Equim wrote:

Partially. But tentatively as a ctb mapper/modder I can tell you that jumps as you said "irregularly" are still experimental, and ctb stream mapping is also so hard to be creative that only one or two ctb mappers can handle this. It's easy to SAY, but it's hard to DO.
Remember, everyone has his own way to enjoy this game. Some enjoy jumps, while some enjoy patherns like you; some enjoy pp, while some don't. While I was a beginner I did enjoy hyperdashes which I thought were cool, and I tried my best to do well in them. Anyways you need to respect to the diversity.
Also, whether you love pp or hate, it's always that you care pp. I suggest you don't always judge a ctb specific map according to pp, since in this way you are likely to ignore what painstaking care the mapper pay in the map. In this point of view I can't agree more with Spectator (see his ask http://ask.fm/Spectat0r/answer/128855444064 and http://ask.fm/Spectat0r/answer/128938149472). In addition, sooner or later the pp arithmetic system will changed, but the quality of a map will still remain.

I respect all mappers,spec's ctb maps can lead new players to join us quickly.His maps should be respected,but I just wanna new PP system decreases this type of ctb maps a little to guide new players correctly.CTB maps can be made into different kinds.Some maps contains so many red
fruits,some maps contains hard patterns and you needs high APM.

Overall,I sincerely bless all CTB mappers to show your own style,don't be shy!

Now PP system has something unfair,I just point it .

How did you find where I show anything rude to any mappers?

Mappers need to hurt their brains to share their ideas , just the new system needs to judge their maps more correctly.

That is what I expect. I love this game and I wanna this game more perfectly.

You know ,the official rules can control your mind ,we need to fix the pp system and make the game filled with charm !

I am weak at Ctb but ! I am not a coward ! I can reply to you : I don't care about PP ! because I just wanna be a HD DT pro player, why should I care about PP ? Challenges are enough to me!

I say again : I love this game deeply and wanna her more perfectly.

We are members of this warm family,shouldn't we love her and take care of her ?

Random speech, but I have to agree with you. Also, any (un)ranked map is good as long as at least one player enjoys the map.

hey there, is this still alive?

yea, but there is just not much to add anymore, since we almost mentioned every problem...
plus Tom is focusing on University as far as i know, so he won't have the time to do anything about it

Mr pepsi did say last year he had a group of people looking into the pp system as a whole. I guess it's still too early in the year to go asking for in-depth updates about it.

I think spins should reward 0% to 5%pp dependent on how close to max spin. High accuracies should also reward 0 to 5%pp dependent on how close to 100% Accuracy.

It doesn't have to cap at 5% though, just a arbitrary number D:

He Ang Erika wrote:

I think spins should reward 0% to 5%pp dependent on how close to max spin. High accuracies should also reward 0 to 5%pp dependent on how close to 100% Accuracy.

It doesn't have to cap at 5% though, just a arbitrary number D:

Gonna have to disagree about spins giving pp due to the fact that I'm sure it's really hard for the game to determine how good your spin actually is and I doubt it can calculate what the best humanly possible spin is.

Accuracy already gives bonus pp.

Zak wrote:

He Ang Erika wrote:

I think spins should reward 0% to 5%pp dependent on how close to max spin. High accuracies should also reward 0 to 5%pp dependent on how close to 100% Accuracy.

It doesn't have to cap at 5% though, just a arbitrary number D:

Gonna have to disagree about spins giving pp due to the fact that I'm sure it's really hard for the game to determine how good your spin actually is and I doubt it can calculate what the best humanly possible spin is.

Accuracy already gives bonus pp.

I'm not sure about how the current system can account for spinners, but mathematically it's very easy to calculate max spin using basic trigonometry. In fact a good spinster can count max spin just based on pure human judgement which is almost flawlessly correct in nearly all patterns. I would say that based on how salad! catches bananas it seems that the system doesn't really include a max spin calculator, but hopefully whenever the engineers find their time they can add on an undisplayed bananas counter and make some pp. I think that the CTB spinner is very unique and heavily skill/experience oriented, to have it just affect score is rather under-appreciated D:

I think the possibility of a good spin making you go from not even top 8 all the way to rank 1 is plenty enough.

