What I meant was that the issues with Score V1 have been known long before PPv2 was conceived, and this was the reason they decided not to let it influence PP in any way. Considering that osu! still uses the score system to determine your "best score", there's no other reason to exclude it if it weren't for its flaws.

So Do you prefer ScoreV1 or ScoreV2 ??

Vote here : https://goo.gl/rkPSDM

Result here : https://goo.gl/OQo1Wa

Seijiku wrote:

:?: So Do you prefer ScoreV1 or ScoreV2 ??

Vote here : https://goo.gl/rkPSDM

Result here : https://goo.gl/OQo1Wa

And this is what we call absolutely pointless

Seijiku wrote:

:?: So Do you hate change or do you blindly accept the future ??

fixed

...Seriously though, we have no idea how ScoreV2 is actually going to be. This is pretty pointless.

I think slider ends should break combo like slider ticks do.

AFAIK, they're the only gameplay element that adds to combo but doesn't break combo. That isn't consistent.

It doesn't make sense that you can play through a song without a combo break OR full combo

JTF195 wrote:

I think slider ends should break combo like slider ticks do.
AFAIK, they're the only gameplay element that adds to combo but doesn't break combo. That isn't consistent.
It doesn't make sense that you can play through a song without a combo break OR full combo

Objectively, I think you're right, but good luck convincing people to change that lol

JTF195 wrote:

I think slider ends should break combo like slider ticks do.

AFAIK, they're the only gameplay element that adds to combo but doesn't break combo. That isn't consistent.

It doesn't make sense that you can play through a song without a combo break OR full combo

I like leniency-abusive slider maps the way they are.

JTF195 wrote:

I think slider ends should break combo like slider ticks do.

AFAIK, they're the only gameplay element that adds to combo but doesn't break combo. That isn't consistent.

It doesn't make sense that you can play through a song without a combo break OR full combo

According to this recent reddit thread: https://www.reddit.com/r/osugame/comments/518xwg/since_notch_hell_is_getting_ranked_slider_ends there's also the fact that slider ends are apparently checked 36ms early. I'm sure there's a reason for that, but I'm not sure what it might be, or if it might be worth revisiting


To clarify, I'm not saying that either of those mechanics should be changed in standard gameplay anytime soon, or ever.

But, ScoreV2 and/or osu!next present an opportunity to fix those inconsistencies without affecting the existing experience.

Except for maybe spinners being worth too much, I think scorev2 is vastly superior to scorev1 for both tournaments and singleplayer. Consistency is not only maintaining combo. I consider a player who can SS everything and get a shitmiss in the middle of the map more consistent than someone who just full combos with very low acc, and scorev2 rewards that much more than scorev1. It's just way more logical. Scorev2 can also create more hype for tournaments as there's always the chance for a comeback unlike combo-based scoring, especially for 1v1 (which IMO makes more sense than team-based tournaments).

And no, scorev2 doesn't overrate accuracy in my opinion. Low acc does not only come from being bad at acc. Barely aiming, running out of stamina, misreading etc can all lower your accuracy.

I might not be against any of the changes except those which are about sliders.

The current slider settings make the maps nowadays varied, so if any of the changes is about to make it more strict, I'll definitely disagree. (loosen it is dumb af too but I know you won't do that.)

should better check the forum page here and I think it's definitely worth to be discussed as well: t/427264

I would rather use pp over scorev2. On the same map, pp can distinguish the value of scores well.

As someone who has played watched my team play in many tournaments, I can say that seeing scorev2 in a tournament made most of us roll our eyes. It just seems like random points pulled out of thin air. For spectators it may seem more intense because the scores are generally much closer throughout the match, in reality you could attain the same thing by dividing scorev1 by 10. The added appeal of scorev2 to spectators is the "hype" when two teams are closely tied. The issue with scorev2 is that the scores aren't really close, it's just a clusterfuck of confusion over who will win because it doesn't make any sense half the time. It's just misleading to spectators and annoying.

