount of notes (you can only get very close, when the amount of notes is high and the timing windows are tight), this might be seen as a problem, but it actually this solves another bigger problem. Without that penalty, getting all rainbows with EZ or No-mod would both give 1 million score, so there would be no reason to ever use HR if you want to maximize your score; with the penalty, using more strict timing windows makes it possible to get closer to 1 million score, giving reason to players to use HR if they can handle it.
Examples of scores (using a standard deviation -> score function that scales similarly to Exscore): Note that those plays were done with Scorev1 mechanics (only 1 judgment per LN), but the idea remains.
https://osu.ppy.sh/b/647965 Drop - Granat [5K EZ] (a very short map)
All Rainbows No-Mod: 993,738 score.
All Rainbows EZ Mod : 966,397 score.
All Rainbows HR Mod: 999,830 score.
https://osu.ppy.sh/b/552745 Soleily - Renatus [Hard], all plays done by me in the same day: Rainbows/300/200/100/50/Miss
1040/641/145/16/4/7 No-Mod: 650,298 score.
1343/446/52/7/0/5 EZ Mod: 646,645 score.
710/745/316/63/3/16 HR Mod: 657,208 score.
https://osu.ppy.sh/b/866194 deadmau5 - Orange File [Emptiness]
235/137/31/1/0/0 No-Mod: 729,438 score.
293/101/10/0/0/0 EZ-Mod: 723,009 score.
170/138/86/9/0/1 HR-Mod: 704,914 score.
All Rainbows No-mod: 999,143 score.
https://osu.ppy.sh/b/746951 UNDEAD CORPORATION - The Empress scream off ver [MX]
677/530/162/23/6/30 No-Mod: 519,410 score.
432/564/295/82/8/47 HR Mod: 544,990 score.
930/408/55/14/1/20 EZ-Mod: 501,934 score
All Rainbows No-mod: 999,757 score.
I suggest that hitting a 50 should break your combo because the combo bonus is supposed to reward consistency.
Quoting smoogipooo
Not breaking combo at 50s would undoubtedly cause people to be rewarded by indiscriminate mashing and punish high accuracy players as low accuracy players would have a chance to overtake high accuracy players through combo points awarded by unskilled mashing.
It is very easy to mash any map in the whole game, as shown by drace's play below. (check t/224534&start=0 for more details)
On the contrary, if the combo is broken at 50s, it can ensure that players are rightfully rewarded by their consistency in hitting the right notes at the right time, which was the objective of the game before any changes.
That is why I strongly suggest that combo to break at 50 to preserve the value of combo as a way to sort good players from bad players.
Quoting smoogipooo
We want to value the more accurate players (accuracy) whilst applying a small reward for consistency (combo).Consistency is the ability to hit good judgments continuously, and therefore hitting bad judgments does not show this skill and should not be awarded extra points due to consistency. Other more established rhythm games such as O2Jam also break your combo at a BAD (roughly equivalent to Osu!mania's 50). This implies that breaking combo at 50 is a tried and tested move and is a more sensible scoring system.
Not breaking combo at 50s would undoubtedly cause people to be rewarded by indiscriminate mashing and punish high accuracy players as low accuracy players would have a chance to overtake high accuracy players through combo points awarded by unskilled mashing.
It is very easy to mash any map in the whole game, as shown by drace's play below. (check t/224534&start=0 for more details)
On the contrary, if the combo is broken at 50s, it can ensure that players are rightfully rewarded by their consistency in hitting the right notes at the right time, which was the objective of the game before any changes.
That is why I strongly suggest that combo to break at 50 to preserve the value of combo as a way to sort good players from bad players.
In regards to the combo section of what scorev2 currently is, a lot of people have recommended doing a percentage of the max combo, to find the combo cap. I think that instead of having the combo cap linearly related to the max combo, there should be a square root relationship.
I propose: Combo Cap = floor(sqrt(Max Combo) * 10)
This means:
-500 max combo has a 223 combo cap
-1000 max combo has a 316 combo cap
-2000 max combo has a 447 combo cap
-5000 max combo has a 707 combo cap
but then again id rather combo not matter at all so lol.
I propose: Combo Cap = floor(sqrt(Max Combo) * 10)
This means:
-500 max combo has a 223 combo cap
-1000 max combo has a 316 combo cap
-2000 max combo has a 447 combo cap
-5000 max combo has a 707 combo cap
but then again id rather combo not matter at all so lol.
I strongly disagree. Making the combo cap dependent on the max combo is problematic, as it couples the difficulty of getting a high score to the length of a map.Cuber wrote:
I think that instead of having the combo cap linearly related to the max combo, there should be a square root relationship.
The current proposal has the combo cap fixed at 400, so now the scoring system is only flawed for songs with less than 400 notes. This may not sound like a big deal, but this means it's going to be easier on maps like Drop - Granat to get a higher score. Noodle maps with very little notes yet considerable difficulty also come to mind as songs that will be easier to score on with ScoreV2.
Neither do I. Holding combo shouldn't be awarded, especially in osu!mania where key mashing goes unpunished. Hitting accurately should be rewarded, hitting poorly should be punished and hitting that which does not exist should be punished if not severely discouraged. All of these things can be achieved without involving combo.Cuber wrote:
but then again id rather combo not matter at all so lol.
Just a tangential point: I legitimately hate it when people use that video of all examples to prove that osu!mania is easy to mash. He's literally smashing 1/2 notes on a song where the spacing between 1/2 notes is larger than the individual timing windows themselves. Of course he's not going to break combo.
The odds of running into a chart like that are far lower than they are running into something that's actually got a legitimate structure/melody.
The odds of running into a chart like that are far lower than they are running into something that's actually got a legitimate structure/melody.
I hope you're extending edge cases of players who ghost tap in between notes to keep a solid rhythm (and not necessarily are mashing to hold combo).Cuber wrote:
but then again id rather combo not matter at all so lol.
Neither do I. Holding combo shouldn't be awarded, especially in osu!mania where key mashing goes unpunished. Hitting accurately should be rewarded, hitting poorly should be punished and hitting that which does not exist should be punished if not severely discouraged. All of these things can be achieved without involving combo.
There's no point in punishing for hitting keys when there's nothing to play within 'x' ms, where 'x' is the time in ms when a miss is usually triggered (or something similar). Pressing 5 keys on a 3 note chord is a whole different story, of course.Halogen- wrote:
I hope you're extending edge cases of players who ghost tap in between notes to keep a solid rhythm (and not necessarily are mashing to hold combo).
You're getting closer. You'd want to consider x in a time per lane as well. I see no issue hitting 5 keys on a 3 note chord if that chord is alone and by itself. Likewise, if a song has a high tempo but slow repeated notes, players might feel inclined to fill in a rhythm on the other hand to keep steady (I do this all the time).Drojoke wrote:
There's no point in punishing for hitting keys when there's nothing to play within 'x' ms, where 'x' is the time in ms when a miss is usually triggered (or something similar). Pressing 5 keys on a 3 note chord is a whole different story, of course.Halogen- wrote:
I hope you're extending edge cases of players who ghost tap in between notes to keep a solid rhythm (and not necessarily are mashing to hold combo).
???
These are these score multiplier I would do
- HD 1.00x
- FL 1.05x
- FI 1.08x
- HR 1.10x
- DT adjust to 110%/120%/.../150% with score bonus increments of 0.05x (lowest is 110% with 1.05x multiplier and max is 150% with 1.25x score multiplier)
- HT adjust to 75%/80%/.../95% with score bonus increments of 0.10x (lowest being 75% with 0.50x score multiplier and max as 95% with 0.9x multiplier)
Reason why I think FL should give a slight score multiplier is because it does make vision a bit tighter so note will be lumped up also they would have to use about half their normal nomod scroll speed. It's lower than FI because if you get FI too the max, it's below the middle of the screen (middle being where most average/top players focus on looking while FL is always in the middle). The higher the bpm the harder it can get because of density. I use to play with FL only for months and I think FI is more challenging a bit.
I want to give HR to be 1.10x and not 1.20x is because are you trying to say HR = 140% rate of song?
HR is higher than visual mods is because accurate players should be rewarded more than visual mods since they doesn't even do much (except for FI/FL slightly)
DT max 150% should be 1.25x because it's makes the map way more challenging then playing with HR or a visual mod.
HT increments is different from DT because HT makes maps way too easy. Since it makes the map easier than intended, it should be more lower of course.
my opinion on score multipliers:
HD - 1.00x
FI - 1.00x
FL - 1.00x
HR - 1.15x
DT - 1.30-1.35x
HT - 0.5x
NF - 0.8x (it's just no fail)
EZ - 0.8x (doesn't make a significant difference really)
HD - 1.00x
FI - 1.00x
FL - 1.00x
HR - 1.15x
DT - 1.30-1.35x
HT - 0.5x
NF - 0.8x (it's just no fail)
EZ - 0.8x (doesn't make a significant difference really)
If we can have all visual mods be 1.00x I would love that
even though they said they would remove score bonus on all visual mods but instead kept it for FL
even though they said they would remove score bonus on all visual mods but instead kept it for FL
Setting a fixed score multiplier for HR and EZ, with the current accuracy formula (accuracy percentage scaled with x^(1+4x)) makes it so there exists accuracy ranges that determine which mod is the most convenient (maximizes the score part of the total score) to use depending on the level of accuracy of the player.
The table is interpreted like in the following example:
If in a map a player gets in average 99.5% accuracy percentage with no mod, then he would get in average more accuracy score using HR if it's multiplier is x1.024 or more (if it is lower than that, then no mod would give in average more score), similarly, the player would get more score in average with EZ compared to No-Mod if the multiplier was x0.894 or more.
IMO, those ranges where in average players get more score by using a specific timing window mod shouldn't exist, the player should get the same score if he plays the notes the same way, regardless of the mod used (and make the mods just a change of scale of the worth of the judgments, analogous as being able to measure the mass of something with either kilograms or pounds). Since the accuracy part of the score only has information about the judgment counts (and not the exact timing of each hit), this is not possible in the edge case where the player gets 100% accuracy or close to it (but it can be accounted for correctly, by making rainbows different to 300s in the accuracy part, and considering the amount of notes to penalize based on the probability the accuracy obtained was far from the expected accuracy of a play; or even better, make the timing windows of rainbows constant regardless of OD and Mods).
Edit: Another possible idea:
Add a new judgment beyond Rainbows (300g), with a tight fixed timing window (for example, +/-10ms), which is required to hit for every note to get 100% of the possible score.
Advantages of the judgment:
The table is interpreted like in the following example:
If in a map a player gets in average 99.5% accuracy percentage with no mod, then he would get in average more accuracy score using HR if it's multiplier is x1.024 or more (if it is lower than that, then no mod would give in average more score), similarly, the player would get more score in average with EZ compared to No-Mod if the multiplier was x0.894 or more.
