forum

[resolved] Revise tied-scores PP calculations. (discuss)

posted
Total Posts
35
This is a feature request. Feature requests can be voted up by supporters.
Current Priority: +1
Topic Starter
[suicide
Why?
To make PP distribution fair.

http://osu.ppy.sh/p/beatmap?b=113749&m=2
24 people are rewarded with a #1 ranking in that song because they're all the same score, so if you got #24, you would get the same amount of PP if you were early and got #4 in that song. Tied scores = same rank = same reward for the rank. (depending on your PP ranking)
If that doesn't make sense, I can clarify further.

Here are a couple more examples.
http://osu.ppy.sh/b/65417?m=2
http://osu.ppy.sh/b/94845?m=2

And the list is VERY big, Peppy is aware that there isn't suppose to be this many tied scores.

My suggestion is to temporarily have 1 missed droplet = #2 record instead of it being the record stacked behind however many people SS'd the song.
2 droplets missed = #3 record and etc. (Until there's a better solution to the problem)

Here's a more thorough explanation
Since all of the Red scores are counted as a #1 record in PP calculation, why not have the blue counted as #2 instead of #25?

In technical terms, if you SS the song it would say you would have #1, but really you may be 24 or something. You would be rewarded with a #1 record in PP (depending on your ranking), but if you get a 99.92% in the song, you wouldn't have #2, you'd be whatever rank before the last person who SS'd the song.

There's a problem here. There's a BIG gap in rankings you can possibly obtain. the gap being #2-24, it's Impossible to obtain those places.

Either have a better variance to eliminate this, or calculate the PP differently for tied-scores.


Think of suggestions to fix this please. I would like a mod feedback as well!
Soaprman
Those maps probably have no spinners in them. If you look at the scores for standard, the 100% acc SS nomod scores are all the same. There are more of these scores in CTB because CTB doesn't have different accuracy possibilities with each object... you either caught the fruit or you didn't. Not much anyone can do about that.
Topic Starter
[suicide

Soaprman wrote:

Those maps probably have no spinners in them. If you look at the scores for standard, the 100% acc SS nomod scores are all the same. There are more of these scores in CTB because CTB doesn't have different accuracy possibilities with each object... you either caught the fruit or you didn't. Not much anyone can do about that.
um yes I suggested a way to do something about that.

I also have another suggestion that nobody would like.
Have catching accuracy ! having the fruit being caught in the middle of the plate, any fruit caught on the edges would be badd.
There are a few things bad with that idea.
BUT HEY IM BRAINSTORMING.
Kurokami

Manchineel wrote:

Have catching accuracy ! having the fruit being caught in the middle of the plate, any fruit caught on the edges would be badd.
Hell no. When you jump/hyperjump you can't controll the ryuuta the way you want. Except if you fuckin' badass. Therefore there are fruits which ones nobody will able to catch perfectly. This add another rule to the ranking criteria and so on. Bad idea.

The OP is also. Who cares if he/she got a tied SS? At least able to maintain that rank. And also most of the maps are contains at least one spinner, just because you able to show some example, not means it is true in most cases. And also, if you take a look at a PP ranking, you will see, ppl with equal pp is rare and most cases they are not equal in skills too or something. no support.

Edit: I recently realize I wrote slider instead of spinner. o.o Corrected.
Drafura
Accuarcy is actually a joke for CtB.

Anyway, I don't understand this :
Sliding scale, so sliding multiplier.
Example: From not so hidden .3 to very hidden .6
From a little harder .6 to much harder .12
And so on.
Or maybe go map CtB diffs and use sliders to make them fun to play ?
Topic Starter
[suicide
Hell no. When you jump/hyperjump you can't controll the ryuuta the way you want. Except if you fuckin' badass. Therefore there are fruits which ones nobody will able to catch perfectly. This add another rule to the ranking criteria and so on. Bad idea.

The OP is also. Who cares if he/she got a tied SS? At least able to maintain that rank. And also most of the maps are contains at least one spinner, just because you able to show some example, not means it is true in most cases. And also, if you take a look at a PP ranking, you will see, ppl with equal pp is rare and most cases they are not equal in skills too or something. no support.

Edit: I recently realize I wrote slider instead of spinner. o.o Corrected.
There are hundreds of maps like this, do I have to find them all? http://osu.ppy.sh/p/beatmap?b=73193&m=2 #1 rank PP rewarded for 15 people.
Also I was thinking about that hyper jump, could be exempt from the thing, and I told you, this is only an idea of how we could stop the problem, I would like to hear others.

Drafura wrote:

Accuarcy is actually a joke for CtB.

