I've got a question about aim pp/stars

How exactly does CS affect it?
let's assume we have a 200 BPM jump map that has a certain "base" aim difficulty defined by the speed and spacing of the jumps
Now we have CS4 on this map and we get the star rating 5.2 or something
What exactly in the calculation changes when we switch to CS5 or higher (e.g. HR) now?

As it seems now higher CS only gives a linear, minor boost to the base difficultym which woule explain why it is so underrated when it actually adds a lot to the difficulty

As far as I know the whole map is scaled up so that CS becomes the same, and the distances get bigger. It's pretty fair imo, since you can do this on your own anyway - you can play with bigger hitcircles but bigger distances too by increasing your osu resolution.

Whenever there's a pp update, there really needs to be a refresh in rankings. It's extremely annoying trying to aim for pp scores to get my rank back to where it was, then losing ranks when I get a high pp score.

play a map you havent played to easily set a new score and refresh your rank
personally i play a tv size [Easy] with nofail doubletime and wave my cursor around for 50 seconds and bam update

Riince wrote:

play a map you havent played to easily set a new score and refresh your rank
personally i play a tv size [Easy] with nofail doubletime and wave my cursor around for 50 seconds and bam update

It doesn't work unless everyone's refreshed already.

afaik it does update, it just takes a couple of days because of the amount of people/scores it has to update

Tess wrote:

pielak wrote:

Hidden still overrated, >>>flashlight still underrated<<<

Whoever says that flashlight is easy if you know the map should shoot themselves.
It's the hardest mod in the game and deserves to be treated as such.

ayy lmaoo

Drezi wrote:

As far as I know the whole map is scaled up so that CS becomes the same, and the distances get bigger. It's pretty fair imo, since you can do this on your own anyway - you can play with bigger hitcircles but bigger distances too by increasing your osu resolution.

No, it doesn't scale to it.. Resolution and Circle Size don't actually relate to each other.
I agree with Zare, I made a few CS5 jump maps and their star rating is absurdly low compared to the actual difficulty due to the circle size being underrated.

Priti wrote:

.
I agree with Zare, I made a few CS5 jump maps and their star rating is absurdly low compared to the actual difficulty due to the circle size being underrated.

You're just not used to cs5. Every map that comes out these days are either cs4 or cs5. You're so accustomed to it that anything 5 or higher is small to you. After I learned to play cs6.5 I think the star rating is actually balanced. The difficulty is mostly from just keeping a full combo on a long map with small circles..

Priti wrote:

Drezi wrote:

As far as I know the whole map is scaled up so that CS becomes the same, and the distances get bigger. It's pretty fair imo, since you can do this on your own anyway - you can play with bigger hitcircles but bigger distances too by increasing your osu resolution.

No, it doesn't scale to it.. Resolution and Circle Size don't actually relate to each other.
I agree with Zare, I made a few CS5 jump maps and their star rating is absurdly low compared to the actual difficulty due to the circle size being underrated.

Yes they do relate as long as you are not playing fullscreen and thus re-scaling resolutions which do not match your monitor back up.

It's true that resolution and circle size are related like that, but any incidental movement (eg shaky hands) has its effect amplified more the smaller the circle size is. Yes, increasing the tablet size solves this issue in theory, but one can only play with a tablet area so large, and most people play with roughly the largest area they feel comfortable playing with. And the reason people try to play with larger areas is just that; the larger the area, the less of an impact shakiness has. Thus, I think on the higher end (cs6/7+) the circle size really tends to get underrated.


The other thing I want to say regards the recent changes. The reason that longer maps should (and now do) have the scaling factor is because your chance to miss and the amount of time required to fc drastically increase as the map length increases. This is irrelevant if the map is a fairly routine full combo, causing long maps that are easy for the player to become overrated. This is not important in most cases, as a map that is a routine full combo for the player will not be worth a significant amount of pp to them anyway.

