forum

Performance Points feedback and suggestions (Standard)

posted
Total Posts
2,750
show more
nooblet

Tom94 wrote:

I can check up exactly nothing without you telling me which users this is about. :P

That being said my theory would be, that the 2nd guy has far more scores in general than the first guy, ending up with a bit more pp due to the base-pp which every score gives.
I forgot to mention this on the wiki-page:

Every score gives a small base pp amount of 0.25, decreasing by a bit with every new highscore that you make. With 1000 highscores you get around 200pp and with 5000 around 400pp. The cap is 416.666pp at the moment, assuming infinitely many highscores are possible.
(pp gain = 0.25 * sum 0.9994^i, i=0 to amount of highscores)

This gets less and less relevant the higher your pp goes and is meant to both encourage playing more maps for the lower-level players and prevent big rank losses after making a new score at the lower ranks.
In the mid-high ranks it is pretty much irrelevant, since 99% of the people have enough highscores to be less than 50pp away from the cap.
Ah thank you, that's definitely it then. Guess it's time to download more beatmaps :) I don't even have that many beatmaps (including other modes) in total yet :?
Zare
After checking the algorithm for speed value of a score I have a question.

According to this, the OD of a map is not taken into account. Why is that? I mean sure, OD >supposedly< only affects accuracy, but if OD is completely ignored while calculating speed difficulty, or the value of a score, wouldn't that mean that the system thinks that every player who gets an 98% score on a random OD6 or OD7 200 BPM stream map is actually able to stream that fast, even when they're effectively just streaming 170 BPM? they can still SS relatively long streams if they start clicking early and end late when the map is low OD like that.

Or is this covered/prevented by the Accuracy factor which is also mentioned?
Topic Starter
Tom94

Zare wrote:

After checking the algorithm for speed value of a score I have a question.

According to this, the OD of a map is not taken into account. Why is that? I mean sure, OD >supposedly< only affects accuracy, but if OD is completely ignored while calculating speed difficulty, or the value of a score, wouldn't that mean that the system thinks that every player who gets an 98% score on a random OD6 or OD7 200 BPM stream map is actually able to stream that fast, even when they're effectively just streaming 170 BPM? they can still SS relatively long streams if they start clicking early and end late when the map is low OD like that.

Or is this covered/prevented by the Accuracy factor which is also mentioned?
The accuracy factor also depends on OD, so yeah, it is effected by this.
mcdoomfrag
- Length of the beatmap in hit objects / hit circles
- Example: Draining time is 2 minutes; 1,000 objects > 500 objects.
- Likewise: 1,000 objects; Drain time 2 minutes > Drain time 5 minutes
What exactly is going on here? I'm having a hard time understanding what you mean by this part of the wiki :S

Edit:

Luna wrote:

Basically, high note density = hard, low density = easy
10 notes per second is harder than 5 notes per second
Simple stuff like that
Oh, now I get it. I wouldn't have guess that the figures were used to explain such a simple concept. I feel dumb now :(
Luna
Basically, high note density = hard, low density = easy
10 notes per second is harder than 5 notes per second
Simple stuff like that
rexcannon
Just another quick question here.

I must have the wrong idea how this works because I look at this score I got here,



Doesn't this mean it's supposed to give me 17% of 106pp (18pp)? Because I 'm pretty sure it didn't give me any pp at all.
Luna
The score did give you 18pp, but you didn't see the full effect because by getting the score, you also pushed down a lot of other scores into lower weightings.
So while this score gave you a bunch of pp, the slightly worse scores you already had previously lost some of their value. It evened out and you probably gained slightly less than a full pp after everything was calculated.
rexcannon
That's not very encouraging honestly.
Luna
It's to avoid people farming pp by getting scores around their average skill all the time. You only really get noticable increases from your absolute top performances.
Horolynn
It's not supposed to encourage you. It's there to reflect your skill level at the current time. The pp amount grows when you keep getting better and better scores, meaning that you are getting better and able to set those scores. Fun gameplay and getting better should be the encouragement and if isn't fun for you, you shouldn't really be playing it cause you will get burnt out and frustrated really quickly.
Topic Starter
Tom94

Luna wrote:

Basically, high note density = hard, low density = easy
10 notes per second is harder than 5 notes per second
Simple stuff like that
That's not what that part of the wiki is supposed to say at all. It should just say, that beatmap length is measured is based on amount of hitobjects, not on time. It has exactly nothing to do with density, since a 5 minute map with 1000 hitobjects would get the same length bonus as a 1 minute map with 1000 hitobjects.