IMO, those ranges where in average players get more score by using a specific timing window mod shouldn't exist, the player should get the same score if he plays the notes the same way, regardless of the mod used (and make the mods just a change of scale of the worth of the judgments, analogous as being able to measure the mass of something with either kilograms or pounds). Since the accuracy part of the score only has information about the judgment counts (and not the exact timing of each hit), this is not possible in the edge case where the player gets 100% accuracy or close to it (but it can be accounted for correctly, by making rainbows different to 300s in the accuracy part, and considering the amount of notes to penalize based on the probability the accuracy obtained was far from the expected accuracy of a play; or even better, make the timing windows of rainbows constant regardless of OD and Mods).
Edit: Another possible idea:
Add a new judgment beyond Rainbows (300g), with a tight fixed timing window (for example, +/-10ms), which is required to hit for every note to get 100% of the possible score.
Advantages of the judgment:
- Scores that were close to the maximum score possible in the previous system would be farther away from it. Since they are farther away, the difference between the median accuracy and the mean accuracy of a player that repeats the map becomes smaller, making it not as necessary to compensate for map length to make a system that balances between different timing window mods fairly. This is specially important for easy charts, where scores of 1 million or very close to it are common.
- Even if there are scores that end up being close to 100% anyways, having the timing windows of the tightest timing window the same regardless of mods makes the differences between mods smaller for the high accuracy range (virtually the same for 100% and close to it), so there wouldn't be need to compensate for map length even in that case.
I've done some testing on group stage maps from last year's MWC, me Feerum and Xylo played on diffrent mods, first I played FL, Feru HR and xylo nomod and then me and Feru nomod and Xylo HR. I've played FL for about a week on and off, so my acc is still meh on it, but I've noticed some things:
1) Sometimes on HR and even on FL you randomly get a fail on the end of the map, but most of the time it doesn't show it in mp link while it does in your game. For example Xylo and Feru had that on their respective HR plays of Good-Bye Tears, I had that on my FL play of Analyze. Feru's fail on Analyze was also the same thing (he had HP on the end and it just dropped to zero for no reason) and it did show up on mp link.
2) The mods have too high multiplier for sure.
First example is Vision (we had a wrong mode setup, but w/e, that just doesn't show all the stats).
Me vs Xylo - I had 95,82%, 9 misses, 371 max combo and similar ratio while he had 99,00%, 1 miss, 978 combo and I somehow still managed to win by 22k points despite having score objectively a shitton worse.
The scores are 951,783 vs 929,027, my score without a multiplier would be about 856k and really, score that bad should lose despite having FL on.
Second example is Good-bye Tears
Me and Xylo had similar accuracy (I had 0,3% less), 3 misses and I had a slightly higher combo, while Feerum on HR had half of our max combos, negative ratio, 9 misses and almost 2% worse accuracy and he still managed to win with Xylo by 70k points and only lost to me by 18k. The diffrence between our scores wasn't that big, and while I should've still won since I had a similar score on FL, Feru shouldn't have won by this much. Even counting in the diffuclty of scoring with HR, it's giving too much of an edge, it should be a lot closer.
The scores were 1,003,659 (FL) vs 985,018 (HR) vs 915,042 (NM) and without multipliers it would be 912,4k (FL) vs 820,8k (HR) vs 915k (NM). The fact that you need to score about 830k to beat any nomod score is not fair imho.
I'd suggest making the mods, especially HR riskier to pick by lowering the multipliers to 1,06x for FL and 1,1 for HR. That way if you want to edge out nomod plays you need to score around 940k with FL alone, around 909k with HR alone and about 857k with FL+HR which in my opionin is a reasonable threshhold.
Here's the mp link for said testing match: https://osu.ppy.sh/mp/26344409
1) Sometimes on HR and even on FL you randomly get a fail on the end of the map, but most of the time it doesn't show it in mp link while it does in your game. For example Xylo and Feru had that on their respective HR plays of Good-Bye Tears, I had that on my FL play of Analyze. Feru's fail on Analyze was also the same thing (he had HP on the end and it just dropped to zero for no reason) and it did show up on mp link.
2) The mods have too high multiplier for sure.
First example is Vision (we had a wrong mode setup, but w/e, that just doesn't show all the stats).
Me vs Xylo - I had 95,82%, 9 misses, 371 max combo and similar ratio while he had 99,00%, 1 miss, 978 combo and I somehow still managed to win by 22k points despite having score objectively a shitton worse.
The scores are 951,783 vs 929,027, my score without a multiplier would be about 856k and really, score that bad should lose despite having FL on.
Second example is Good-bye Tears
Me and Xylo had similar accuracy (I had 0,3% less), 3 misses and I had a slightly higher combo, while Feerum on HR had half of our max combos, negative ratio, 9 misses and almost 2% worse accuracy and he still managed to win with Xylo by 70k points and only lost to me by 18k. The diffrence between our scores wasn't that big, and while I should've still won since I had a similar score on FL, Feru shouldn't have won by this much. Even counting in the diffuclty of scoring with HR, it's giving too much of an edge, it should be a lot closer.
The scores were 1,003,659 (FL) vs 985,018 (HR) vs 915,042 (NM) and without multipliers it would be 912,4k (FL) vs 820,8k (HR) vs 915k (NM). The fact that you need to score about 830k to beat any nomod score is not fair imho.
I'd suggest making the mods, especially HR riskier to pick by lowering the multipliers to 1,06x for FL and 1,1 for HR. That way if you want to edge out nomod plays you need to score around 940k with FL alone, around 909k with HR alone and about 857k with FL+HR which in my opionin is a reasonable threshhold.
Here's the mp link for said testing match: https://osu.ppy.sh/mp/26344409
Just to clarify things...
With hard rock enabled, any diff with OD8 and HP8 will be adjusted to OD10 and HP10, which is easily done, despite with a huge accuracy drop from players. The HR gives score boost that isn't quiet necessary. Reduce the score multiplier to 1.05 would be just fine.
I also consider to rank the random mod if enabled with HR, as it will be much more harder than before, if RD and HR combined, it will be 1.10 with completely random notes on each tick.
With hard rock enabled, any diff with OD8 and HP8 will be adjusted to OD10 and HP10, which is easily done, despite with a huge accuracy drop from players. The HR gives score boost that isn't quiet necessary. Reduce the score multiplier to 1.05 would be just fine.
I also consider to rank the random mod if enabled with HR, as it will be much more harder than before, if RD and HR combined, it will be 1.10 with completely random notes on each tick.
hey guys i know it's a weird request but can someone make sure that this never becomes anything more than a tournament score?
That is only half of the story, HR also makes timing windows tighter than OD10's by a decent about:O2MasterFX wrote:
Just to clarify things...
With hard rock enabled, any diff with OD8 and HP8 will be adjusted to OD10 and HP10, which is easily done, despite with a huge accuracy drop from players. The HR gives score boost that isn't quiet necessary. Reduce the score multiplier to 1.05 would be just fine.
This is OD 10 nomod
This is OD 10 with HR
EZ/HR multiplies/divides the timing windows by exactly 1.4. The UI has a weird method of rounding everything to 0.5ms, but even when assuming the numbers shown by the UI are correct you're left with an insignificant margin of error.-Kamikaze- wrote:
... HR also makes timing windows tighter than OD10's by a decent about: ...
just my random input, consider that hp7 hr is hp 9.8 which is easy enough to die on during tournament finals and stuff, for the high risk i would at least go higher then 1.1 or make it scale somewhat with difficulty but idk how you would do that
Last try at this.
Currently accuracy does make a good metric of how hard the map is. I can tell you straight up what the star rating of a map would be according to my accuracy. I feel the score should reflect that. However, there is a point, around between 986k and 1M scores, where it is more of a matter of consistency. Playing a score for 990K+ often feels like playing standard. A lot of retries and a lot of hope for more MAX. If anything, I think that's where you should focus combo based scoring at (if you are persistent about it).
Instead of the combo portion being dependent on misses, have it be dependent on not have a MAX and adjust so it is worth around 100k out of 1M or whatever most would agree on. That way players who are trying to get a better play have to worry about their overall performance by acc'ing the map like now, and players who are going for the SS and beyond would have to worry more about consistency.
Just don't have the element of luck forced on players unless they are really going for a perfect play. And I do say luck because consistency involves taking the probability that you may fuck up somewhere along the line.
Currently accuracy does make a good metric of how hard the map is. I can tell you straight up what the star rating of a map would be according to my accuracy. I feel the score should reflect that. However, there is a point, around between 986k and 1M scores, where it is more of a matter of consistency. Playing a score for 990K+ often feels like playing standard. A lot of retries and a lot of hope for more MAX. If anything, I think that's where you should focus combo based scoring at (if you are persistent about it).
Instead of the combo portion being dependent on misses, have it be dependent on not have a MAX and adjust so it is worth around 100k out of 1M or whatever most would agree on. That way players who are trying to get a better play have to worry about their overall performance by acc'ing the map like now, and players who are going for the SS and beyond would have to worry more about consistency.
Just don't have the element of luck forced on players unless they are really going for a perfect play. And I do say luck because consistency involves taking the probability that you may fuck up somewhere along the line.
Gameplay: [smoogipooo] Fix osu!mania ScoreV2 LNs not correctly capping to 50 score when the hold starts after the LN start was fully missed.Why not remove the ability to repress missed LN's? I've always thought it was weird that completely missing a LN gets punished, but missing it and, repressing it waaaaay too late and not releasing until seconds after the LN has ended somehow gives you points and pepe points.
Gameplay: [smoogipooo] Reduce osu!mania ScoreV2 FL multiplier to 1.06x (prev. 1.10x).I still stand for FL giving 1.0x points, for reasons many people have pointed out already.
I would like to look at the v2 when it comes to harder LN maps with HR. As a pretty decent LN player I tried some things out and I was astonished with the results. On the screenshot below you can see that I had a pretty good acc and nice score. However I FAILED IN THE MIDDLE because LN releases on hr are so incredibly tight that you get a miss on what would have been 200 (or even 300) on v1 nomod. This is so hilarious when you hokd a good acc then the denser sections come and you fail so easily because of high HP on HR. All but 1-2 of the misses were because of stupid release windows. Despite my efforts to release as precisely as I can, I still died. The release windows on HR should be more lenient, especially when it comes to window for miss. I practically have not received other release judgments than MAX, 300, 200 or miss.
My suggestion: either increase the miss window for releases (with making the 100/50 windows wider as well), or change all of the misses to 50s which will prevent failing the map so easily.
My suggestion: either increase the miss window for releases (with making the 100/50 windows wider as well), or change all of the misses to 50s which will prevent failing the map so easily.
btw as we discovered today while we had a friendly Poland vs Argentina match on score v2 the maxes are disgustingly underrated for some reason.