However, it causes problems in the extreme case (~2500+ combo). The scaling on long maps continues to increase, meaning that a player can gain a significant amount of pp from a map that they will practically never miss on, given that it is long enough. This is clearly seen in the world's end map that pielak referenced earlier. The pp awarded is pretty clearly too high. With that said, I think the changes worked wonders for anything in the 1000-1500 combo range. The bonus for map length simply needs to begin to taper off around 2k combo, and I think the algorithm will be in an excellent spot in terms of accounting for map length.

pooptartsonas wrote:

And the reason people try to play with larger areas is just that; the larger the area, the less of an impact shakiness has. Thus, I think on the higher end (cs6/7+) the circle size really tends to get underrated.

shakiness is almost completely irrelevant for mouse players though. once they snap their cursor goes dead still.
i say almost because obviously shaking can have some effect on the snap itself but a hovering tablet player will almost always shake either way, a mouse player depends on nervousness.

shakiness seems like a poor reason so buff high CS

pooptartsonas wrote:

However, it causes problems in the extreme case (~2500+ combo). The scaling on long maps continues to increase, meaning that a player can gain a significant amount of pp from a map that they will practically never miss on, given that it is long enough. This is clearly seen in the world's end map that pielak referenced earlier. The pp awarded is pretty clearly too high. With that said, I think the changes worked wonders for anything in the 1000-1500 combo range. The bonus for map length simply needs to begin to taper off around 2k combo, and I think the algorithm will be in an excellent spot in terms of accounting for map length.

fair point though, this should definitely be considered. even low acc fcs of that are giving well above 300 pp just wow

Riince wrote:

shakiness is almost completely irrelevant for mouse players though. once they snap their cursor goes dead still.
i say almost because obviously shaking can have some effect on the snap itself but a hovering tablet player will almost always shake either way, a mouse player depends on nervousness.

Eh, mouse players have other things to worry about than nervousness.

my top scores

















wat

Your accuracy being lower in the non-DT scores doesn't exactly help anything.

You'd be surprised how much adding 1% to your score when your accuracy is that high can do to your pp.

lagomorph worth same pp as da nu nuttah lol
OD too stronk

Riince wrote:

lagomorph worth same pp as da nu nuttah lol
OD too stronk

Da Nu Nuttah isn't hard, I find lagomorphic's jumps harder

CalignoBot wrote:

Your accuracy being lower in the non-DT scores doesn't exactly help anything.

You'd be surprised how much adding 1% to your score when your accuracy is that high can do to your pp.

even with the same accuracy, those DT scores would give more pp than the Nomod scores, which is just stupid

Zare wrote:

CalignoBot wrote:

Your accuracy being lower in the non-DT scores doesn't exactly help anything.

You'd be surprised how much adding 1% to your score when your accuracy is that high can do to your pp.

even with the same accuracy, those DT scores would give more pp than the Nomod scores, which is just stupid

You're undervaluing accuracy/overvaluing the difficulty of these maps imo
Honestly they're all (this top 5) about the same difficulty

GoldenWolf wrote:

Riince wrote:

lagomorph worth same pp as da nu nuttah lol
OD too stronk

Da Nu Nuttah isn't hard, I find lagomorphic's jumps harder

well i find da nu nuttah harder and thats just my opinion
my score on both, the star diff, and even the scoreboards themselves seem to agree

im pretty sure lagomorph was farmed heavily back with tp because of the high aim value it had.

perhaps. but the # of complete playthroughs on them is even, with da nu nuttah edging out by a bit more.
but maybe you're right. i mean how many of those plays are ponyfags who found osu and played renard maps?

-GN wrote:

Who plays FL for pp anyway? It's so ball-bustingly hard to pull off anything above 250pp that it's not worth it in any case. FL is played for interesting scores, not to get points for it, since anyone who wanted the latter would find more effective ways to get it than playing a single map 200 times to get some high rank.

It could probably use a small buff, just to make the FL scores stand even taller compared to HD/nomod plays, but other than that I don't get the point, honestly.

The point of buffing it would be exactly that. Making it worth it to play FL, because right now it is /heavily/ underrated. For a mod where you can not possibly get a sightread FC/SS without cheating, the fact that it isn't the most rewarding mod is ridiculous.

If you're going to tell me that this play



is easier to achieve than this



you're crazy.

pooptartsonas wrote:

However, it causes problems in the extreme case (~2500+ combo). The scaling on long maps continues to increase, meaning that a player can gain a significant amount of pp from a map that they will practically never miss on, given that it is long enough. This is clearly seen in the world's end map that pielak referenced earlier. The pp awarded is pretty clearly too high. With that said, I think the changes worked wonders for anything in the 1000-1500 combo range. The bonus for map length simply needs to begin to taper off around 2k combo, and I think the algorithm will be in an excellent spot in terms of accounting for map length.