I'll fix the last line of that section of the wiki. No idea what the person who added it was thinking when saying, that shorter beatmaps would be weighted more with the same amount of hitobjects. I didn't put that in originally. :P
Dexus
I made a thread but then realized there was this thread so I nuked the other one. Just going to copy/paste it into here.

So how do you calculate star diffiuclty with mods, or is this a ppv2 wiki/future topic we will have to wait for? I'm curious as to how much it affects the star rating when adding DT / HR / HD so I can gauge more accurately how difficult a map is. Like say someone was to play this map https://osu.ppy.sh/s/132095 [Hard] with DT the star rating would go from 2.05 to ? as compared to playing HR+HD. HD and DT have static multipliers to the difficulty while HR has that ceiling effect (excluding CS) so I'd assume it's not something to be so easily calculated without using a program/the game?
hehe

Dexus wrote:

I made a thread but then realized there was this thread so I nuked the other one. Just going to copy/paste it into here.

So how do you calculate star diffiuclty with mods, or is this a ppv2 wiki/future topic we will have to wait for? I'm curious as to how much it affects the star rating when adding DT / HR / HD so I can gauge more accurately how difficult a map is. Like say someone was to play this map https://osu.ppy.sh/s/132095 [Hard] with DT the star rating would go from 2.05 to ? as compared to playing HR+HD. HD and DT have static multipliers to the difficulty while HR has that ceiling effect (excluding CS) so I'd assume it's not something to be so easily calculated without using a program/the game?
http://osutp.net/beatmaps
Dexus
Those aren't weighted the same due to various difficulty settings actually being factored into the star difficulty.

This will be easier to figure out when the game's internal star system gets updated and to work dynamically with the mods selected.
hehe

Dexus wrote:

Those aren't weighted the same due to various difficulty settings actually being factored into the star difficulty.

This will be easier to figure out when the game's internal star system gets updated and to work dynamically with the mods selected.
basically DT > HR > FL > HD. i'm not exactly sure what you mean by "various difficulty settings actually being factored into the star difficulty" since ppv2 is pretty much based of tp and hence the song's difficulty is the same. if you want to see what the mods do, tick the boxes and you can see how the level goes up. HD and FL just give bonus aim points (there are 3 components, aim(cursor), speed(mostly bpm) and accuracy (self explanatory)). an perfect FC give maximum speed and aim points, and more accuracy gives more accuracy points.

DT: increases OD, AR and BPM. all three stats are increased dramatically.
HR: increases CS, AR, OD, HP. HP does not really matter for FCs. higher OD increase than DT, thus more points given for better accuracy. CS and AR affect aim, but not as much as BPM and AR in DT.
HD and FL: bonus aim points.
EZ: extremely low AR is also awarded bonus acc.
HT: opposite of DT.
Topic Starter
Tom94
tastystew is correct. However there can be edge-cases where DT / HR might have a different ordering or even be below FL or HD. Really depends on the maps. Detailed explanations on how the difficulty is computed are a future wiki topic.
hehe

Tom94 wrote:

tastystew is correct. However there can be edge-cases where DT / HR might have a different ordering or even be below FL or HD. Really depends on the maps. Detailed explanations on how the difficulty is computed are a future wiki topic.
this is good to hear.
Dexus
I know what the mods do in general to aim/speed/etc , you're not really understanding me. I really just want to see how the star rating is affected through mods. From what I can see the "level" in osutp isn't the same as it only seems to factor the aim/speed while "star rating" includes OD, AR, CS, etc. So looking at the lists there are differences in how it is sorted. It's fine for now as I can see in the future the star rating changes due to mod usage on a map may possibly be shown.

