PyaKura wrote:

Even if you say that it's already too late to implement it and I can hardly imagine a different "ranking system" just for a single game mode (especially since o!m is only 2 years old) since osu!'s been around for quite a while now.

Doesn't matter how long o!m's been out, peppy wouldn't change it anyway. :/

Honestly, PP isn't the worst way to rank players... Imagine how dumb ranks would be if they relied on "ranked score" instead of PP (oh wait, they did years ago, and it was awful).

Ranked score was better than pp, because it didn't pretend to be a proper ranking system. It was a motivation to play because it was going up no matter if you did well or badly, doing well just made it go up faster - unlike pp, which awards you points at first and then slows down, leaving players frustrated and demotivated. Why do we need a proper ranking system? Just a quick look at a few leaderboards tells you who's the best at the game. To fix lowdiff farming, ranked score could be scaled just like mania charts were.

Tear wrote:

Ranked score was better than pp, because it didn't pretend to be a proper ranking system. It was a motivation to play because it was going up no matter if you did well or badly, doing well just made it go up faster - unlike pp, which awards you points at first and then slows down, leaving players frustrated and demotivated. Why do we need a proper ranking system? Just a quick look at a few leaderboards tells you who's the best at the game. To fix lowdiff farming, ranked score could be scaled just like mania charts were.

If pp (or rank score, or any other value used for rank) went up noticeably whether or not you did well, then lowdiff farming wouldn't be avoided (unless the scaling from difficulty increases extremely fast with difficulty, for example, the easiest map in the game gives 1 point, a mid-tier map gives 10^35 points, and a top-tier map gives 10^45 points). The purpose of the decrease of pp gain rate from plays is limiting the amount of pp obtained from playing many maps of similar difficulty without increasing the difficulty over time. If the difficulty calculation of the pp system is accurate, then players will keep getting pp as long as they increase the difficulty (or performance) of the maps they play (because of that, a good difficulty calculator for osu!mania maps is very important for making the pp system better).

Looking at beatmap leaderboards of hard maps can give you an idea of who are the best players in the game, but you can't use that method to know the skill ranking of players that aren't as good as those top players (a well designed pp system should be able to give a good idea of that).

There's one very simple change that would makes all modes of osu fun, addictive and still rewarding:

Disable pp weighting.

Let it accumulate as you play. Even if you didn't do your best, you still get a few points. Unlike score, pp goes up exponentially so lowdiff farming would be severely reduced.
"But I can just play all the easies in the game for #1!" And pros can play all the insanes for #1, what's your point?
Any attempt to make an objective skill rating is doomed to fail, so why not make the game fun instead? Note that pp without weighting is identical to DJ Points in LR2 and IIDX. Also note that my solution does not render Tom's work obsolete at all, pp as a measure of a score's worth is still used to reward skill and make you go up faster than playing easier stuff.

Tear, that's just plain wrong. You want everyone to play tons of autoconverts to get their score up? That's going back to ppv1, which as we know sucked dicks. Current pp system is okay-ish, if we'd get some attention and some fixes to formula, we'd have a realy good system. But oh,well.

Don't we all agree that autoconverts should be unranked, lol

Unrank autoconverts and bam, current system is 2 times better. Also, you should know by now that no one is going to change anything

Yeah it's just pointless banter, I get bored.

As frustrating as it is, I actually like the weighting on PP. It ensures that you don't have inflation on PP as players play more and more songs, making the ranking system much more stable, which is a major requirement for any half-decent ranking system.

One of the real flaws in the system is that it only uses your top scores in calculations. This incentivizes those players who are willing to sit there and spam a song until they get that fluke personal best. But at least that takes much more effort than just spamming out a bunch of low PP scores (which would happen if the weighting were removed with no other changes).

I actually think that current system is a good base onto something really awesome. Good and wasted base.

-Kamikaze- wrote:

I actually think that current system is a good base onto something really awesome. Good and wasted base.

Can't say I disagree with that. At least I can say I tried.