Zak wrote:

I think the possibility of a good spin making you go from not even top 8 all the way to rank 1 is plenty enough.

Often on very hard songs where there are only a few NoMod scores or very easy songs with a bunch of HDHRDT scores, spinners can give you good ranking on the scoreboard, but as for the rest having good spinner has to be coupled with good mods usually if not always involving HD. Otherwise neither getting pp nor top score is very viable. E.g. A NoMod score with ridiculous spin on [Overdose] will be easily pushed off the Top 50 within like a month. DT spinner can hardly outplay or even level HD spinners. HR will bring your score far up, but however good a HR spin it will be dominated by HDHR in the end. To have a ridiculous spin just squeezed away is really unfortunate, at the least a little pp will do some good. Spinner rewards should not be greater or even close to HD rewards or any other score multiplying mods, but I still think that other than 1100 score which is tiny compared to mods, slight pp boost will make pushed-off-the-board mad spins feel more deserving of praise and appreciation.

I would like to see this topic revived, there is not dumb that this pp-sytem is just bad.

BoberOfDarkness wrote:

I would like to see this topic revived, there is not dumb that this pp-sytem is just bad.

Not much to say that wasn't said already.

pp still needs many changes D: The discussion is still important.

When can I see players like Clara Kyptoric VelperK CLSW wampir Emiru geraldwch e0486 Setomi etc in tops :C

Knowledge is important, but having good data to work on is importanter.

The main reason we don't need discussion at this time is because it won't get much done, we have to wait until Tom (or someone else) who has the time and ability to reshape the system again comes along to do so, but at the same time, most issues that are currently there have already been brought up and will likely be looked over by whoever changes the system again.

Also the reason you don't see those players at the top is because they don't gain pp, that's how the system works, unless you keep gaining, your rank drops further and further.

Zak wrote:

they don't gain pp, that's how the system works

They don't gain pp, that's why the system doesn't work looool

By the way I'm 666th pp commentor C: I'm the king noww

He Ang Erika wrote:

Zak wrote:

they don't gain pp, that's how the system works

They don't gain pp, that's why the system doesn't work looool

By the way I'm 666th pp commentor C: I'm the king noww

What, sure they system isn't accurate but it does work in the sense that if you don't play anything to gain pp, you're going to get passed by people below you.

wampir is too lazy for being on top

Someone said that I should post this in here instead of in the Feature Requests forum. This is about star rating more than performance points, but I hope it still applies.
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
I think that the circle size and approach rate extremes should be weighted more in the Catch the Beat star difficulty rating.

I currently have almost 1500 SS ranks in Catch the Beat. Most of those scores are from maps in the 1.8 star and below difficulty range. Around 98% of all of the SS ranks that I have obtained took me anywhere from one to three tries. A few took anywhere from four to fifteen. I was fine with the amount of tries even if it took ten or so. Everything was going great until I ran into Pixel Studio - Gravity [ZOMGWTF]. I've tried that map 226 times and I still haven't even gotten half way through with the perfect mod on. If you don't know what this map is then you are in for a treat because it has a CS of 7, an AR of 10, an OD of 10, and a HP of 9. With Catch the Beat, the map requires fantastic precision and memory.

Is my lack of ability to get a SS rank on this map because I'm simply just not good enough? Yes, but that doesn't take away from the fact that this map should be, in no way, only 1.49 stars in Catch the Beat. Now, I don't claim to know the intricacies of the formula for the Catch the Beat star difficulty rating, but I think something needs to be changed. Weighting approach rate and especially circle size extremes more for the star rating should be able to bump this map way up to at least closer to where it should be.

I do recognize that changing a formula that applies to 40,000 maps to address a problem in only one map would cause more problems than it would fix, but this formula change wouldn't just effect this map since it would beneficial for all small circle size maps. It is consistent that the small circle size maps take me much longer to SS than maps with a normal circle size. In fact, all of the maps that took me more than around six tries were the under-weighted high-CS maps. Through this, I genuinely believe that this change will improve the level of polish in osu!, if only slightly.

Thank you for reading! Have a good day.