It's just weird because two different maps, Map A having a higher star rating while B has a lower star rating (they are relatively close) . Both are done with mods and get the same acc yet map B rewards more. I wanted to see if possibly when mods are on they show that map B may possibly have a higher rating than map A.

I hope that makes sense.
Topic Starter
Tom94

Dexus wrote:

I know what the mods do in general to aim/speed/etc , you're not really understanding me. I really just want to see how the star rating is affected through mods. From what I can see the "level" in osutp isn't the same as it only seems to factor the aim/speed while "star rating" includes OD, AR, CS, etc. So looking at the lists there are differences in how it is sorted. It's fine for now as I can see in the future the star rating changes due to mod usage on a map may possibly be shown.

It's just weird because two different maps, Map A having a higher star rating while B has a lower star rating (they are relatively close) . Both are done with mods and get the same acc yet map B rewards more. I wanted to see if possibly when mods are on they show that map B may possibly have a higher rating than map A.

I hope that makes sense.
The star rating at the moment is pretty much equivalent to the osu!tp level, just scaled in another way. Therefore the same changes through mods apply.
pooptartsonas
So star rating does not currently take OD into account? Also, I know that HP Drain is not relevant to pp calculations due to how they work, but I think it could be relevant for something like star system, especially for newer players who are not worried about full combos? I guess a star system rating how hard maps are to pass would be entirely different, but maybe it could be useful as a very small factor or something.

I may be misremembering, but I seem to recall that you said something about OD eventually being incorporated into the star system. If this is the case, a feature showing star system with different mods (probably best displayed after integration with the client) would be very useful imo, even if that is not the case currently.
electrolytes
I'm curious: the new star system looks like it compresses the previous 1-to-just-below-5 range into maybe a third or less of what it was (perhaps 1 to ~2.2). Is anything planned to allow searching for ranges of difficulty, similarly to with OD, AR, HP, and CS? (ex. "sd>1.9 sd<2.2"). Maybe something like this (other than the difficulty sort, which groups things by a full star apart) already exists and I just don't know about it. If it doesn't though, it seems like it might become much harder to search for beatmaps around the same level of difficulty when the star difficulty changes are brought into the client. This is all assuming the 1-to-5 star range stays around of course.
mcdoomfrag

electrolytes wrote:

I'm curious: the new star system looks like it compresses the previous 1-to-just-below-5 range into maybe a third or less of what it was (perhaps 1 to ~2.2). Is anything planned to allow searching for ranges of difficulty, similarly to with OD, AR, HP, and CS? (ex. "sd>1.9 sd<2.2"). Maybe something like this (other than the difficulty sort, which groups things by a full star apart) already exists and I just don't know about it. If it doesn't though, it seems like it might become much harder to search for beatmaps around the same level of difficulty when the star difficulty changes are brought into the client. This is all assuming the 1-to-5 star range stays around of course.

Tom94 wrote:

Keeby wrote:

It would be nice to implement the star system in-game, if possible.
All planned.
Its in the works.
silmarilen
is this map https://osu.ppy.sh/s/155691 really supposed to be 5.00 stars? tp rates it at lvl 67
Topic Starter
Tom94

silmarilen wrote:

is this map https://osu.ppy.sh/s/155691 really supposed to be 5.00 stars? tp rates it at lvl 67
Fixed. It didn't get updated with its actual difficulty value for some reason. The internal pp difficulties were stored correctly. (rrtyui only got 211 for his #1)
KogureKun
Curious about how much of a bonus is applied when playing at low AR's (sub 4.5/5). Is it possible to use EZ mod on hard+/insane songs and get a higher PP than if you were using nomod? Not sure how EZ mod is calculated for a difficulty and if there's any bonus for the huge clusters of notes
Dexus
How does weighting add up to such a high amount of pp for players when stuff exponentially yields less in comparison to your Top most performance. A mass amount of scores are weighted as 0%. Am I reading the list wrong and the weighting is already applied to the shown pp amount?