I mean take a look at this:
Poland won despite with 6 less 200s overall while having 220 maxes less, and by over 5k points to top it off which is just ridiculous.
Also if you take a look at Juan's and Hudonom's scores: Hudonom got a score 18 points better than juan despite having 98 less maxes just because he had one less 200. That is really not okay.
There were other cases where scores were ridiculously close for no good reason:
Grubonom vs Juan vs aluu, I know that Grubonom missed right at the end but still
MAX'es really should be weighted much more that they are now.
mp link for this match: https://osu.ppy.sh/mp/26530910
I mean take a look at this:
Poland won despite with 6 less 200s overall while having 220 maxes less, and by over 5k points to top it off which is just ridiculous.
Also if you take a look at Juan's and Hudonom's scores: Hudonom got a score 18 points better than juan despite having 98 less maxes just because he had one less 200. That is really not okay.
There were other cases where scores were ridiculously close for no good reason:
Grubonom vs Juan vs aluu, I know that Grubonom missed right at the end but still
MAX'es really should be weighted much more that they are now.
mp link for this match: https://osu.ppy.sh/mp/26530910
Topic Starter
That sounds impossible. Are you sure everyone was actually using Stable (latest) / Beta / CE?
Right now, percentage is the only number that factors into the Accuracy calculation,, so there's no differentiation between 300g/300. Therefore, there is only a very tiny portion of the score actually affected by the difference (mainly the combo).
Actually, when looking at the code, I'm not even sure that is true right now, though that's probably not intended. I'm a dumbass I found it.
Also as a minor side note: Maybe consider starting the combo counter at max combo? (if possible). That way, it doesn't diminish the weighting of the first 400 notes of the map.
Also as a minor side note: Maybe consider starting the combo counter at max combo? (if possible). That way, it doesn't diminish the weighting of the first 400 notes of the map.
This is correct - 320s are very much underweighted because the only component of the scoring system that takes into account 320 accuracy is the combo component, which only has a 20% prominence. Add on to the fact that the difference between a 300 and 320 is so small and that the absolute difference between juan and Hudo's 320 count isn't that significant, it would make sense that 320s are really underweighted at the moment.Ciel wrote:
Right now, percentage is the only number that factors into the Accuracy calculation,, so there's no differentiation between 300g/300. Therefore, there is only a very tiny portion of the score actually affected by the difference (mainly the combo).
You could mitigate this by including 300gs into accuracy, but from what I've experimented it might create too much emphasis on MAX accuracy with charts that players have issues getting 96%+ on (and as a result would not be an accurate assessment of skill). You can try having a weightage of 310 instead of 320 for accuracy instead, but it's up to you.
But by including the rainbow component into accuracy, I don't really see much purpose in having a combo component in the first place, but I'm not quite sure how to justify it and put it in words.
Alternatively, you can avoid including MAXes in the accuracy component and just increase the importance of MAXes to like 360 to increase the emphasis of it by a noticeable but not overpowering amount in the combo component, but that requires a bit more experimentation.
Also as a minor side note: Maybe consider starting the combo counter at max combo? (if possible). That way, it doesn't diminish the weighting of the first 400 notes of the map.Also supporting this.
Also, emphasis on underjoy's post. He brings up a pretty good point about LNs, and it wouldn't make sense for him to have so many misses with a distribution like that. Here's a HRv1 and HRv2 comparison as well for another chart. Given the 50/100 windows of v1, it wouldn't make much sense for the miss count in v2 to be that high in the first place, unless there's some bug that scorev1 has with LNs.
Yes I am sure, look at freemod picks, everyone had multipliers on mods. Also on score v1 100 maxes more would edge out 1x200 less by about 3,5k on Ambitious.smoogipooo wrote:
That sounds impossible. Are you sure everyone was actually using Stable (latest) / Beta / CE?
You can also confirm that by looking at ideally same combos on FCs which is nearly impossible to do on v1.
We all agreed to play that match on score v2 so everyone was prepared for it
EDIT: Also since the timing windows for MAX scales with OD on v2, could you maybe give us a small table with how wide are they for major OD and HR?
Make MAX worth 100% and 300 worth 99% 
Topic Starter
HR was being added twice to the hit windows. I've pushed a fix (only on CE for now) that rectifies this and modifies all other timing windows sliiiiiightly so they align with MAXs a bit better by interpolating from OD0-OD5 and OD5-OD10 instead of linear scaling from OD0-10._underjoy wrote:
I would like to look at the v2 when it comes to harder LN maps with HR. As a pretty decent LN player I tried some things out and I was astonished with the results. On the screenshot below you can see that I had a pretty good acc and nice score. However I FAILED IN THE MIDDLE because LN releases on hr are so incredibly tight that you get a miss on what would have been 200 (or even 300) on v1 nomod. This is so hilarious when you hokd a good acc then the denser sections come and you fail so easily because of high HP on HR. All but 1-2 of the misses were because of stupid release windows. Despite my efforts to release as precisely as I can, I still died. The release windows on HR should be more lenient, especially when it comes to window for miss. I practically have not received other release judgments than MAX, 300, 200 or miss.
My suggestion: either increase the miss window for releases (with making the 100/50 windows wider as well), or change all of the misses to 50s which will prevent failing the map so easily.
Please test again.
I don't know what to say.
With the current system not using HR after you reach some level of accuracy is a bad idea (unless the map has some part that would make you fail with HR, even with good overall accuracy). Additionally, the FL multiplier is more significant the closer the play is to a perfect play.Todestrieb wrote:
I don't know what to say.
There's another issue with the current proposal I don't think has been addressed in this thread. Bad judgements like 50s, 100s (and 200s to some extent) have significantly less impact on score compared to ScoreV1.
In ScoreV1, bonusscore is severely reduced by anything worse than a 300. Bonusscore accounted for 50% of your total score, so getting a ton of bad judgements is going to seriously impact your total score in ScoreV1. ScoreV2 replaces bonus score, which punishes bad judgements, with combo score, which only punishes misses. Mashing is a common complaint in osu!mania already, but this is going to get worse by rewarding players for keeping combo without taking bad judgements into account.
To give a slightly over the top example: scoring a 200 on every note in a song is going to give you an accuracy of 66.67% and a score of about ~335k. The score is really low because you barely get any bonus score. The same play would net you a somewhat respectable 733.3k in the current proposal. Add FL to this, and you get a score of 777.3k. In practice, it's going to be less extreme than this, but it's definitely present.
I don't think this is acceptable at all. Inaccurately mashing your way through hard parts of a song while keeping combo is going to give really good scores. Not just because of the combo score, but also because bad judgements give too much score.
In ScoreV1, bonusscore is severely reduced by anything worse than a 300. Bonusscore accounted for 50% of your total score, so getting a ton of bad judgements is going to seriously impact your total score in ScoreV1. ScoreV2 replaces bonus score, which punishes bad judgements, with combo score, which only punishes misses. Mashing is a common complaint in osu!mania already, but this is going to get worse by rewarding players for keeping combo without taking bad judgements into account.
To give a slightly over the top example: scoring a 200 on every note in a song is going to give you an accuracy of 66.67% and a score of about ~335k. The score is really low because you barely get any bonus score. The same play would net you a somewhat respectable 733.3k in the current proposal. Add FL to this, and you get a score of 777.3k. In practice, it's going to be less extreme than this, but it's definitely present.
I don't think this is acceptable at all. Inaccurately mashing your way through hard parts of a song while keeping combo is going to give really good scores. Not just because of the combo score, but also because bad judgements give too much score.
Yeah, maxes should be weighted a lot more imolim38 wrote:
lmao.
Have been talking to other people about scorev2 and its issues. The main three issues that people have said are:
- Rainbows are weighted too little.
- HR on v2 is ridiculously difficult to get rainbows on (on Stable at least, on Cutting Edge it seems as easy as accumulating rainbows on nomod - but more on that later).
- LNs generate a lot of misses if they have really tricky releases. They work fine for other types of LNs.
All of those problems can definitely be amended.
Rainbow Accuracy
HardRock
LNs
I think that should be all, feel free to ask any questions if you're uncertain about a couple of things that I've pointed out/suggested.
- Rainbows are weighted too little.
- HR on v2 is ridiculously difficult to get rainbows on (on Stable at least, on Cutting Edge it seems as easy as accumulating rainbows on nomod - but more on that later).
- LNs generate a lot of misses if they have really tricky releases. They work fine for other types of LNs.
All of those problems can definitely be amended.
Rainbow Accuracy
SPOILER
I've been experimenting with weightages and discussing with people about how much a 200 should be worth compared to a 300. I initially thought that 310 would be fine (and a 200 would be worth 11 300s), but when it came to matches like this, if accuracy was the only factor, Argentina would win by 21,000 points. I do think that Argentina should win and it's a step in the right direction, but 21,000 seems extremely overwhelming since it undermines the fact that Poland had overall, noticeably less 200s. I tried it with harder charts too and they seem to favour rainbow accuracy a little too much for my liking - especially since when it comes to harder charts (where players struggle with), good rainbow accuracy is usually caused by variance rather than a higher skill level. 200s and worse judgements should determine performance for that.
I wanted to use 307 afterwards, but it still gave a bit too much emphasis for my liking, about 12,500 points for that Argentina/Poland match. I went down to 305, and the difference is about 6,800. I think that's ultimately the most reasonable assessment, and others I've talked to seem to agree with the prenotion that a 200 is about 21 normal 300s. Ignoring the bad judgements (since those values are pretty much set in stone at this point), this is probably (part of) the ideal solution. This does mean that only full rainbow scores are SSs, but I don't see that as a problem as frames of reference can be shifted.
Getting rid of the difference between a rainbow and a normal 300 in the combo scoring component is probably ideal too, since that should be in the accuracy component, not the combo component. If rainbows are included into accuracy, the combo component does not need a rainbow component.
I also wanted to soften the exponential curve a tiny bit when it comes to including rainbows, mainly because at a certain point extremely good accuracy is more caused by variance rather than a very high skill level - unless the performance is consistently done, which is not measurable with just one match and one attempt. The exponential I had in mind was Accuracy^(2 + 2 * Accuracy), but it's essentially Accuracy^4 - so 1 power down.
Similar note, wanted to respond to this:
tl;dr: Embed rainbows into accuracy with a weightage of 305 instead of 320, change the accuracy curve to Accuracy^(2 + Accuracy * 2), remove the differentiation between rainbows and normal 300s in the combo component (both of them should have a HitValue of 30).