You're underestimating how hard it is to get a 2K+ combo. And, for that matter, anyone saying that World's End is overrated is /completely/ ignoring the ridiculously long spaced streams, the constant jumps that can sometimes get pretty damn big and the fact that it's a 6 minute+ beatmap. It is not overrated at all.

Tess wrote:

The point of buffing it would be exactly that. Making it worth it to play FL, because right now it is /heavily/ underrated. For a mod where you can not possibly get a sightread FC/SS without cheating, the fact that it isn't the most rewarding mod is ridiculous.

this is still not a play-style that should be promoted. it isn't fun or interesting, and because it doesn't make the song harder in any technical aspect it provides a negligible pp bonus like hidden and that's how it should be. there's a difference between a maps difficulty and how hard it is to achieve the score, the second one isn't used in pp calculations.

Tess wrote:

You're underestimating how hard it is to get a 2K+ combo. And, for that matter, anyone saying that World's End is overrated is /completely/ ignoring the ridiculously long spaced streams, the constant jumps that can sometimes get pretty damn big and the fact that it's a 6 minute+ beatmap. It is not overrated at all.

and you're underestimating the significance of 350+ PP. that's probably one of the easiest score Rucker has achieved on his top 10 yet it gives more pp than almost all of them.

Riince wrote:

this is still not a play-style that should be promoted. it isn't fun or interesting, and because it doesn't make the song harder in any technical aspect it provides a negligible pp bonus like hidden and that's how it should be. there's a difference between a maps difficulty and how hard it is to achieve the score, the second one isn't used in pp calculations.

This is a really stupid thing to say, though. First of all, whether it's fun or interesting or not depends largely on the player. A lot of things that I find fun other players find unnecessarily hard or tedious, and vice versa. That doesn't change the fact that those maps deserve to be rewarded. Also, it does make the map harder because it makes it harder to read. That a map isn't faster or the map's settings don't change doesn't mean that it's not harder when you make 90% of it fucking invisible. Lastly, this line is particularly dumb

Riince wrote:

there's a difference between a maps difficulty and how hard it is to achieve the score

I think you're confusing subjective difficulty with objective difficulty. Yes, if you happen to have trouble with squares, nailing a square jump in an easy map won't give you any pp, despite that score being difficult to achieve for you. But FL is difficult for everyone. Just like DT/HR aren't mods that just anyone can perform well with. So the difficulty in FL should give a significant pp bonus, especially since most players can't even play with it. People only hate FL because they suck at it, because it is hard.

Riince wrote:

and you're underestimating the significance of 350+ PP. that's probably one of the easiest score Rucker has achieved on his top 10 yet it gives more pp than almost all of them.

This doesn't nullify my statement, though. I said you're underestimating how hard the map is and you say you're not, but don't really give any reasonable explanation as to why. "It's the easiest score Rucker achieved on his top 10" - but what makes it easier? What about this map is easier than the others he played that gave him equal or less pp? Unless you actually answer that you're not really saying anything at all.

I think the main point is that FL shouldn't be made necessary by rewarding it too much, just because it's hard. By that logic you could make a mod where there's no sound at all, and only a black screen, so FC-ing a map like that would be insanely hard, but should that style of play be promoted by giving it insanely high rewards?

You see getting good FL scores requires specific practice on the given map, while learning the other mods gives you the ability to play them on any map, so it would be bad for everyone if getting high ranks would require you to practice individual maps with FL, instead of focusing on improving your general skillset.

Drezi wrote:

You see getting good FL scores requires specific practice on the given map, while learning the other mods gives you the ability to play them on any map, so it would be bad for everyone if getting high ranks would require you to practice individual maps with FL, instead of focusing on improving your general skillset.

more or less what i meant by "not a play-style that should be promoted", thank you for saying this

I understand your points, but it remains a fact that FL is a mod within the game and FL is difficult. The ability to memorize a map and FC it with FL (with decent accuracy) is just as much skill as any other osu skill (reading, accuracy, consistency, stamina, etc.). Since it is in the game, and it is difficult, it should be rewarded. If you think it shouldn't be in the game you should consult peppy, rather than not rewarding players for their skill and effort. Keep in mind - pp is short for performance points. This means that your rank is based on your performances and not your skill.