If it isn't then I would like for the top performances if possible to show the pp*weighting amount automatically; and then next to the % show the current displayed mount of pp without the percentage affecting it; that way instead of showing a bunch of 400pp, 300pp, 300pp, it would show the right amounts 400pp, 285pp, 270pp, etc.

The way the weighting is displayed and how much you actually have is pretty confusing to me personally.
Topic Starter
Tom94

Dexus wrote:

How does weighting add up to such a high amount of pp for players when stuff exponentially yields less in comparison to your Top most performance. A mass amount of scores are weighted as 0%. Am I reading the list wrong and the weighting is already applied to the shown pp amount?

If it isn't then I would like for the top performances if possible to show the pp*weighting amount automatically; and then next to the % show the current displayed mount of pp without the percentage affecting it; that way instead of showing a bunch of 400pp, 300pp, 300pp, it would show the right amounts 400pp, 285pp, 270pp, etc.
That would mislead many people into thinking the scores were considered "worse" instead of only having a weight applied.

If you want to know how this pp amount adds up just check the math behind it and verify it yourself. It is correct.
Dexus
Well then can you show the weighted amount of pp NEXT to the weight percentage then? Why fool people into thinking pp for a map is worth something when it really isn't worth anything?

Edit:
GoldenWolf

Dexus wrote:

sounds cool, i'm too lazy to do the math most of the time so yeah
nooblet
I was wondering how people had such high PP's when their top scores didn't have many high scores, too. It's most likely due to this

Tom94 wrote:

Every score gives a small base pp amount of 0.25, decreasing by a bit with every new highscore that you make. With 1000 highscores you get around 200pp and with 5000 around 400pp. The cap is 416.666pp at the moment, assuming infinitely many highscores are possible.
(pp gain = 0.25 * sum 0.9994^i, i=0 to amount of highscores)

This gets less and less relevant the higher your pp goes and is meant to both encourage playing more maps for the lower-level players and prevent big rank losses after making a new score at the lower ranks.
In the mid-high ranks it is pretty much irrelevant, since 99% of the people have enough highscores to be less than 50pp away from the cap.
I really like the idea of showing how much PP it's worth after weighing though. Perhaps to a decimal or two for the ones giving base PP? :)
Topic Starter
Tom94
Fair enough, it does show the actual pp it gives now. I'm still against the decimals though, since it clutters things too much.
Dexus
Best solution would probably be title= whatever in the div so you can get a lot more information into a tooltip. That would make it so you could just have the pp numbers listed and then hover over for more detail on weightings, mods, etc.

Thanks for adding the weighted amounts!
MPGHThunder
I'm extremely confused about the ranking system. I mean, take rrtyui for example. He's #1 on The Quick Brown Fox yet it only gave him 275pp. His best performance is Saiya and that gave him 542pp. I mean, like dafaq?
silmarilen
newsflash: big black is not the hardest thing in existence
MPGHThunder

silmarilen wrote:

newsflash: big black is not the hardest thing in existence
Newsflash: That wasn't the point.
silmarilen
so... what else would be the point?
MPGHThunder

silmarilen wrote:

so... what else would be the point?
It's a significantly more difficult map. Not to mention he's got like a 8 million lead...
Zare
That's the point silmarilen addressed tho.
Remote Control on HDDT is harder to FC than Big Black Nomod.
MPGHThunder

Zare wrote:

That's the point silmarilen addressed tho.
Remote Control on HDDT is harder to FC than Big Black Nomod.
165 * 2 = 330 as opposed to 360.3 BPM on The Big Black. If you add in the sudden jumps and hidden it probably is a more difficult map, but that's subjective really. 7,802,730 point lead and 3.80% more accurate than HappyStick. The map has also been played much more than Remote Control. The PP system is horribly vague...
silmarilen
please, if you dont know what you're talking about, dont talk about it. just the fact that you think double time doubles the bpm already shows you dont, not to mention all the other flaws in your argument.
show more
Please sign in to reply.

New reply