Anyway, I don't understand this :

Or maybe go map CtB diffs and use sliders to make them fun to play ?
For example you enable FLASHLIGHT mod, but then only use half the mod. so you would get a .6 multiplyer and would be able to barely see the fruits falling down. I don't really see this as a good idea, but Im desperate for ideas and I need new ones.
Kurokami

Manchineel wrote:

For example you enable FLASHLIGHT mod, but then only use half the mod. so you would get a .6 multiplyer and would be able to barely see the fruits falling down.
What? O.o

I know there is more mapset out there but still, one from every 10th? That isn't too much.

On the other hand, kinda annoying to see equal scores worth the same pp. Especially when one of them (2 player) is a pro (around #100 with pp) and for the other player took time to get that place. But its hard to do something agains this, so maybe put "[discussion]" in the title.

I was thinking about something for CtB only maps. Maybe, if we set a rule "a map must contains a spinner If it is fit somewhere". But this will only work for, as I said, CtB specific maps since there is no exact ranking criteria yet.
Mithos
Just suggest that every ranked ctb needs a spinner.
TheVileOne
How about we make maps that have too many tied ranks worth less pp? If so many people can tie in it, it obviously isn't a very significant performance.

I thought of this without any balancing in mind:

pp = amount/# of people tied at rank.
Drafura

Mithost wrote:

Just suggest that every ranked ctb needs a spinner.
There's a rule/guideline thread to discuss about this. And we allready try to find spot to add spinners when modding a map with 0 spinners or only small spinners.

For example you enable FLASHLIGHT mod, but then only use half the mod. so you would get a .6 multiplyer and would be able to barely see the fruits falling down. I don't really see this as a good idea, but Im desperate for ideas and I need new ones.
That would be the same... people using Half-FL mods will get the same score as HD, DT and other Half-FL scores....
benguin
It seems people are oversimplifying how the pp calculation works. People that get the same score on the same song DON'T gain the same amount of performance points, you mostly gain pp from geting scores that are a little better than what you normally get. For example, someone in CtB who is in the top 20 for pp ranking will probably get 0 pp for a no mods normal SS while someone who just started playing the game might get like 30 pp for that. So I don't see the problem here, if someone you consider a "noob" works hard and pushes themselves to earn that tied SS, give them the reward they deserve!
Drafura
This doesn't change the fact that getting tied scores on ctb happend too frequently when there's no spinners or only short spinners benguin.
benguin
True, I just didn't like how the OP made it seem like PP is the reason why score variance is a problem in the first place. If we take pp out of the equation, why is score variance such an issue in the first place? I don't see why there should be concern for it.
Drafura
Score variance make the score calculation more accurate on how well you played the map. ofc this is hard to judge (and I think this will lead to nothing) since ctb doesn't have a good accuarcy system wich really affect your score (10 point non multiplied for a droplet ... getting top40 with a D is so easy). The only thing could fix this imo is the catching accuarcy (get the fruit in the middle of the ryuuta), and this will never come, so many people doesn't want it, a lot of maps will become impssible to SS, and it's going to be hell for ctb modding to test if a map is SSable.
TheVileOne
Lets ignore these maps and play the maps that actually are balanced. Problem solved.
benguin
I think I see your point now. However, it's often the case that easy maps with no spinners often have this problem with lots of tied scores and you don't get pp anyways from playing maps that are easy for your skill level so I don't see this hurting pp too much. But I guess it is still somewhat of an issue since people enjoy trying to rank in on songs regardless of difficulty just for fun. I don't see this issue of score variance as a crisis though, so I think other feature request should have higher priority, at least until someone comes up with a viable solution for this.

As for having a judge based on catching the fruit in the center of the catcher, that's not going to work imho. Part of the being able to play this mode well has to do with having the skills and abilities needed to learn how to catch fruit on the very edges of the plate; in several maps (perhaps most), there is at least one fruit that if you catch it in the middle of the catcher, it's pretty much a guarentee that you'll miss the next fruit.
Laharl

Manchineel wrote:

More than 50+ people get the same amount of PP in that song because they're all the same score, so if you got #50, you would get the same amount of PP if you were early and got #4 in that song. Tied scores = same rank = same PP reward.
If that doesn't make sense, I can clarify further.
I don't really get what's wrong about that.
MrSake
Maybe this would work, if the catching location wouldn't affect to accuracy but only to the score. This way maps would be SS-able like now, but there would be some score and rank difference between players who got the same combo and accuracy.

Well, old scores would be problematic, because if the default fruitscore in them would be max, nobody could beat those, and if it is minimum, beating those would be maybe too easy.
Drafura

benguin wrote:

I don't see this issue of score variance as a crisis though, so I think other feature request should have higher priority, at least until someone comes up with a viable solution for this.
Well, yes.