Let's take a hypothetical situation where a lower ranked player (30-50K), by pure luck, SSes Freedom Dive. Now, obviously this is not possible, but let's say that there is a 1 in 100 quadrillion chance that somebody SSes Four Dimensions by pure luck, rather than skill. They would still receive the same pp that a skilled player would for that play, because they're judged by their performance and NOT by their skill. So, regardless of how often you need to practice a map, if you achieve a score that others can't achieve easily, you should be rewarded for that performance, just like with any other mod.

Discouraging repeated practice on the same map is like discouraging osu. rrtyui played Big Black around 3000 times before SS-ing it. Does that make his SS any less impressive? And does the fact that he played it that often (because it's a difficult map) mean that his score shouldn't be rewarded?


The answer is no

Well sure, but you have to admit that there's still a difference between having to retry a lot to be able to play a map with FL, vs retying a lot until you make a better play on a map you can already do. In the first case you probably don't have technical difficulties with the map, but still you have to retry a lot to memorize it no matter what, while in the second case you're just retrying to set a score that's slightly above the usual, average level of your play, and you could set the same score later after improving some, without having to retry as much.

FlashLight currently gives a whopping 50% pp bonus on aim pp. That means on an aim-heavy map you can expect around 25% more total pp compared to a non-FL play with the same outcome. I wouldn't call that low or underrated.

The main problem with FL scores is that they usually are way sub-par compared to the same player's other best performances.

You should buff FL more so I get more pp from doing nothing.

Playing FL does make you better at playing FL, though. You basically get better at reading follow points and predicting/guessing right, and aiming at nothing (which btw, is harder than aiming at something). It still requires a bunch of retries though.

Oh yeah and accuracy tends to be harder because your mind is focused on recalling parts of the map rather than paying attention to when to click.

.

I feel like the difference between OD7, OD8, and OD9 for accuracy pp is too large. There's basically no difference between OD7 and OD8, and OD9 isn't that much harder. I don't think high OD's should be worth any more, but I think the lower OD's shouldn't be worth so much less.

-Soba- wrote:

I feel like the difference between OD7, OD8, and OD9 for accuracy pp is too large. There's basically no difference between OD7 and OD8, and OD9 isn't that much harder. I don't think high OD's should be worth any more, but I think the lower OD's shouldn't be worth so much less.

From the wiki


I disagree.

Tess wrote:

-Soba- wrote:

I feel like the difference between OD7, OD8, and OD9 for accuracy pp is too large. There's basically no difference between OD7 and OD8, and OD9 isn't that much harder. I don't think high OD's should be worth any more, but I think the lower OD's shouldn't be worth so much less.

From the wiki


I disagree.

even if the windows for getting 300 is smaller, most people are already clicking well within the OD8 window in OD7 anyway is the point I was making. I don't think OD8 maps should give so much more pp for free just because they're OD8.

personally i get around 99-100% on od7 and around 96-97 on od8 because im no accuracy player so im unsure what 'most people' are, unless its just people much better than me [like you] and obviously there are less of those than the other way around so i dont see why people who are more accurate shouldnt be rewarded for their accuracy.

In the absence of per-hitobject data, the pp calculation system can't distinguish between someone who SSed an OD7 map by using the entire 300 window and someone who would have SSed the map even if it was OD9 - pp calculations can't assume you're hitting "well within the OD8 window in OD7". If you're capable of SSing OD8 maps, then doing so will get you the extra pp that the lenient timing window of OD7 cheated you out of.

There are definitely people out there who have noticeably worse accuracy on OD8 than on OD7 (Edit: Hi Riince!). They feel the same to you because you find both of them so easy.

I can get >99% with OD9.8 maps and around 97-98% with OD10 and I still notice a difference between OD7/8/9/10. If you don't notice a difference you're just not paying attention. Even if you're a high accuracy player, if you try SSing 10 OD7 maps and then 10 OD8 maps subsequently, I can guarantee you that the 10 OD8 maps will take longer. Try it and prove me wrong.

Maybe I was assuming too much to say most players, sorry. And sure I'll prove you wrong later Tess if you really want me to.

Make sure to record it and post it here. If you can't record just make a nice collection of images, one showing an in-game timestamp from when you start playing, and then each subsequent SS, and you have to do one OD at a time. So, either 10 OD7 SS first, or 10 OD8. I'd like to see this.