I initially wanted to increase the rainbow judgement weightage without embedding rainbows into accuracy, but no matter how much I changed it, the difference is very minor (~600-1,200 points) and a 200 will almost always be too powerful compared to a rainbow 300. So I scrapped that idea and thought that embedding rainbows into accuracy with a reasonable weightage and maybe making the curve more lenient would be the best idea.Shoegazer wrote:
320s are very much underweighted because the only component of the scoring system that takes into account 320 accuracy is the combo component, which only has a 20% prominence. Add on to the fact that the difference between a 300 and 320 is so small and that the absolute difference between juan and Hudo's 320 count isn't that significant, it would make sense that 320s are really underweighted at the moment.
You could mitigate this by including 300gs into accuracy, but from what I've experimented it might create too much emphasis on MAX accuracy with charts that players have issues getting 96%+ on (and as a result would not be an accurate assessment of skill).Alternatively, you can avoid including MAXes in the accuracy component and just increase the importance of MAXes to like 360 to increase the emphasis of it by a noticeable but not overpowering amount in the combo component, but that requires a bit more experimentation.
I've been experimenting with weightages and discussing with people about how much a 200 should be worth compared to a 300. I initially thought that 310 would be fine (and a 200 would be worth 11 300s), but when it came to matches like this, if accuracy was the only factor, Argentina would win by 21,000 points. I do think that Argentina should win and it's a step in the right direction, but 21,000 seems extremely overwhelming since it undermines the fact that Poland had overall, noticeably less 200s. I tried it with harder charts too and they seem to favour rainbow accuracy a little too much for my liking - especially since when it comes to harder charts (where players struggle with), good rainbow accuracy is usually caused by variance rather than a higher skill level. 200s and worse judgements should determine performance for that.
I wanted to use 307 afterwards, but it still gave a bit too much emphasis for my liking, about 12,500 points for that Argentina/Poland match. I went down to 305, and the difference is about 6,800. I think that's ultimately the most reasonable assessment, and others I've talked to seem to agree with the prenotion that a 200 is about 21 normal 300s. Ignoring the bad judgements (since those values are pretty much set in stone at this point), this is probably (part of) the ideal solution. This does mean that only full rainbow scores are SSs, but I don't see that as a problem as frames of reference can be shifted.
Getting rid of the difference between a rainbow and a normal 300 in the combo scoring component is probably ideal too, since that should be in the accuracy component, not the combo component. If rainbows are included into accuracy, the combo component does not need a rainbow component.
I also wanted to soften the exponential curve a tiny bit when it comes to including rainbows, mainly because at a certain point extremely good accuracy is more caused by variance rather than a very high skill level - unless the performance is consistently done, which is not measurable with just one match and one attempt. The exponential I had in mind was Accuracy^(2 + 2 * Accuracy), but it's essentially Accuracy^4 - so 1 power down.
Similar note, wanted to respond to this:
A 66.67% score nets you about 306K (181K for accuracy, 125K for combo). Adding FL increases it to 324K, you might've miscalculated. In any case, I do agree with the fact that bad judgements (non-200/300 judgements) should be penalised more, but I don't think it's necessarily what they have in mind at the moment, since the values are carved in stone. MAX judgements are not.Drojoke wrote:
To give a slightly over the top example: scoring a 200 on every note in a song is going to give you an accuracy of 66.67% and a score of about ~335k. The score is really low because you barely get any bonus score. The same play would net you a somewhat respectable 733.3k in the current proposal. Add FL to this, and you get a score of 777.3k. In practice, it's going to be less extreme than this, but it's definitely present.
tl;dr: Embed rainbows into accuracy with a weightage of 305 instead of 320, change the accuracy curve to Accuracy^(2 + Accuracy * 2), remove the differentiation between rainbows and normal 300s in the combo component (both of them should have a HitValue of 30).
HardRock
SPOILER
Accumulating rainbows on HR on most charts is really strict already, but since the rainbow window is stricter in scorev2 in ODs beyond 8, it gets even stricter and probably way too difficult. On anything above OD7.5, the HR will be boosted to 10 - which means that it has a rainbow window of +-13ms. Add on the general effects of HR (which makes windows 40% tighter), it gets knocked down to 9ms, truncated. Considering that any hits within +-5ms is caused by computer performance variance, having only 4ms of "controlled" timing is very very low - especially since the difference between an 6ms controlled window (HRv1 window) and a 4ms controlled window is huge. While it is true that certain modes have windows this tight on HR (and maybe DT), 85% of charts used in MWC are OD8, whereas it's much less common in other game modes to have something this tight (in Standard it's some absurd DTHR with a decent OD, and in Taiko it's some absurdly high OD with DTHR, and I'd argue that timing on both games are easier than Mania).
There's also the fact that with the rainbow weightage aspect included, wins on FreeMod are very variance-based rather than performance-based.
My main suggestion is to keep the 40% tighter window (except for early miss judgements) on HR, but not increase OD with HR. 11ms (OD8-OD9 is standard), while I still think is noticeably harsh, is the norm and the 40% seems to fit the other judgement windows quite well anyway. In fact, v1 already does this, this is a OD5 chart on HR. While it appears as if it looks like OD7+HR, is actually aligned with OD5 with 40% tighter windows. Ignore the additional 0.5ms, as it's some byproduct of woc's janky coding.
But that's one thing. The other thing I noticed is that the difference between rainbow difficulty between scorev2 NoMod and scorev2 HR on on Cutting Edge is insignificant. I've experimented this with juan, and his performance on NoMod and HR is similar, variance included. Pictures included:
A 25x300 difference for something that should be a 6ms gap (15ms - 9ms) is absurd, and is most definitely not caused by variance in performance.
I told juan to play on the Stable build as well, and then he noticed a massive difference in accuracy with scorev2 HR. He can usually get a 6:1 to 7:1 ratio on scorev2 HR in CE (and is comparable to his nomod scores consistently), but he can barely break 2.5:1 on Stable. He didn't seem to mention any performance issues either. Here are pics of his scores on the Stable build:
Note that these two charts are also comparable in difficulty as well, both would've been in Group Stages last year (and Sakura Mirage was in last year).
I know you mentioned a HR rainbow fix earlier in the thread, but I'm not sure if it did what it was supposed to do. Seems like it created issues rather than fixed a problem. underjoy's ratio on HR before the fix made sense to me given the relatively low OD and all, my main concern was with the miss count (which I'll talk about in the next section).
tl;dr: Don't increase OD when HR is switched on but keep the stricter timing windows, Accumulating rainbows on OD10+HR is way too strict because the window where a player can control is really really small and it affects an overwhelming majority of charts in MWC. There might be a problem in Cutting Edge where HR currently is as easy as NoMod.
There's also the fact that with the rainbow weightage aspect included, wins on FreeMod are very variance-based rather than performance-based.
My main suggestion is to keep the 40% tighter window (except for early miss judgements) on HR, but not increase OD with HR. 11ms (OD8-OD9 is standard), while I still think is noticeably harsh, is the norm and the 40% seems to fit the other judgement windows quite well anyway. In fact, v1 already does this, this is a OD5 chart on HR. While it appears as if it looks like OD7+HR, is actually aligned with OD5 with 40% tighter windows. Ignore the additional 0.5ms, as it's some byproduct of woc's janky coding.
But that's one thing. The other thing I noticed is that the difference between rainbow difficulty between scorev2 NoMod and scorev2 HR on on Cutting Edge is insignificant. I've experimented this with juan, and his performance on NoMod and HR is similar, variance included. Pictures included:
A 25x300 difference for something that should be a 6ms gap (15ms - 9ms) is absurd, and is most definitely not caused by variance in performance.
I told juan to play on the Stable build as well, and then he noticed a massive difference in accuracy with scorev2 HR. He can usually get a 6:1 to 7:1 ratio on scorev2 HR in CE (and is comparable to his nomod scores consistently), but he can barely break 2.5:1 on Stable. He didn't seem to mention any performance issues either. Here are pics of his scores on the Stable build:
Note that these two charts are also comparable in difficulty as well, both would've been in Group Stages last year (and Sakura Mirage was in last year).
I know you mentioned a HR rainbow fix earlier in the thread, but I'm not sure if it did what it was supposed to do. Seems like it created issues rather than fixed a problem. underjoy's ratio on HR before the fix made sense to me given the relatively low OD and all, my main concern was with the miss count (which I'll talk about in the next section).
tl;dr: Don't increase OD when HR is switched on but keep the stricter timing windows, Accumulating rainbows on OD10+HR is way too strict because the window where a player can control is really really small and it affects an overwhelming majority of charts in MWC. There might be a problem in Cutting Edge where HR currently is as easy as NoMod.
LNs
SPOILER
LNs in v1 seem to be bugged - making them much easier than they should be. While I understand how the LN mechanics work, it doesn't seem to work that way for LNs that you don't let go but you hit the LN head perfectly. For some reason, no matter when you let go of the note, as long as you hit the head perfectly, you will get a 200. If you hit the head a bit earlier or later, you get a 100 instead. Here is video demonstration of this. This is probably (partially) why LNs in v1 are so easy compared to ones in v2 - particularly ones with very tricky releases. Players subconsciously don't let go of LNs properly and they don't get punished for it. In v2, the punishment becomes noticeable. In v2, players get a miss if they don't let go.
Getting rid of this bug is definitely a good start, but since scorev2 is implemented at such short notice (85% of participants probably haven't used scorev2 yet, though you can argue that it's their own fault) and MWC is used as testing grounds, you'd want to make LN releases more lenient than they currently are at the moment for easier transitioning - as players are getting a lot of misses already, even on NoMod. I think a LN leniency of 1.8x would be fine, but this is a bit of an arbitrary figure. I don't really know the effects of this because I don't play LN charts that much, and you're better off asking players like juankristal or _underjoy instead.
tl;dr: Increase LN leniency to 1.8x for easier transitioning, as LNs in v1 contain a bug that makes LN releases much much easier than they should've been in the first place.
Getting rid of this bug is definitely a good start, but since scorev2 is implemented at such short notice (85% of participants probably haven't used scorev2 yet, though you can argue that it's their own fault) and MWC is used as testing grounds, you'd want to make LN releases more lenient than they currently are at the moment for easier transitioning - as players are getting a lot of misses already, even on NoMod. I think a LN leniency of 1.8x would be fine, but this is a bit of an arbitrary figure. I don't really know the effects of this because I don't play LN charts that much, and you're better off asking players like juankristal or _underjoy instead.
tl;dr: Increase LN leniency to 1.8x for easier transitioning, as LNs in v1 contain a bug that makes LN releases much much easier than they should've been in the first place.
I think that should be all, feel free to ask any questions if you're uncertain about a couple of things that I've pointed out/suggested.
Gameplay: [smoogipooo] Make osu!mania ScoreV2 early miss window unchanged by HR.Such a small detail, but very effective in punishing spamming. I'm digging this!