Sleepteiner wrote:

Someone said that I should post this in here instead of in the Feature Requests forum. This is about star rating more than performance points, but I hope it still applies.
-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
I think that the circle size and approach rate extremes should be weighted more in the Catch the Beat star difficulty rating.

I currently have almost 1500 SS ranks in Catch the Beat. Most of those scores are from maps in the 1.8 star and below difficulty range. Around 98% of all of the SS ranks that I have obtained took me anywhere from one to three tries. A few took anywhere from four to fifteen. I was fine with the amount of tries even if it took ten or so. Everything was going great until I ran into Pixel Studio - Gravity [ZOMGWTF]. I've tried that map 226 times and I still haven't even gotten half way through with the perfect mod on. If you don't know what this map is then you are in for a treat because it has a CS of 7, an AR of 10, an OD of 10, and a HP of 9. With Catch the Beat, the map requires fantastic precision and memory.

Is my lack of ability to get a SS rank on this map because I'm simply just not good enough? Yes, but that doesn't take away from the fact that this map should be, in no way, only 1.49 stars in Catch the Beat. Now, I don't claim to know the intricacies of the formula for the Catch the Beat star difficulty rating, but I think something needs to be changed. Weighting approach rate and especially circle size extremes more for the star rating should be able to bump this map way up to at least closer to where it should be.

I do recognize that changing a formula that applies to 40,000 maps to address a problem in only one map would cause more problems than it would fix, but this formula change wouldn't just effect this map since it would beneficial for all small circle size maps. It is consistent that the small circle size maps take me much longer to SS than maps with a normal circle size. In fact, all of the maps that took me more than around six tries were the under-weighted high-CS maps. Through this, I genuinely believe that this change will improve the level of polish in osu!, if only slightly.

Thank you for reading! Have a good day.

I understand what you're getting at, and i surely think that high Circle Size (with difficult patterning) should get reconsideration in the future,
but with your Example of "Gravity", that map in itself has a quite easy pattern, the only thing that makes it more difficult is just the small circle size combined with high AR, Plus the little amount of weightening for Accuracy is understandable since its such a short song.

But there are a bunch of other olders maps that are being valued "less?" then they should be
^<-(hard to compare with Difficult maps on high AR&low CS that give a comparable high Star Rating)
(most of them being high CS maps with +DTHR added)
(+HD is not valued enough on these cases if its added...because there is a huge increase of difficulty when High CS Maps are being Played HDDTHR)

like for example:

ArmCannon - Borrow Mega Nuke (Eureka Bong Worm) [Exeter Modem]+DTHR = 4,67Stars
Chrono Cross - Yasunori Mitsuda - Time's Scar [The Whole Mapset]+DTHR = 4-5Stars
Daft Punk - Technologic [Insane]+DTHR = 4,23Stars
DragonForce - Through The Fire And Flames [Hard]+DTHR = 4,46Stars
Mindless Self Indulgence - Bullshit [BULLSHIT!]+DTHR = 4,73Stars
Sasaki Nozomi - Cosplay no Kokoroe [Insane]+DTHR = 4,76Stars

But to be honest, im not even sure if those +4Star Ratings are fine, they seem fair...but on special occasions it just feels WAY harder than everything else in its comparable Star Difficulty Range

I think HD should give you less pp and with DT you should get more pp with

Osten112 wrote:

I think HD should give you less pp and with DT you should get more pp with

Yea also I think non-hyper jumps should give more pp

Osten112 wrote:

I think HD should give you less pp and with DT you should get more pp with

I've played so many maps that are downright cancer with HD, but can be played well on DT just after a few tries :X E.g. Fallen Angel [SHD] stuff like that. In general HD would give less pp than DT, but certainly should be more than NoMod.

It's almost off topic but the reason HD is a problem in CTB is because a lack of viable pp maps

If you are a solid player on nomod and farm the popular pp maps (Hanabi Umarudose Xeno etc etc), you land at a solid rank of 80~ as of today
Player B turns out to be a HD player and plays those maps on HD, with an exception of a couple, with the addition of 6~7 HD scores Player B is rank 60~40.
If CTB had more maps that people didn't fc in 15 minutes and move on (Towayori, Genryuu, Dreamless, L) there would be more variety to ranking and also a benefit to becoming better as a player to be able to fc more maps rather than spending 5 plays fcing nomod, 50 on HDing it, then the next 300 trying to HR.