@Sake : Pretty good idea.
TheVileOne
That is a good idea. But I don't know if I would want such a big change. I mean if we're talking about consistency, it would be even more frustrating to get perfect score, and it would be annoying to have a mod used score with SS not beat a no mod score with SS only because I was more sloppy in the "harder" mod. I think there should be an adjustment for leniency for mod increases like mania has.
DJ Angel
stop QQing and rank too, it isn't hard to improve pp on ctb because of the decreased number of players, meh

edit: >I'm not a SS player
nvm, I take my words from before, you suck and you don't want to improve your scores, which is what pp is for (improve ranks and not a ''okay'' rank)
please, improve before talking about other people that can do what you can't

and...the suggestion is retarded not because of the suggestion itself, but because of the examples and facts you gave

Not talking about the hidden scalling itself, but you are giving the suggestion as your egocentric point of view.

"I can't do SS so I don't earn PP and don't appear on the tied scores, but others earn PP because they actually could do it."
so, you made this mostly useless thread to express your opinion

I will be objective, the idea isn't bad but the way you putted it was retarded, and calling someone pp abuser is a imature act of jealously which you should be ashamed of, despite that, my suggestion is first, stop thinking that it was easy for him to make those ranks. I myself were a ctb player and did really lots and lots of tied scores, which weren't easy at all to achieve.

Second, you aren't trying enough. You said you had problems with SS, what's about to actually try more than 5~10 times?

And third, clean the main page, keep 1 or 2 beatmaps as a actual example of tied SSH scores but take all of the rest off, for the sake of the suggestion, too.

ps: Might actually star this if my last requirement is fulfilled
Topic Starter
[suicide

DJ Angel wrote:

stop QQing and rank too, it isn't hard to improve pp on ctb because of the decreased number of players, meh

edit: >I'm not a SS player
nvm, I take my words from before, you suck and you don't want to improve your scores, which is what pp is for (improve ranks and not a ''okay'' rank)
please, improve before talking about other people that can do what you can't

and...the suggestion is retarded not because of the suggestion itself, but because of the examples and facts you gave

Not talking about the hidden scalling itself, but you are giving the suggestion as your egocentric point of view.

"I can't do SS so I don't earn PP and don't appear on the tied scores, but others earn PP because they actually could do it."
so, you made this mostly useless thread to express your opinion

I will be objective, the idea isn't bad but the way you putted it was retarded, and calling someone pp abuser is a imature act of jealously which you should be ashamed of, despite that, my suggestion is first, stop thinking that it was easy for him to make those ranks. I myself were a ctb player and did really lots and lots of tied scores, which weren't easy at all to achieve.

Second, you aren't trying enough. You said you had problems with SS, what's about to actually try more than 5~10 times?

And third, clean the main page, keep 1 or 2 beatmaps as a actual example of tied SSH scores but take all of the rest off, for the sake of the suggestion, too.

ps: Might actually star this if my last requirement is fulfilled
I don't know how to put this in a way in a better way. if somebody could put it in a way that users would put their stars worth that would be very nice. I see some people agreeing with whats in this thread, but not a lot of agreeing with what I've put. Please edit it, and I'll copy pasta. I really don't mind.
Lets make this a community effort.
Topic Starter
[suicide
bump
I have a new idea/ edits to the thread/
Since all of the Red scores are counted as a #1 record in PP calculation, why not have the blue counted as #2 instead of #25?

In technical terms, if you SS the song it would say you would have #1, but really you may be 24 or something. You would be rewarded with a #1 record in PP (depending on your ranking), but if you get a 99.92% in the song, you wouldn't have #2, you'd be whatever rank before the last person who SS'd the song.

There's a problem here. There's a BIG gap in rankings you can possibly obtain. the gap being #2-24, it's Impossible to obtain that place.

Either have a better variance to eliminate this, or calculate the PP differently for tied-scores.
Soaprman
There's not a problem there at all. The blue scores are #25 because 24 people did better than they did.
Kitokofox
What he is trying to imply, is that all the players in the red group got the same score with the same mods, and should therefore get an equal amount of PP for their equivalent efforts, instead of being late and getting maybe the 15th hidden double-time SS and not getting as much PP as that who got the score first. So he is insisting that the red group, which is one score step lower, would be treated as if they got rank 2 instead, to sort of make it fair for those who came in late as well as more fair for the larger tie below. This seems like an issue that would apply to all maps that lack spinners, CtB more than other modes. There is no real way of making this sort of play fair, because if someone gets maybe a mod high combo or misses some droplets, some players would have slightly less score and therefore "add" to the number of ties. It's very sketchy to do something like this. What I would probably consider is have a player get the PP as if they were on top of the tie, which is probably what osu! does already, so I'm not exactly concerned. I'd feel that the PP would slightly degrade itself because of such a massive tie, adding an ease-to-get factor. But this ease-to-get factor would already exist in the "Get a tie, get PP for top of the tie", since if there's an even larger tie above, it means it should be theoretically "easier" to join that higher tie, and so therefore you'd get less PP for being in said lower tie. This means if you just fail getting that SS and get a score between the ties, you'll get that rank and PP between those ties. Simple as that.