Hello,
I decided to do some testing of LN maps on v2 system. As you know, there was previously an issue that caused LN maps on HardRock to be barely playable due to the excessive amount of misses made on LN releases. However, today I have observed that this problem is also really visible on v2 nomod play. We've tested some maps, here is the mp link and some screenshots from our play that show the problem very well:
https://osu.ppy.sh/mp/26731666
https://osu.ppy.sh/ss/5743076
http://puu.sh/qlPHd/c3d92da6d5.png
As you can see, there is one thing those screenshots have in common: abnormal amount of misses and negligible amount of 50 and 100 judgments. As a pretty decent LN player I can easily notice when I'm hitting a LN as I should and when I'm releasing it too early/too late. My observations are that for releasing a LN on normal terms you can get only MAX, 300, 200 and miss. Apparently there is a bug that doesn't allow the other judgements to appear and converts them into misses.
This is very visible on the Funny Shuffle map, which has really bizarre LN endings. If the system worked correctly, I'd get a lot of 50s and 100s for uncorrect releases - however, I got only one 50. How? It seems that while I cannot see a way to get a 100 from releasing a long note, there is a possibility (and only possibility it seems) to get a 50, if you release the LN much too early and repress it.
Anemone LN is another difficult LN map with unintuitive long note patterning, and releases don't follow the music. I struggled with the map, not only because it's quite hard, but mostly because the lack of my precision on releasing gave me a high number of misses which caused me to fail really quickly as well as tank the score and be unable to recover. The amounts of 50 and 100 also look very suspicious. In this map I also tried to repress the LNs as said above, and I got 50s. This leads to an interesting conclusion: it's more beneficial to repress a LN, than to hold it too long (to avoid a miss).
About maps like Inside DT, where the numbers look okay: 100 and 50 judgments are caused by wrong note presses, not by releases. I still felt got two seemingly random misses near the start of the map. At the ending fast jacks combined with noodles made me obtain bad judgments.
It's really easy to observe that LN releases are at the moment not working as intended, and while this might not be so blatant in early stages of the MWC, it will surely cause a lot of trouble in higher-end maps.
I decided to do some testing of LN maps on v2 system. As you know, there was previously an issue that caused LN maps on HardRock to be barely playable due to the excessive amount of misses made on LN releases. However, today I have observed that this problem is also really visible on v2 nomod play. We've tested some maps, here is the mp link and some screenshots from our play that show the problem very well:
https://osu.ppy.sh/mp/26731666
https://osu.ppy.sh/ss/5743076
http://puu.sh/qlPHd/c3d92da6d5.png
As you can see, there is one thing those screenshots have in common: abnormal amount of misses and negligible amount of 50 and 100 judgments. As a pretty decent LN player I can easily notice when I'm hitting a LN as I should and when I'm releasing it too early/too late. My observations are that for releasing a LN on normal terms you can get only MAX, 300, 200 and miss. Apparently there is a bug that doesn't allow the other judgements to appear and converts them into misses.
This is very visible on the Funny Shuffle map, which has really bizarre LN endings. If the system worked correctly, I'd get a lot of 50s and 100s for uncorrect releases - however, I got only one 50. How? It seems that while I cannot see a way to get a 100 from releasing a long note, there is a possibility (and only possibility it seems) to get a 50, if you release the LN much too early and repress it.
Anemone LN is another difficult LN map with unintuitive long note patterning, and releases don't follow the music. I struggled with the map, not only because it's quite hard, but mostly because the lack of my precision on releasing gave me a high number of misses which caused me to fail really quickly as well as tank the score and be unable to recover. The amounts of 50 and 100 also look very suspicious. In this map I also tried to repress the LNs as said above, and I got 50s. This leads to an interesting conclusion: it's more beneficial to repress a LN, than to hold it too long (to avoid a miss).
About maps like Inside DT, where the numbers look okay: 100 and 50 judgments are caused by wrong note presses, not by releases. I still felt got two seemingly random misses near the start of the map. At the ending fast jacks combined with noodles made me obtain bad judgments.
It's really easy to observe that LN releases are at the moment not working as intended, and while this might not be so blatant in early stages of the MWC, it will surely cause a lot of trouble in higher-end maps.
Topic Starter
Thanks to _underjoy and a few others that notified me of this - yes LN release timings weren't accounting for the 1.5x lenience. They have been fixed now. Please test again!
Holding LNs also gradually restores HP again now.
Holding LNs also gradually restores HP again now.
ok this is too funny not to comment on:
was playing a very long song, too hard, so i stopped playing. watched as my score dropped for a while
2 minutes later into the song, and I haven't hit any more notes, and my score is going up now
was playing a very long song, too hard, so i stopped playing. watched as my score dropped for a while
2 minutes later into the song, and I haven't hit any more notes, and my score is going up now
Topic Starter
I've checked to make sure what was causing this, and it is intended. The reasoning is simple - towards the lower end of the accuracy model, there is less of a difference between 30% accuracy and 20% accuracy (as an example), but this is not the only factor that defines the accuracy portion of your score. To keep things looking progressive it also takes into account how much of the map has been completed, so for very short maps this value increases much more drastically than your accuracy decreases at those lower accuracies.
To simplify it, no, the calculations are not incorrect and you aren't getting "more score by missing", it's just your accuracy being recalculated to consider how much of the map has been completed.
To simplify it, no, the calculations are not incorrect and you aren't getting "more score by missing", it's just your accuracy being recalculated to consider how much of the map has been completed.
The fix that fixed LN misses does not seem to have made its way to the beta build. A fix before the tournament would be much appreciated
Topic Starter
The fix will propagate all the way to Stable (latest) by Friday in time for MWC. Please use Cutting Edge until then.
Edit: It's been propagated back to Beta now.
Edit: It's been propagated back to Beta now.
Thank you. The main issue was that it was only available for supporterssmoogipooo wrote:
The fix will propagate all the way to Stable (latest) by Friday in time for MWC. Please use Cutting Edge until then.
Edit: It's been propagated back to Beta now.
Hello, so i noticed a super major issue on scorev2 related to LNS (I'm using beta btw)
Even with this new update about the ''hold gives hp'' this is mere nothing, like i didnt feel nothing.
Its hard to talk about this, but like on score v1, a map with HP 9, if you KEEP missing and hitting your hp will be like dropping a little and recovering FAST.
On scorev2 a map on HP 8.5, if you keep missing and hitting or even missing a little and start combo'ing the DRAIN is pretty big and the recovery is super small.
https://osu.ppy.sh/ss/5756787 - score v2
https://osu.ppy.sh/ss/5756802 - score v1
i even stopped played for like 2secs on v1 and still got my hp to recover in like 1~2secs, on v2 if i rush any ln pattern my hp will be drained so much and the recovery is super small i cant even reach max hp, score v1 replay ~ http://puu.sh/qo0SF/86fb41302b.osr (score v2 replay is lost, my osu! didnt saved idk why)
so please take a look again on LN HP Drain, and LN HP Recovery.
Thanks
i dont consider me a good LN player, but all players who knows me always tell me i'm a good LN player, so well you should at least read my opnion.
here some scores that maybe be a proof about my ln skills
- https://osu.ppy.sh/ss/5756872
- https://osu.ppy.sh/ss/5756878
- http://osu.ppy.sh/ss/5756892
- https://osu.ppy.sh/ss/5756897
- https://osu.ppy.sh/ss/5756900
Even with this new update about the ''hold gives hp'' this is mere nothing, like i didnt feel nothing.
Its hard to talk about this, but like on score v1, a map with HP 9, if you KEEP missing and hitting your hp will be like dropping a little and recovering FAST.
On scorev2 a map on HP 8.5, if you keep missing and hitting or even missing a little and start combo'ing the DRAIN is pretty big and the recovery is super small.
https://osu.ppy.sh/ss/5756787 - score v2
https://osu.ppy.sh/ss/5756802 - score v1
i even stopped played for like 2secs on v1 and still got my hp to recover in like 1~2secs, on v2 if i rush any ln pattern my hp will be drained so much and the recovery is super small i cant even reach max hp, score v1 replay ~ http://puu.sh/qo0SF/86fb41302b.osr (score v2 replay is lost, my osu! didnt saved idk why)
so please take a look again on LN HP Drain, and LN HP Recovery.
Thanks
i dont consider me a good LN player, but all players who knows me always tell me i'm a good LN player, so well you should at least read my opnion.
here some scores that maybe be a proof about my ln skills
- https://osu.ppy.sh/ss/5756872
- https://osu.ppy.sh/ss/5756878
- http://osu.ppy.sh/ss/5756892
- https://osu.ppy.sh/ss/5756897
- https://osu.ppy.sh/ss/5756900
Topic Starter
@FelipeLink: Have you tried making a simple map of few notes and testing? I have and the HP looks almost exactly the same. The difference is that LNs have two judgements now instead of one, so the punishment is doubled for missing a complete LN. It is also slightly harder to achieve perfect scoring on LNs due to both the reduced leniency of LNs and that they have two judgements.
All according to plan in my opinion?
All according to plan in my opinion?
Pretty sure he might tested that in stable while the change was only in cutting edge implemented. Not sure thosmoogipooo wrote:
@FelipeLink: Have you tried making a simple map of few notes and testing? I have and the HP looks almost exactly the same. The difference is that LNs have two judgements now instead of one, so the punishment is doubled for missing a complete LN. It is also slightly harder to achieve perfect scoring on LNs due to both the reduced leniency of LNs and that they have two judgements.
All according to plan in my opinion?
Topic Starter
Nah he specifically stated beta.
I believe his main concern was the fact that the HP recovery while holding on to LNs in v1 was far more lenient than what it currently is in v2.smoogipooo wrote:
@FelipeLink: Have you tried making a simple map of few notes and testing? I have and the HP looks almost exactly the same. The difference is that LNs have two judgements now instead of one, so the punishment is doubled for missing a complete LN. It is also slightly harder to achieve perfect scoring on LNs due to both the reduced leniency of LNs and that they have two judgements.
All according to plan in my opinion?
Its hard to talk about this, but like on score v1, a map with HP 9, if you KEEP missing and hitting your hp will be like dropping a little and recovering FAST.From what he told me, the HP drain of LNs in v2 was reasonable, he was talking about the HP recovery itself that is too harsh (regenerates too slowly) and causes him to fail much more easily than on v1. Maybe that's intended?