Clana wrote:

It's almost off topic but the reason HD is a problem in CTB is because a lack of viable pp maps

If you are a solid player on nomod and farm the popular pp maps (Hanabi Umarudose Xeno etc etc), you land at a solid rank of 80~ as of today
Player B turns out to be a HD player and plays those maps on HD, with an exception of a couple, with the addition of 6~7 HD scores Player B is rank 60~40.
If CTB had more maps that people didn't fc in 15 minutes and move on (Towayori, Genryuu, Dreamless, L) there would be more variety to ranking and also a benefit to becoming better as a player to be able to fc more maps rather than spending 5 plays fcing nomod, 50 on HDing it, then the next 300 trying to HR.

I agree with that. There are some maps which gives more than 350pp if FC/SS no mod which are easy to do. My actual top score is around 230pp but I'm able to FC Akatsuki no Tsuki Oriental, Piranha or MegaMan map (or 1/2 miss, but I only want SS), which all gives above 350pp, that's ridiculous. I normally don't have the skill to do a >5 stars map so easily with my rank. (#850) I don't say that there're only easys maps above 5 stars but some are easiers than others, for example as Clana said, L is an hard map for me.
I don't mean they're bads maps, no, they're goods and funnys, but they give so much pp for the difficulty that we need to FC them.
+ if we put Hidden or for the really good players HR, we will be easily on the top 100 "without efforts", and that's why a lot of people who plays other mods than CTB said that this mod is easy.
I hope in the future there's some hards maps that even the top 20 have difficulties to get a S on, like an Odoru Mizushibuki [Death Dance] CTB version, it will be very nice I guess !

It's called hdash overweight. faster hdash = higher stars, but there is some pattern that make hdash easier or harder to pass.
This thread is about pp feedback and suggestions, i think we already said everything about pp and stars problem but it's like devs don't read this thread. It's my feeling, like for many thread in feature request that have no answer for ages.

I think when Tom or someone else who can actually come help fix things comes around we can get somewhere if they simply put a heavier weight on timing jumps as most players will agree those are usually much harder to hit than hypers a majority of the time. With weighting put to both hyper and non-hyper patterns put properly lots of converted maps will be giving similar pp to ctb maps and rankings will be much more varied.

CelegaS wrote:

It's called hdash overweight. faster hdash = higher stars, but there is some pattern that make hdash easier or harder to pass.
This thread is about pp feedback and suggestions, i think we already said everything about pp and stars problem but it's like devs don't read this thread. It's my feeling, like for many thread in feature request that have no answer for ages.

It's not that the devs don't read this thread, it's that there was pretty much only a single person capable of creating the pp formula we currently use, and that person was Tom and he's currently busy with real life. We pretty much need to wait for him to be able to come back or for someone else to come along who is able to adjust the formula accordingly.

I am like pushing and wishing my hardest for a pp revamp, while I know for certainty, deep down in my heart that when Genryuu Kaiko gets dethroned I will cry myself to sleep and never wake up again QAQ

.
.
.

It's 'kay I will get better at putting mods on standard converts map, it'll happen when I wake up from those tears

He Ang Erika wrote:

I am like pushing and wishing my hardest for a pp revamp, while I know for certainty, deep down in my heart that when Genryuu Kaiko gets dethroned I will cry myself to sleep and never wake up again QAQ

.
.
.

It's 'kay I will get better at putting mods on standard converts map, it'll happen when I wake up from those tears

That specific map hopefully will bring awareness to a need to a new change.
I had a dream where #132 has 10004 pp after 5 or 6 of those maps gets ranked ;(

About the open sourced thing: The difficulty calculator wasn't open sourced, it's just the pp part after the difficulty was already calculated. (did i word that right?)