Much prefer scores v1 over score v2. You should consider using pp based scoring and see how that looks.

^ The thing is, I'm pretty sure that PP functions more like a post-play calculation, so I don't know how well something like that would fare in a real-time environment. If anyone found a way to implement a score system that used it while still keeping scores cumulative (as in, misses count as 0 rather than causing players to lose points), I could see something like that working out.

The difference is that pp only looks at your highest combo while in score v2 you will still receive some score for additional lower combos.
I think it would look like bullshit if players would suddenly stop getting points because they have to get back to their highest combo to have the multiplier roll out again.

To use the pp formula for score:
1) During play, use the pp formula assuming everything not played yet is a Miss.
2) Modify the formula so the values go up more linearly during play (or another kind of growth that is easier to follow during play).

The pp formula is (with ScoreV2 Slider mechanics):
Length = Amount of notes in the map.
Acc% = Accuracy percentage, range: [0,1]. (6 * Count300 + 2 * Count100 + Count50) / 6 / (Length)
AimBase = A map constant that depends on the Aim Difficulty, Length, AR, OD, Visual Mods.
SpeedBase = A map constant that depends on the Speed Difficulty, Length, OD.

ODBase = A map constant that depends on OD.
AccBase = A map constant that depends on Length and Visual Mods (not OD)

MaxComboRatio = The maximum combo achieved divided by the maximum combo possible on the map.

pp = Constant * (AimValue^1.1 + SpeedValue^1.1 + AccValue^1.1)^(1/1.1)

AimValue = AimBase * 0.97^CountMiss * MaxComboRatio^0.8 * (0.5 + Acc% / 2)
SpeedValue = SpeedBase * 0.97^CountMiss * MaxComboRatio^0.8 * (0.5 + Acc% / 2)
AccValue = AccBase * ODBase^(Acc%^24)


If all Values (Aim, Speed and Acc) have roughly the same magnitude (they should with end-of-play values), and each one increases linearly during play, then score increases linearly during play. So we have to modify each of the Value formulas in a way they have the same magnitude at the end of play, and scale linearly during play.

ProgressRatio = Amount of played notes divided by the total amount of notes.


Note: While using those formulas, assume every note not played is a miss, for example: CountMiss = Number of missed notes so far plus number of notes remaining. MaxComboRatio = Longest combo so far divided by Length. The values increase linearly as long as the performance through the map is consistent.

FAimValue = AimBase * ProgressRatio^(1/5) * 0.97^( (CountMiss + Length * (ProgressRatio - 1) ) / ProgressRatio) * MaxComboRatio^0.8 * ( 0.5 + Acc% / (2 * ProgressRatio) )

FSpeedValue has the same form as FAimValue.

FAccValue = AccBase * ProgressRatio * ODBase^( (Acc% / ProgressRatio)^24 )

Then, the Constant of the pp formula can be raised so the values shown on score are higher.

sure, pp scoring is fine too and it's also pretty trivial to implement as Full Tablet pointed out. ScoreV2's slider acc would also nicely punish people who are sloppy on slider maps.

Only thing I'm against is slider accuracy

oh, as for the slider changes, I think they're great! with the current system a lot of accuracy is wasted on slider heavy maps and it's a real shame. players are less sloppy than you might think on sliders. it pains me to see bad unstable rate scores getting the same acc % as actually accurate plays. it will also prevent the development of bad slider habits in new players.

Franc[e]sco wrote:

it will also prevent the development of bad slider habits in new players.

Pretty much this. Sliders should never have ignored accuracy to begin with, IMO. It just promotes bad habits.

The only problem with slider acc to me is that, right now, they are great for BPM changes as you do not have to be nearly as close, which would feel awkward and ruin acc if sliders had accuracy. Imo, the best way to handle this is make sliders 1.5x or even 2x as lenient as circles so things such as BPM changes would still require acc without being unfair.
Also, if you add slider acc, rip all maps with variable timing


Edit: Another solution would, as stated before (I don't remember who said it tho ), make slider acc a diff-increasing mod. That way, if there are variable timing maps, you aren't forced into extremely difficult acc due to timing changes.