On scorev2 a map on HP 8.5, if you keep missing and hitting or even missing a little and start combo'ing the DRAIN is pretty big and the recovery is super small. i even stopped played for like 2secs on v1 and still got my hp to recover in like 1~2secs, on v2 if i rush any ln pattern my hp will be drained so much and the recovery is super small i cant even reach max hp
Also reposting this I guess, because the overemphasis on 200s is still a major problem and this would fix the problem:
Shoegazer wrote:
Rainbow AccuracySPOILERI initially wanted to increase the rainbow judgement weightage without embedding rainbows into accuracy, but no matter how much I changed it, the difference is very minor (~600-1,200 points) and a 200 will almost always be too powerful compared to a rainbow 300. So I scrapped that idea and thought that embedding rainbows into accuracy with a reasonable weightage and maybe making the curve more lenient would be the best idea.Shoegazer wrote:
320s are very much underweighted because the only component of the scoring system that takes into account 320 accuracy is the combo component, which only has a 20% prominence. Add on to the fact that the difference between a 300 and 320 is so small and that the absolute difference between juan and Hudo's 320 count isn't that significant, it would make sense that 320s are really underweighted at the moment.
You could mitigate this by including 300gs into accuracy, but from what I've experimented it might create too much emphasis on MAX accuracy with charts that players have issues getting 96%+ on (and as a result would not be an accurate assessment of skill).Alternatively, you can avoid including MAXes in the accuracy component and just increase the importance of MAXes to like 360 to increase the emphasis of it by a noticeable but not overpowering amount in the combo component, but that requires a bit more experimentation.
I've been experimenting with weightages and discussing with people about how much a 200 should be worth compared to a 300. I initially thought that 310 would be fine (and a 200 would be worth 11 300s), but when it came to matches like this, if accuracy was the only factor, Argentina would win by 21,000 points. I do think that Argentina should win and it's a step in the right direction, but 21,000 seems extremely overwhelming since it undermines the fact that Poland had overall, noticeably less 200s. I tried it with harder charts too and they seem to favour rainbow accuracy a little too much for my liking - especially since when it comes to harder charts (where players struggle with), good rainbow accuracy is usually caused by variance rather than a higher skill level. 200s and worse judgements should determine performance for that.
I wanted to use 307 afterwards, but it still gave a bit too much emphasis for my liking, about 12,500 points for that Argentina/Poland match. I went down to 305, and the difference is about 6,800. I think that's ultimately the most reasonable assessment, and others I've talked to seem to agree with the prenotion that a 200 is about 21 normal 300s. Ignoring the bad judgements (since those values are pretty much set in stone at this point), this is probably (part of) the ideal solution. This does mean that only full rainbow scores are SSs, but I don't see that as a problem as frames of reference can be shifted.
Getting rid of the difference between a rainbow and a normal 300 in the combo scoring component is probably ideal too, since that should be in the accuracy component, not the combo component. If rainbows are included into accuracy, the combo component does not need a rainbow component.
I also wanted to soften the exponential curve a tiny bit when it comes to including rainbows, mainly because at a certain point extremely good accuracy is more caused by variance rather than a very high skill level - unless the performance is consistently done, which is not measurable with just one match and one attempt. The exponential I had in mind was Accuracy^(2 + 2 * Accuracy), but it's essentially Accuracy^4 - so 1 power down.
tl;dr: Embed rainbows into accuracy with a weightage of 305 instead of 320, change the accuracy curve to Accuracy^(2 + Accuracy * 2), remove the differentiation between rainbows and normal 300s in the combo component (both of them should have a HitValue of 30).
Topic Starter
That's the thing - it isn't any more or less lenient, the code is identical. It's just that there are other factors that cause you to lose more HP in ScoreV2. The actual rate of recovery of HP is unchanged. Perhaps if this starts to matter more in the later rounds I can issue a quick update, but I don't think it's good to make last minute changes until that's determined to be a problem by the map selectors.
And likewise I'm pretty sure it's too late now to improve on the weighting of MAXs. That will definitely be looked into for MWC7K, but probably not for MWC4K unless there are very significant results that show a fix is 100% needed. Let's wait until after the first round!
And likewise I'm pretty sure it's too late now to improve on the weighting of MAXs. That will definitely be looked into for MWC7K, but probably not for MWC4K unless there are very significant results that show a fix is 100% needed. Let's wait until after the first round!
Hmm, actually when we played with underjoy on updated v2 he said that hp recovery was too easy and he was playing on HR so no idea
Am I the only one who thinks it's silly to have a system potentially changed through the tournament (That it'a specifically made for) because of being unfinished?
Unless you redo round1 after a fix if that ever was needed.
Unless you redo round1 after a fix if that ever was needed.
I really don't like the idea of waiting for results of the round where the max:300 aspect matters the most by far, but eh, nothing I can do about it at this point.
@smoogi: Fair enough on the HP recovery.
I would much rather see the change now, especially since group stages is the round where MAX accuracy matters the most, and cases like this and this (lim vs bumpinho) are going to be very common. I don't think the competitors will mind a change in MAX weightage especially since it'd be similar to scorev1 would look like (which makes for easier transitioning) and from what I've talked to with other players, they certainly wouldn't mind having a change in MAX weightage if it meant for a more accurate assessment of ability. The frame of reference of what defines an SS will be malleable, especially since they know that these context of an SS is now situational.
But that's how I see it really, I w ould much rather have short-term discomfort (for a long-term gain) rather than an implementation afterwards when in the stage that matters, it had a noticeable problem for certain facets of the tournament. Ultimately up to you.
I do agree that having system changes in midst of a tournament is a bit silly, but I do feel that if it's to ensure a more accurate assesment of ability, I don't think it's a problem in itself. The fact that it is unfinished does justify the change.Khelly wrote:
Am I the only one who thinks it's silly to have a system potentially changed through the tournament (That it'a specifically made for) because of being unfinished?
Unless you redo round1 after a fix if that ever was needed.
I would much rather see the change now, especially since group stages is the round where MAX accuracy matters the most, and cases like this and this (lim vs bumpinho) are going to be very common. I don't think the competitors will mind a change in MAX weightage especially since it'd be similar to scorev1 would look like (which makes for easier transitioning) and from what I've talked to with other players, they certainly wouldn't mind having a change in MAX weightage if it meant for a more accurate assessment of ability. The frame of reference of what defines an SS will be malleable, especially since they know that these context of an SS is now situational.
But that's how I see it really, I w ould much rather have short-term discomfort (for a long-term gain) rather than an implementation afterwards when in the stage that matters, it had a noticeable problem for certain facets of the tournament. Ultimately up to you.
@smoogi:
I made a ''test'' map for trying how LN v1 works vs LN v2 and well i just found that the HP recovery is the same but well the HP drain is kinda a problem, with score v2 LN you can fail twice faster than v1, so well yeah thats the only problem i can find on lnv2, the recovery is not significant because the drain is too big.
And well, on hard LN maps even good skilled players can do rushs and miss like 1~3 lns in a row and on v2 this is like 100% dead and on v1 this isn't a big problem.
Oh and i miss my Hit Lightning, after the new beta update they just disappeared ):
(i talked with some players who use's beta and all of they hit lightning disappeared too)
I made a ''test'' map for trying how LN v1 works vs LN v2 and well i just found that the HP recovery is the same but well the HP drain is kinda a problem, with score v2 LN you can fail twice faster than v1, so well yeah thats the only problem i can find on lnv2, the recovery is not significant because the drain is too big.
And well, on hard LN maps even good skilled players can do rushs and miss like 1~3 lns in a row and on v2 this is like 100% dead and on v1 this isn't a big problem.
Oh and i miss my Hit Lightning, after the new beta update they just disappeared ):
(i talked with some players who use's beta and all of they hit lightning disappeared too)
I'm going to have to agree with Shoegazer on this. At first when I heard people were talking about how 200s were overemphasis, I thought it wasn't change a lot but when I saw those picture, that is crazy. I'm pretty sure every mania player would want this too be fix because we all know accuracy is key in mania. Just because someone has a rainbow ratio of 11:1 and someone has 4:1, that doesn't mean the person with the 4:1 ratio should win just because he has 1 less 200... (juan vs Hudonom). Need to be fix asap! It's not too late to change this before MWC starts.Shoegazer wrote:
I would much rather see the change now, especially since group stages is the round where MAX accuracy matters the most, and cases like this and this (lim vs bumpinho) are going to be very common. I don't think the competitors will mind a change in MAX weightage especially since it'd be similar to scorev1 would look like (which makes for easier transitioning) and from what I've talked to with other players, they certainly wouldn't mind having a change in MAX weightage if it meant for a more accurate assessment of ability. The frame of reference of what defines an SS will be malleable, especially since they know that these context of an SS is now situational.
What the players above are saying is true. accuracy is CRITICAL in group stages, and this is where the MAX counts really matter. It's much better to slightly give more weight to the MAXs than keep the current system and wait it out.
Matches haven't happened yet. It's not too late to make this change to the score system -- the ratios that players will get on the songs in groups will not change based off of this, but the physical scores and the representation of their abilities well.
Topic Starter
It is too late, the changes are untested, which is especially dangerous since the proposal modifies MAX's value, and there are other things to do to get osu! ready for MWC. Sorry but it's not happening just yet, and won't happen until MWC 7K. I discussed with Loctav and we both agreed the impact as a result of not putting this change through would be unnoticed in most cases.
Droves of people participating in MWC are taking note about how ridiculously impactful non-300 judgments are. We saw lim post a score where the presence of an additional 200 caused them to lose against a person who had 90x more 300s, which is absolute insanity -- this issue was actually brought to light by both map selectors: Kamikaze with the screenshot of a mock match between Poland/Argentina that showed juankristal losing to Hudonom by 18 points when he had less than half of the 300 count but a single 200 added on (and just so this doesn't get looked at the wrong way, it was juankristal's 79 to Hudonom's 177 - we're not talking small values here), and Shoegazer with his very usual, eloquent explanations as to why the formula is flawed.
This has been mentioned for the past few pages, I might add. It hasn't gone unnoticed in the slightest.
This has been mentioned for the past few pages, I might add. It hasn't gone unnoticed in the slightest.
How do the map scoreboards work if most people are playing with ScoreV1 but those with Cutting Edge have V2?
V2 scoring is only available as an option in multiplayer.numbermaniac wrote:
How do the map scoreboards work if most people are playing with ScoreV1 but those with Cutting Edge have V2?
Also, The V2 scoring for osu!mania is available in the Stable version right now.
Only 2pts separation gap for only the difference between 2 300 xD 
If It's HR and 2 max = 2 points, Sounds of Summer have 1407 objects, only 1.407/1.100.000 scores is rewarded for the MAXs? 
Combo, lim got more 300/200 on <400combo because of score multiplierlim38 wrote:
Only 2pts separation gap for only the difference between 2 300 xD
Consider the following scores from the group stages:
Cheetose has a 28.6% higher MAX/300 ratio compared to WindyS. Cheetose's score should be a fair bit higher, but somehow WindyS got a higher score??? Also notice how just a couple of 200s and 2 misses create a pretty large gap between juankristal and the entire Korean team. There are many more examples like this in the group stages. Players are barely rewarded for significantly better MAX/300 ratios, while players getting just a couple of 200s/100s/50s are severely punished. This has one very serious implication in the MWC4K: teams are severely punished for having a (relatively) bad player on their team, but not proportionally awarded for having really good players on their time.