Long story short, I think the system is fine just the way it is. Even better if it does what I thought it did.
Topic Starter
[suicide

Kitokofox wrote:

What he is trying to imply, is that all the players in the red group got the same score with the same mods, and should therefore get an equal amount of PP for their equivalent efforts, instead of being late and getting maybe the 15th hidden double-time SS and not getting as much PP as that who got the score first. So he is insisting that the red group, which is one score step lower, would be treated as if they got rank 2 instead, to sort of make it fair for those who came in late as well as more fair for the larger tie below. This seems like an issue that would apply to all maps that lack spinners, CtB more than other modes. There is no real way of making this sort of play fair, because if someone gets maybe a mod high combo or misses some droplets, some players would have slightly less score and therefore "add" to the number of ties. It's very sketchy to do something like this. What I would probably consider is have a player get the PP as if they were on top of the tie, which is probably what osu! does already, so I'm not exactly concerned. I'd feel that the PP would slightly degrade itself because of such a massive tie, adding an ease-to-get factor. But this ease-to-get factor would already exist in the "Get a tie, get PP for top of the tie", since if there's an even larger tie above, it means it should be theoretically "easier" to join that higher tie, and so therefore you'd get less PP for being in said lower tie. This means if you just fail getting that SS and get a score between the ties, you'll get that rank and PP between those ties. Simple as that.

Long story short, I think the system is fine just the way it is. Even better if it does what I thought it did.
no that's not what i'm trying to say.
People in the red group get rewarded with a #1 record and is distributed with PP depending on their rank however, if you miss one droplet then you get rewarded for placing a #25 record in that song, and not obtaining #2 record.
Topic Starter
[suicide

Soaprman wrote:

There's not a problem there at all. The blue scores are #25 because 24 people did better than they did.
The problem is they technically placed #2 because everyone above them is #1 according to the system, however they're not rewarded as having a #2 record, the blue group is all rewarded as having a #25 record.
Kitsunemimi

Manchineel wrote:

The problem is they technically placed #2 because everyone above them is #1 according to the system
How do you know this is the case?
Topic Starter
[suicide

Kitsunemimi wrote:

Manchineel wrote:

The problem is they technically placed #2 because everyone above them is #1 according to the system
How do you know this is the case?
When you tie a score, it actually tells you you're #1 in-game.

In this picture it tells me i'm #25 because I have the same score as that person. If you were to SS that song, you would be displayed as #25 because it goes in order that way, but it would tell you that you were #1 in that little circle.

Kitsunemimi
That's just what it has ALWAYS displayed in game, and it has no correlation to what it's processed at server-side.

I'm guessing that everyone is awarded pp based on whatever rank they achieve, time of score factored in.
peppy
Maps with ties like that already result in very low pp yield.

ALso please do not PM me with links to feature request threads. It gets old. I read this forum and answer where appropriate already.
Topic Starter
[suicide

peppy wrote:

Maps with ties like that already result in very low pp yield.

ALso please do not PM me with links to feature request threads. It gets old. I read this forum and answer where appropriate already.

Very low PP yield?? That's ridiculous because this user is yielding GREATLY on ranking with tied-scores.
http://osu.ppy.sh/u/1267642

ALL of his top ranks are tied plays! which means he's getting a lot of PP.
and he's a pretty high rank.
benguin
I still don't see the problem here. It's not like anyone is being disadvantaged by tied-scores gving the same/less/more/whatever pp. If you are annoyed that someone is getting pp for a tied score; well there is nothing stopping you or anyone else from getting that same tied score, everyone has that same opportunity. It really seems like we're making a big deal out of nothing in this thread imho.
Tsukimi Luna
Well there isnt much of a problem here since maybe... 1 out of 1000 songs are like that
Topic Starter
[suicide

benguin wrote:

I still don't see the problem here. It's not like anyone is being disadvantaged by tied-scores gving the same/less/more/whatever pp. If you are annoyed that someone is getting pp for a tied score; well there is nothing stopping you or anyone else from getting that same tied score, everyone has that same opportunity. It really seems like we're making a big deal out of nothing in this thread imho.
It's just that the people who didn't SS the song deserve to be the real rank before #1, that way they get at least some PP for being #2, instead of say of the system sees them at say rank 25.

what's the point of retrying a difficult song if you missed one droplet? we want to move on the next song or it gets boring.

Also, it's not 1 out of 1000
that's exaggerating, Because I can list hundreds if I wanted to.
And if it were asked for, I WOULD DO IT.
Please sign in to reply.

New reply