I'm really fond of slider accuracy, but the score out of 1 million to me is a huge problem for ScoreV2.

On an average map, it can work, but very drastic changes happen to the scores of the players when you get into longer maps such as the marathons typically used in a tiebreaker. Due to the length of the map, it becomes significantly more difficult for players to hold combo, and as seen even in the tiebreaker of the grand final of OWC last year, if the map is also difficult, accuracy ends up being far more important towards your end score than combo a majority of the time. This is even with the current score ratio of 7:3 combo:accuracy.

I guess having max combo score scale in some way with then length of a map would help? Maybe linearly though, so it doesn't throw off the balance too in favour of combo once again.

Looking nice! But what about the multiplier on difficult reduction mods such as EZ?

^
They discussed about it here. p/5570799

This isn't really a huge problem but I'm going to complain anyway XD

If only 500 points are awarded per tick in a spinner, I don't think I'll feel as satisfied when I do spinners lol.

But I really would love more accuracy contributing to the score over combo because its a rhythm game! Osu is virtually the same as playing a musical instrument; what's the point of playing a piano or a guitar in a consistent rhythm, but playing the wrong notes? And its a real pain to try and beat low-accuracy but high combo scores.

I'm looking forward to score v2. Good luck with it!

Just popping in here to inform everyone that ScoreV2 is now available as an unranked(?) mod on cuttingedge. Might be worth giving a try.

Is there anything somewhere realy explaining how Score v2 works ? I've noticed when trying on Cutting Edge that my score was actualy decreasing when I was doing 100's which feels very unnatural. Score v2 plays seams to be failing to be submitted too but score v2 isn't considered as UNRANKED by the game client.

Slider accuracy is another topic as well, I just know I'll probably lose more than 1% of accuracy once score v2 is the only thing used because of it. Changing a very big game mechanic when the game is almost 10Yo with some players that are active since many years and used to how it works is a bit wierd.

I'm really enjoying the new slider accuracy judgement, it really keeps you on your toes more than the old system. It hits the hardest on wubwub maps like Bassdrop Freaks because now you actually have to pay attention to the slider patterns instead of zxzxzxzing your way through.

But I'm not sure if there's any replacement hitobject for mappers that want to hide a tempo change, other than being able to configure the OD through timing sections I take that back - just played Roze in Scorev2 and trying to account for all the changes in the tempo was actually really fun. Got 10 100s and scored a 606,074

I'm not sure I like how spinners are scored. Spinners add to the maximum score of 1,000,000 points, which means that the 1 million figure is kind of arbitrary. On maps without spinners an SS gets 1 million points, but on maps with lots of spinners you can score more than 1,000,000 points even if you didn't SS, so the 1 million figure loses its meaning.

But I can't think of a better way to score spinners other than making them not count at all, which is dumb. Maybe the difficulty to clear a spinner could be MASSIVELY increased so that it's actually really hard to spin a 300, and then remove spinnerbonus? Although that's an even worse solution. Idk.

quaternary wrote:

But I can't think of a better way to score spinners other than making them not count at all, which is dumb. Maybe the difficulty to clear a spinner could be MASSIVELY increased so that it's actually really hard to spin a 300, and then remove spinnerbonus? Although that's an even worse solution. Idk.

The only thing I could think of would be to count the spinner bonuses toward the 1mil cap (as spinners are capped at 477rpm) and adjust the other values accordingly, but that'd probably make it even more arbitrary, as it'd make it impossible to get 1mil without a spinbot.

What will happen to fallback? Is it going to be implemented to fallback at least at any point in the future?

Probably been pointed out but if you are going to add a judgment to the slider head you should add some immediate feedback. At the moment if I hit a 100 on a sliderhead I won't know until the end. Having feedback on the slider end is still important but less so. Perhaps just having separate judgment/feedback for sliderhead vs sliderbody (includes tail) judgment. While this would further inflate how much of your score is dictated by sliders it seems fairly standard (ie. lazers in KSM are overweighted in terms of score and I'm pretty sure this applies to SDVX too).