Without changing ScoreV2 too much, having MAX's included in the accuracy component would make most sense. Thinking about it, completely ignoring the most important judgement of accuracy in the accuracy component of ScoreV2 is more than a little odd.
Cheetose has a 28.6% higher MAX/300 ratio compared to WindyS. Cheetose's score should be a fair bit higher, but somehow WindyS got a higher score??? Also notice how just a couple of 200s and 2 misses create a pretty large gap between juankristal and the entire Korean team. There are many more examples like this in the group stages. Players are barely rewarded for significantly better MAX/300 ratios, while players getting just a couple of 200s/100s/50s are severely punished. This has one very serious implication in the MWC4K: teams are severely punished for having a (relatively) bad player on their team, but not proportionally awarded for having really good players on their time.
Without changing ScoreV2 too much, having MAX's included in the accuracy component would make most sense. Thinking about it, completely ignoring the most important judgement of accuracy in the accuracy component of ScoreV2 is more than a little odd.
Can we please get rid of the combo component? These kind of wonky differences are going to be a thing as long as combo is involved in the scoring system.Tidek wrote:
Combo, lim got more 300/200 on <400combo because of score multiplierlim38 wrote:
Only 2pts separation gap for only the difference between 2 300 xD
The second thing is a bit too much to ask for since they've already decided that that's going to be a major feature and that's final.
Is it still too much to ask for when this major feature is inherently flawed? Because this clearly is the case with the combo component, as can be seen in some of the scores posted in this thread. The scoring system's combo aspect can most definitely be improved upon without completely removing it, but it's always going to be susceptible to producing undesirable scores.Khelly wrote:
The second thing is a bit too much to ask for since they've already decided that that's going to be a major feature and that's final.
I want ScoreV2 to be the BestThingEver™ as much as everyone else here, but I'm pretty sure osu!mania is even better off with the old bonus score compared to ScoreV2's combo component. If the devs have decided to go with it anyway (have they even stated this?), then that's a pity.
/rant
Has anyone talked about combo as being a mod?
Just like how perfect and sudden death are mods, combo is pretty much just the more lenient form of sudden death where you don't die when you combo break. It could give like a score multiplier or something so people have an incentive to use it.
So on a song that is really hard to FC because of some dense minijack or burst that only a few people can get will be much more rewarding for those who can FC and won't be discouraged if their full combo was beaten by some better accuracy player that completely missed that one impossible section.
This would make it more appealing for tourneys as missing a note midway basically ruins your potential from scoring an even remotely decent score while getting a full combo will be very rewarding.
Just like how perfect and sudden death are mods, combo is pretty much just the more lenient form of sudden death where you don't die when you combo break. It could give like a score multiplier or something so people have an incentive to use it.
So on a song that is really hard to FC because of some dense minijack or burst that only a few people can get will be much more rewarding for those who can FC and won't be discouraged if their full combo was beaten by some better accuracy player that completely missed that one impossible section.
This would make it more appealing for tourneys as missing a note midway basically ruins your potential from scoring an even remotely decent score while getting a full combo will be very rewarding.
As a mod, yes pleaseSquishykorean wrote:
...
What do you mean by combo mod? What would it do?Squishykorean wrote:
Has anyone talked about combo as being a mod?
Just like how perfect and sudden death are mods, combo is pretty much just the more lenient form of sudden death where you don't die when you combo break. It could give like a score multiplier or something so people have an incentive to use it.
So on a song that is really hard to FC because of some dense minijack or burst that only a few people can get will be much more rewarding for those who can FC and won't be discouraged if their full combo was beaten by some better accuracy player that completely missed that one impossible section.
This would make it more appealing for tourneys as missing a note midway basically ruins your potential from scoring an even remotely decent score while getting a full combo will be very rewarding.
He means to have options on which scoring method to use. The combo mod in this sense would use the combo based scoring.Full Tablet wrote:
What do you mean by combo mod? What would it do?
In that case, wouldn't it better if both scoring methods were calculated simultaneously for each play, and both values are recorded in the play? In leaderboards, people would have an option to sort by combo scoring or accuracy score.abraker wrote:
He means to have options on which scoring method to use. The combo mod in this sense would use the combo based scoring.Full Tablet wrote:
What do you mean by combo mod? What would it do?
Accuracy score should be the main value used for the rest of the performance and ranking system (as long there is only one ranking based on the one-dimensional pp value of each player), since the count of the judgments is a more reliable statistic to base inferences about the skill of a player, compared to the length of string of non-misses (mixing both by adding two different formulas together doesn't make it better than a pure accuracy-based formula either, the quality of the formula used for overall score increases the lower the weight of the combo portion of it is).
That said, basing the accuracy score on the current Accuracy Percentage formula is not good either. Accuracy Percentage underweights the importance of Rainbows over regular 300s, and overweights the importance of 200s over 100s/50s/Misses.
"Skill" isn't all about accuracy. Although most people agree it should be the largest factor, players have a variety of other skills which some focus more than others and that should deserve some acknowledgement in their score. In this case, it would be for players with the speed/finger dexterity to hit those few extra notes in the hardest part of a map to keep their full combo compared to accuracy players who will skim through that section and perfect everything else on the map.Full Tablet wrote:
Accuracy score should be the main value used for the rest of the performance and ranking system (as long there is only one ranking based on the one-dimensional pp value of each player), since the count of the judgments is a more reliable statistic to base inferences about the skill of a player, compared to the length of string of non-misses (mixing both by adding two different formulas together doesn't make it better than a pure accuracy-based formula either, the quality of the formula used for overall score increases the lower the weight of the combo portion of it is).
The thing is, combo is not a good way to measure the aspect of skill you mention either.Squishykorean wrote:
"Skill" isn't all about accuracy. Although most people agree it should be the largest factor, players have a variety of other skills which some focus more than others and that should deserve some acknowledgement in their score. In this case, it would be for players with the speed/finger dexterity to hit those few extra notes in the hardest part of a map to keep their full combo compared to accuracy players who will skim through that section and perfect everything else on the map.
For measuring the ability of being able to hit hard notes without missing, miss count is a better statistic than a score based on combo.
If you want to isolate speed/dexterity skill from overall accuracy skill, ideally, you should be able to calculate 2 different scores for each aspect of skill, and have two different leaderboards and rankings for the same plays.
So yeah, accuracy is the main value in both score v1 and v2. Combo is important as your score drops if you miss, whats the point of having a "combo mod" then? I dont really get it.
If you are good at doing combos but your acc is really bad why should you win against a good accuracy with some missees? Into a level where everyone misses then yeah, sure, but adding the combo mod wont change stuff imo.
If you are good at doing combos but your acc is really bad why should you win against a good accuracy with some missees? Into a level where everyone misses then yeah, sure, but adding the combo mod wont change stuff imo.
What you (and I) want is better achieved by increasing the significance of a single miss. Accuracy players will be harshly (but fairly) punished for missing, and speed/dexterity players still stand a chance with their lower accuracy and higher combos. If ScoreV2 were to drop the combo component and go this route, we would have a scoring system that...:Squishykorean wrote:
"Skill" isn't all about accuracy. Although most people agree it should be the largest factor, players have a variety of other skills which some focus more than others and that should deserve some acknowledgement in their score. In this case, it would be for players with the speed/finger dexterity to hit those few extra notes in the hardest part of a map to keep their full combo compared to accuracy players who will skim through that section and perfect everything else on the map.Full Tablet wrote:
...
- still rewards players for keeping their combo.
- treats x consequent misses the same as x isolated misses (ScoreV2 punishes isolated misses more severely)
- cannot produce broken scores where similar performances have unexplainable differences, because relatively more 200s were attained at <400 combo.
- can be significantly simpler to understand and implement.
Kempie wrote:
If ScoreV2 were to drop the combo component and go this route, we would have a scoring system that...:
- rewards players for keeping their combo.
When all you do is remove the combo component, misses aren't sufficiently punished for, i.e. keeping combo (not missing) isn't as big of a deal as it should be. By making misses more punishing, you are "rewarding" (more like not punishing) players that do not miss, and thus keep their combo.Ayaya wrote:
Kempie wrote:
If ScoreV2 were to drop the combo component and go this route, we would have a scoring system that...:
- rewards players for keeping their combo.
It sounds confusing because it's sort off tackling the same problem in the exact opposite way. You're not actually rewarding players for keeping high combos, you're just upping the punishment of misses. Fewer misses generally equal higher combo, with edge cases like FC'ing a song and then missing the last 5 notes being an exception to that rule. Unsurprisingly, these edge cases also produce funny scores in ScoreV1/V2, but not in MIGS scoring for example.
I knew what you meant but you just worded that part weird, also I just wanted to use that jpeg 
Yes I would agree punishing misses more harshly would solve much of the problem. The only thing that might make this slightly unattractive is the fact someone might mess up and cause a chain of misses even though they are hitting the notes correctly. The current combo multiplier doesn't punish as hard when this happens because as long as you recover and don't miss later, you are still given a chance at getting a decent score despite a tiny screw up that the game mechanics happened to snowball on.
Would you reward a player that missed 5 notes spread out throughout the map equally to a player who hit a chord wrong and missed 5 consecutive notes together?
This is purely opinion based but what this seems to be turning into is another stepmania scoring system and with lots of players here coming from that game, it would only be natural to lean towards their scoring system.
Would you reward a player that missed 5 notes spread out throughout the map equally to a player who hit a chord wrong and missed 5 consecutive notes together?
This is purely opinion based but what this seems to be turning into is another stepmania scoring system and with lots of players here coming from that game, it would only be natural to lean towards their scoring system.
So you should cater to the people migrating rather than having them adapt or leave like they should?Squishykorean wrote:
This is purely opinion based but what this seems to be turning into is another stepmania scoring system and with lots of players here coming from that game, it would only be natural to lean towards their scoring system.
Getting a chain of misses in isolation is an extremely rare scenario. In hard sections particularly, it is complemented with 100s and 50s, which is the major proportion of the penalty of a general CB rush/bad judgement rush. It's not misses that are particularly penalising despite the slightly higher penalty per judgement, since they happen less frequently than 50s/100s. A CB rush will almost always punish a player far more than misses in isolation, not necessarily because of the combo multiplier, but because of the accuracy component.Squishykorean wrote:
Yes I would agree punishing misses more harshly would solve much of the problem. The only thing that might make this slightly unattractive is the fact someone might mess up and cause a chain of misses even though they are hitting the notes correctly. The current combo multiplier doesn't punish as hard when this happens because as long as you recover and don't miss later, you are still given a chance at getting a decent score despite a tiny screw up that the game mechanics happened to snowball on.