Stuff i could gather:
- There's a length multiplier which scales with max combo. It scales linearly up to 3000 combo, after which it's reverse logarithmic scale.
At 375 combo, the multiplier is 1. At 3000 combo the multiplier is 1.35 after which it scales slowly
- The formula for the miss multiplier is (0.97^no_of_misses)
- The formula for combo multiplier is ( player_max_combo ^ 0.8 / beatmap_max_combo ^ 0.8 )
- The formula for acc multiplier is ( acc ^ 5.5 )
(a HR fc on owen tag4 would be ~1300pp according to this)
- AR gives a 10% bonus for each point above 9 and 2.5% bonus for each point bellow 8
- HD gives 5% bonus for AR10-11 and an additional 7.5% for each point bellow 10
- FL gives 35% bonus and reapplies the combo multiplier again ( a 3000 combo map gets 82% more for FL, a 375 combo map gets 35% more)
- NF gives 10% penalty;

PakaChan wrote:

About the open sourced thing: The difficulty calculator wasn't open sourced, it's just the pp part after the difficulty was already calculated. (did i word that right?)

Stuff i could gather:
- There's a length multiplier which scales with max combo. It scales linearly up to 3000 combo, after which it's reverse logarithmic scale.
At 375 combo, the multiplier is 1. At 3000 combo the multiplier is 1.35 after which it scales slowly
- The formula for the miss multiplier is (0.97^no_of_misses)
- The formula for combo multiplier is ( player_max_combo ^ 0.8 / beatmap_max_combo ^ 0.8 )
- The formula for acc multiplier is ( acc ^ 5.5 )
(a HR fc on owen tag4 would be ~1300pp according to this)
- AR gives a 10% bonus for each point above 9 and 2.5% bonus for each point bellow 8
- HD gives 5% bonus for AR10-11 and an additional 7.5% for each point bellow 10
- FL gives 35% bonus and reapplies the combo multiplier again ( a 3000 combo map gets 82% more for FL, a 375 combo map gets 35% more)
- NF gives 10% penalty;

The interesting part for me now is what penalty Easy has, and why the Effect of DT overweighs it, is it purely based on the AR? or the Mod-Multiplier

I don't believe AR does anything to Star Rating, so I think it's just how much harder DT makes it in the case of EZ turning hypers into ridiculously hard pixel-jumps and DT making the remaining ones even faster.

AR bonus (no mod) could be for just AR9>
AR bonus (HD) 15% AR0-7
AR bonus (HD) 10% AR8-9.5
AR bonus (HD) 5% AR9.5>

The bonuses for HD are fine as they are aside from around 9.5-10, it goes down to too little at that point, and should maybe only drop to around 7% at AR10

PP has failed us....

love the FL boost to pp its great

Yu_Makari wrote:

love the FL boost to pp its great

FL isn't the easiest thing tho. Do you have any tips for FL?

Does anyone know why the score multiplier of DT is lower than that of HR, considering that an SS with DT gives more pp than an SS with HR - or even HDHR, for that matter.

Just wondering, because if someone gets a DT FC then does a HR FC of the same map, later, they'll lose pp

Cipse wrote:

Does anyone know why the score multiplier of DT is lower than that of HR, considering that an SS with DT gives more pp than an SS with HR - or even HDHR, for that matter.

Just wondering, because if someone gets a DT FC then does a HR FC of the same map, later, they'll lose pp

8 years ago a bad thing happened... More seriously, 8 years ago when the multipliers were decided the mapping meta was quite different with higher CS and lower AR. On these sort of maps HR is harder than DT but mapping has since evolved while scoring and mods have stayed the same.

There was some work done behind the scenes by Tom94 to allow for you to have more than one score on the same map for different mod combinations. Your highest score would show up on the leaderboards but you would have the pp from your highest pp play but I believe this has been shelved due to a large increase in time to retrieve scores ingame.

ScoreV2 currently has a 1.1x multiplier for HR and a 1.2x multiplier for DT. This would mean even a HDHR score would be beaten by just DT but we don't know how far away it is from being implemented into the game fully. In the meantime you just have to avoid overwriting DT scores with HR, though HDDT does beat out HR so if you can play HD not all is lost.

119410501 wrote:

Remove PP rankings.

Fucking 3 years ago still true AF