I think giving the player feedback (note: feedback =/= actual judgment) on all objects and not just circles will lead to an overall creep in accuracy/consistency as it reduces the feedback bias player currently receive.

Anyways, my post is less about whether SV2 is a good thing and more about implementation so hopefully it's useful.

Serious suggestions:

1. After playing ScoreV2 in several types of map, from simple pp farming with almost no sliders, ending on technical slider spam maps, i would recommend to increase leniency of sliders by around 25-30% in od8 up to 50% in od10.

2. Move slider score indicator (300-100-50-miss) into slider head circle instead of slider-end.

Rest is fine in my opinion, with change like above, this could already rollout into live version.

With regards,
Love ScoreV2, game feels so much better with it, with exceptions like above.

den0saur wrote:

What will happen to fallback? Is it going to be implemented to fallback at least at any point in the future?

I'm pretty sure fallback hasn't realy been updated since 2015 (Copyright notice on fallback still says 2007-2015) hidden already works differently and i'm almost sure score v2 will never be added to fallback. You should switch to stable, honestly the only computer where I saw fallback working and stable not is a Pentium 4 computer that is probably 15 years old.

About score v2 itself, I don't get why sliders has been made much more punishing. Not only you need to accurately hit the slider start, but when sliderbreaking by missing the start of a slider, you get a 50 instead of a 100. Missing ticks still work the same as before it seams
Also I'm all for having a score indicator when clicking the slider start, then another one for slider end. Cause now, if you get a 50, you don't realy know, did you hit the start very innacurately or did you miss the start and sliderbroke. Same for 100's and missed slider ticks. Also seing a 100 at the end of a slider on which you didn't break is very unusual, especialy since the slider end will obviously sound on beat. Having a 100/50/Miss when clicking the slider start to know how accuratly you did hit it, then another indicator at the end if you missed ticks or the slider end whould probably be a good idea.

Other than that, score v2 still seams to rely heavily on combo and some people said it tooday and I agree, it should probably be based on the number of misses, cause why the guy who had 900 combo but 20 misses should have a higher score than the one with a single miss at 500 combo in the middle of the map (With similar accuracy of corse), was not that the biggest problem with score v1 ? Right now I think the adventage to score v2 is when you are doing 100's doesn't matter, but it still rely too heavily on combo probably.

den0saur wrote:

What will happen to fallback? Is it going to be implemented to fallback at least at any point in the future?

Fallback was only meant to be for people who had issues with the big OpenGL update of 2015, and since most of the compatibility issues with the current 'latest' version have been ironed out, I don't think it'd ever be updated. There's a new client on the horizon, and it wouldn't surprise me if the current 'fallback' gets disconnected from online services and the current 'latest' becomes the new 'fallback' when the new client gets released.

I personally think scorev2 is some kind of an joke the game will become boring almost linear it will become harder too so less new players will be playing it. and losing points when u dont get 300 thats bullcrap. so if this becomes the new score system im out. now sorry i need to not write this and enjoy osu while its good.

Edit: And basically the game will become osu!mania in the points way

First off, I have no idea how score in osu!mania is a bad thing when applied to osu! standard. This complaint I get the least, because it changes literally nothing about the core gameplay and is really only a "muh spreadsheets" kind of thing, except replace spreadsheets with numbers. It's the same thing.

The real controversy lies within the slider timing window. Personally, I don't see how this is a big issue, as osu! should've had something like this since the very beginning. As it's advertised, osu! is a rhythm game. Woah. This means that playing the song accurately should reward the player, and with sliders now finally getting timing treatment, players need to actually click to the rhythm of the song as opposed to click whenever. If you don't hit a slider accurately, you deserve less accuracy. Honestly as a rhythm game player fanatic, I'd love to see accuracy showcased more, but because osu! also has aiming and combo as features, I can't take those away. The split right now is good, though a bit more of an increase for accuracy wouldn't hurt at all.