Would you reward a player that missed 5 notes spread out throughout the map equally to a player who hit a chord wrong and missed 5 consecutive notes together?
This is purely opinion based but what this seems to be turning into is another stepmania scoring system and with lots of players here coming from that game, it would only be natural to lean towards their scoring system.
The main ways you can get a massive amount of misses and only misses is if you're very negligent in an easy section (misreads), minor misreads in hard sections (which are generally penalised less anyway) or if you're hitting extremely conservatively in hard sections. Being negligent in an easy section is usually minor (1-2 misses), it's very very rare to actually early miss a massive chord in the first place anyway. And even then, that's an extreme edge case - and an extremely rare scenario that is rarely replicated in normal gameplay and can be discarded as an anomaly in gameplay.
Combo multiplier has the issue of not being able to discriminate between 5 misses spread out in easy sections and 5 misses in a hard section, definitely, but so do misses. Having a combo multiplier is just increasing the magnitude of that source of error. This is ultimately why people don't think combo multiplier is a good idea - it's an unnecessary metric that adds nothing if the game itself does not promote the idea of "perfection". osu!mania does not promote perfection. Otherwise, the PP system would favour more towards 1mil scores rather than 700-850K scores, SR would look at only the hardest part of the chart (it technically does this already if the hard section is long enough, but it doesn't address super short spikes completely well), and non-300 judgements would be far far more punishing.
Since there's not much left to discuss about ScoreV2, I'll just post this interesting MWC match to remind the devs on the severity of MAX's being underrated:
Other scores left out for brevity.
Other scores left out for brevity.
Maybe the 1xgood was too bad for halogen
It's pretty telling that we know how harsh it is given we've also seen what the difference is with individual 300 judgments -- the fact that I got 164 less 300s but only had a single 200 and lost by 350 points when our scores were that high is very telling: 300s are not weighted enough, or judgments beneath that are weighted too heavily (or a combination of the two)...
Yeah, holy crap, that should not happen. If the ratio was closer I could see guil edging that out, but with that much difference in 300's that's pretty insane.
essentially, the mechanic of score in the end should be an indicator of which spread of judgements took more skill to hit
in the example of halogen vs. guilhermeziat, i think pretty much anyone can glance at those judgement spreads and say one is better than the other.
that's not being shown in the actual score however.
in the example of halogen vs. guilhermeziat, i think pretty much anyone can glance at those judgement spreads and say one is better than the other.
that's not being shown in the actual score however.
today's match between Japan and Brazil (and a lot of other matches throughout this tourney featuring dense LN maps) shows how ridiculously difficult the HP penalty is for LN maps. Please take this into consideration and rebalance HP losses and gains on long notes for the future.
ScoreV2 in retrospect (similar to my previous long post, you can just read the bolded parts if you can't be assed):
First things first. While some of these issues were pointed out before MWC started, there also wasn't that much time to actually go about fixing these changes. That doesn't mean I'm going to repost these issues here however. Also, I am mostly going to be reposting issues discovered, along with a few other of my own comments, so this will largely be a post combining info.
One of the major issues found, especially pertaining to high level play, comes from the fact that the accuracy component did not differentiate between 300g's and 300's. This resulted in the weights of the non-perfect judgements counting for far more than they should have, as the only advantage a 300g had was in the combo portion of the score, which was typically extremely small. In a previous post, Shoegazer suggested how we could let 300g's be weighted as a 305 in the accuracy portion of the score. (Smaller than scorev1 due to the exponential vs. linear nature of this portion of the score). A naive implementation of this, however, would make SS's pretty much non-existent, which I am not sure is good for the overall state of the game (as getting rid of an entire grade seems like a bad idea). When MWC 7k rolls around next year and people start actually caring about this again, it would be a good idea to start by testing out this change, especially as it significantly affects score in the earlier stages of a tournament.
At the time of my previous post, I mentioned how the ending window of LNs should be loosened a bit. As it turns out, there was never any extra leniency for releases in the first place, and for the most part, it now looks pretty fine. Overall, it is still more difficult to get good accuracy on LNs like before, but that is fine.
However, LNs still run into 2 problems. First of all, the HP drain rate (and health mechanics related to LNs) are generally fucked. This is due to a couple of reasons:
To be perfectly honest, this barely mattered at all during the tournament, as for the most part, the player/team with the better accuracy won. Other than the fact that the notes at the beginning of the song are weighted less than notes after the 400th notes, which can most certainly be fixed in the future, there is not really enough enough experimental evidence, especially with scores that aren't this high, in order to see what effect this really has. This probably has to be put off for a more large scale test in the future, especially considering outliers that have good acc/bad combo, and those that have bad acc/good combo.
First things first. While some of these issues were pointed out before MWC started, there also wasn't that much time to actually go about fixing these changes. That doesn't mean I'm going to repost these issues here however. Also, I am mostly going to be reposting issues discovered, along with a few other of my own comments, so this will largely be a post combining info.
300g's and 300's
One of the major issues found, especially pertaining to high level play, comes from the fact that the accuracy component did not differentiate between 300g's and 300's. This resulted in the weights of the non-perfect judgements counting for far more than they should have, as the only advantage a 300g had was in the combo portion of the score, which was typically extremely small. In a previous post, Shoegazer suggested how we could let 300g's be weighted as a 305 in the accuracy portion of the score. (Smaller than scorev1 due to the exponential vs. linear nature of this portion of the score). A naive implementation of this, however, would make SS's pretty much non-existent, which I am not sure is good for the overall state of the game (as getting rid of an entire grade seems like a bad idea). When MWC 7k rolls around next year and people start actually caring about this again, it would be a good idea to start by testing out this change, especially as it significantly affects score in the earlier stages of a tournament.
LNs
At the time of my previous post, I mentioned how the ending window of LNs should be loosened a bit. As it turns out, there was never any extra leniency for releases in the first place, and for the most part, it now looks pretty fine. Overall, it is still more difficult to get good accuracy on LNs like before, but that is fine.
However, LNs still run into 2 problems. First of all, the HP drain rate (and health mechanics related to LNs) are generally fucked. This is due to a couple of reasons:
- In Scorev1, simply holding an LN wouuld give you a small increase in HP every combo tick. This originally was not present in v2, but this was apparently fixed by smoogi before the tournament. I suspect then, that the main reason for the HP drain being fucked comes from the next point.
- While each normal note can only be "missed" once, you can actually miss a LN up to 3 times. The points where you can miss are namely: the LN start, the LN end, and releasing a held LN (resulting in a HoldBreak or whatever, can't be assed to search it up). In addition, if you don't have the LN start, or break the hold, you automatically reduce the highest judgement for the LN end to 50, hampering HP gain even more. Compare this to Scorev1, where you could only miss twice (one of the contributing reasons to why LN maps are harder to maintain HP in as well).
The Combo Portion
To be perfectly honest, this barely mattered at all during the tournament, as for the most part, the player/team with the better accuracy won. Other than the fact that the notes at the beginning of the song are weighted less than notes after the 400th notes, which can most certainly be fixed in the future, there is not really enough enough experimental evidence, especially with scores that aren't this high, in order to see what effect this really has. This probably has to be put off for a more large scale test in the future, especially considering outliers that have good acc/bad combo, and those that have bad acc/good combo.
Not sure if I understand this correctly, but if I do,
300g has been hit in two ways in scorev2:
1. Not playing a role anymore in multipliers.
The main difference between 300g and 300 other than HitValue was the HitBonus allowing more accurate players to more quickly make up for the score deficit after a mistake. The removal of HitPunishment and HitBonus means there's no difference in the following notes. Your combo multiplier increases the same way whether you get a 200, 300 or 300g. Literally there's a 1/16 raw score difference on a single note between the two judgments, which is in turn affected by:
(as mentioned multiple times)
2. Being confined in the combo portion of scoring.
80% of the score AND the multiplier (the main point!) of the remaining 20% portion have nothing to do with your 300g ratio. This results in a pathetic actual difference between the two judgments, and players with no actual sense of rhythm get scores like this: (forgot mp link, but the others had about 3:1 300g ratio give or take)
hue
Personally I agree with the 300g -> 305 idea, and to preserve the practicality of grades may I dare suggest the ridiculous idea of having TWO accuracy calculated: a displayed accuracy and an actual accuracy. The displayed accuracy does not differentiate between 300g's and 300's while the actual accuracy used in the scoring does. This is kind of like scorev1 though.
300g has been hit in two ways in scorev2:
1. Not playing a role anymore in multipliers.
The main difference between 300g and 300 other than HitValue was the HitBonus allowing more accurate players to more quickly make up for the score deficit after a mistake. The removal of HitPunishment and HitBonus means there's no difference in the following notes. Your combo multiplier increases the same way whether you get a 200, 300 or 300g. Literally there's a 1/16 raw score difference on a single note between the two judgments, which is in turn affected by:
(as mentioned multiple times)
2. Being confined in the combo portion of scoring.
80% of the score AND the multiplier (the main point!) of the remaining 20% portion have nothing to do with your 300g ratio. This results in a pathetic actual difference between the two judgments, and players with no actual sense of rhythm get scores like this: (forgot mp link, but the others had about 3:1 300g ratio give or take)
hue
Personally I agree with the 300g -> 305 idea, and to preserve the practicality of grades may I dare suggest the ridiculous idea of having TWO accuracy calculated: a displayed accuracy and an actual accuracy. The displayed accuracy does not differentiate between 300g's and 300's while the actual accuracy used in the scoring does. This is kind of like scorev1 though.
Where can I see the current characteristics of scoreV2 ? The OP is outdated and I honestly don't want to read through the 25 pages of this post
Also why does FlashLight give a multiplier but FadeIn doesn't when there are some people who perform better with FL than NoMod but not a single one in the world who performs better with FI than NoMod? (I can tell by the "global rankings with active mods" that I am the only one in the world who uses FI for topscores, and even I only do it because it's a fun challenge, it's actually a handicap to my performance)
Also why does FlashLight give a multiplier but FadeIn doesn't when there are some people who perform better with FL than NoMod but not a single one in the world who performs better with FI than NoMod? (I can tell by the "global rankings with active mods" that I am the only one in the world who uses FI for topscores, and even I only do it because it's a fun challenge, it's actually a handicap to my performance)
I think you are like 3 months late. Pretty sure you can use the search function here.LastExceed wrote:
Where can I see the current characteristics of scoreV2 ? The OP is outdated and I honestly don't want to read through the 25 pages of this post
Also why does FlashLight give a multiplier but FadeIn doesn't when there are some people who perform better with FL than NoMod but not a single one in the world who performs better with FI than NoMod? (I can tell by the "global rankings with active mods" that I am the only one in the world who uses FI for topscores, and even I only do it because it's a fun challenge, it's actually a handicap to my performance)