Another thing people need to realize is that this is not the final product. Yes, currently it is implemented in the game as a mod, but this is just to showcase what the future entails and to test the new scoring system for those willing to try it out. When they have constructive criticism against ScoreV2, discussion is raised, and depending on whether other people agree or not, things can be tweaked. No system will ever be perfect, but to get as close to perfection as possible is always the goal, and I believe ScoreV2 honestly is making an attempt at this better than ScoreV1 ever could. Because ScoreV2 isn't complete yet, and can be tweaked (as it has been when large outcries happen such as with Taiko big note hits/mania taking combo more into account were either tweaked or outright removed).

Score V2 is broken right now and SHOULD NOT be put out

If someone is able to do this \/


Then obviously the system should be reworked.... ALOT

I mean how can a .99 star get more score then a song that is near 11 stars.... that is not even logical anymore

Oh and touching on the sliders just a bit...
ima just post what my friend said cause he said it quite well tbh


Now i am not like the other people saying that it just flat out should never come out.... but i mean..... there are clearly some major flaws........

if you do not want people to get triggered do not put on the desc of the mod "try the future scoring system", that lowkey makes it sound final and like NOTHING will be changed, make it something like "Test out the new scoring system". that way people know that its a TEST

The score is supposed to indicate how well you played the map, not indicate how good the play was (that is pp's function).

In A FOOL MOON NIGHT, the play gets all 300s (awarding the base 1 million points + mod bonuses); while in Syrup, it also gets all 300s (awarding the million + mod bonuses), plus score from overspinning.

honestly, i think accuracy is already overrated. the pp difference between 94% and 97% is already huge. i don't enjoy playing a map 200 times to get it around 97% and finally getting a bit of pp. when i get an s, i just want to move on to a different map. not play the same map over and over to get some pp.

Full Tablet wrote:

The score is supposed to indicate how well you played the map, not indicate how good the play was (that is pp's function).

In A FOOL MOON NIGHT, the play gets all 300s (awarding the base 1 million points + mod bonuses); while in Syrup, it also gets all 300s (awarding the million + mod bonuses), plus score from overspinning.

The point with puting those scores was to show how broke the system was. A map with such a low combo and skill requirement should not have a higher score amount (even just getting a 300 on the spinner would give you a higher score).

There should be some kind of score balance for higher maps so you feel like higher maps reward you better then lower maps (aka the higher star maps should NOT be 1m when you can get 1m on a 1 star)

gellandor wrote:

Full Tablet wrote:

The score is supposed to indicate how well you played the map, not indicate how good the play was (that is pp's function).

In A FOOL MOON NIGHT, the play gets all 300s (awarding the base 1 million points + mod bonuses); while in Syrup, it also gets all 300s (awarding the million + mod bonuses), plus score from overspinning.

The point with puting those scores was to show how broke the system was. A map with such a low combo and skill requirement should not have a higher score amount (even just getting a 300 on the spinner would give you a higher score).

There should be some kind of score balance for higher maps so you feel like higher maps reward you better then lower maps (aka the higher star maps should NOT be 1m when you can get 1m on a 1 star)

If you have played a game such as cytus with a max score of 1000000, you should understand this kind of scoring too (although that there's a weighting system between combo and accuracy). Usually a game like that has 3 difficulty level in a song, and each of them maxed out to 1000000. Read the bolded quote words to make sure you understand about what score v2 is.
Otherwise, I think score v2 is pretty good for tournament setting, so that the combo and accuracy balance out. But I won't like to see it in live scoring leaderboard because of:
1. It is sad to remove all the scores from score v1, especially on those old maps where nearly no one wants to play them anymore
2. Score v2 is not really a good score metric in insanely hard map like Mazzerin's, where combo takes near zero role on score v2.

I don't have an idea if a separate leaderboard can be used (without deleting score v1 of course).

I think ScoreV2 is a great idea, I don't understand why so much people are agaisnt it