Okay, so far, in here, I've heard a lot of whining, and not really a lot of meat for suggesting.

bluh bluh ban all autoconverts

First, what ARE autoconverts? I'm not a "mania" purist - I'm casual as crap, I only really started getting a smudge more pp-happy in the past like... ...month or so, and even then, I just want to try to keep over 100K - not an overly hard goal.
Second, assuming that there's mania-specific maps, removing "autoconverts" removes a lot of viable maps, and you may as well kiss 99.9% of the song list goodbye with that kind of mentality. "OH LOOK EVERYONE, LET'S MURDER THE SAME TWENTY SONGS!" isn't something I really relish sitting behind, no offense intended.
osu! is about fun, and osu!mania shouldn't be much different, and singling things out isn't entirely fun. Taiko's singled out, and every attempt I make to play that mode casually ends either in frustration, disappointment, or otherwise feeling inept - the casuals are kinda singled out.

rank _k scores separately

How about "no"? There's debuffs for altering the Ks, and the bigger difference in Ks translates into a bigger dip in overall score. I appreciate the sentiment, but I think that the debuff that we currently have is good enough to keep things from getting too farm-happy for everyone.

Now then, onto my ideas (note, they're guesses, I don't know what the actual algorithm entails):

1) Develop a formula for average density, and pace - to factor into the possible "stars" (let's face it, having "slow parts" will make a difference no matter what, "skipping" them requires database/resubmitting tweaks that I don't think the dev team wants to really do). ie. [*] = note density x (bpm/120)
2) "amplify" the current star rating? I'm seeing literally NOTHING over 6 stars in my list, and I think that several maps are potentially of 7+ star status.
3) (may also be under 'gameplay', and since I don't know the origin of osu!mania, this can be ignored) Alter how 'spinners' function (perhaps similar to Taiko, but limited to just the 4K bindings for spinner input)? I notice that in really spinner-happy maps, they turn sad-easy.

...And I went braindead. Oh well, current base is good.

Omio9999 wrote:

I only really started getting a smudge more pp-happy in the past like... ...month or so, and even then, I just want to try to keep over 100K - not an overly hard goal. Removing "autoconverts" removes a lot of viable maps, and you may as well kiss 99.9% of the song list goodbye with that kind of mentality.osu! is about fun, and osu!mania shouldn't be much different, and singling things out isn't entirely fun.

Why are you talking about pp's if you're playing casually ? The pp system - as half-arsed as it is - is a competitive element in the game. If you're playing casually nothing prevents you from playing autoconverts for fun. By the way, we are talking about unranking the autoconverts so they don't give any pp at all, not straight out remove them.

Omio9999 wrote:

rank _k scores separately

How about "no"? There's debuffs for altering the Ks, and the bigger difference in Ks translates into a bigger dip in overall score. I appreciate the sentiment, but I think that the debuff that we currently have is good enough to keep things from getting too farm-happy for everyone.

Following what I said previously, you can only apply nK-mods to autoconverts. Mania-specific maps are not affected by those mods. (Not to say pretty much everyone farms on mania-specific maps since they usually give way more pp's).

Omio9999 wrote:

2) "amplify" the current star rating? I'm seeing literally NOTHING over 6 stars in my list, and I think that several maps are potentially of 7+ star status.

If your map list is mostly, if not entirely made out of autoconverts, then it's completely normal. They are usually far easier than mania-specific maps (or retardedly hard for a few of them). There is currently only one mania-specific ranked map rated over 7 stars which would Imperishable Night 2006, and a few maps over or around 6 stars (off the top of my head, Akasha, Intersect Thunderbolt, Shuffle Heaven, Zirkfied...). There are loads of 5 stars maps as well.

Anyway, I don't want to offend you but it seems your current o!m experience was mostly based around autoconverts. FYI, autoconverts as you might have guessed, are automatically converted by the game from osu!standard maps. They're okayish for first-timers, but past that point everyone is better off playing mania-specific because of the way they are structured mapping-wise. You may not see what I'm talking about atm but I can assure you that autoconverts feel REALLY WRONG compared to mania-specific maps (pitch relevancy, layering, SVs... autoconverts fail at all of that and more).

I didn't bother replying to your entire post because your view on the whole thing is uuuh... out of place I guess, due to your lack of experience, and thus so are your statements.

That said, the only real downside for mania-specific maps is that the mappool is tiny compared to osu!std, but there are enough ranked maps and tons of unranked which keep the mode alive and fun for most players.

I suggest you play some more o!m on mania-specific maps to give you a better idea of the current situation. Have fun

Bobbias wrote:

One of the real flaws in the system is that it only uses your top scores in calculations. This incentivizes those players who are willing to sit there and spam a song until they get that fluke personal best. But at least that takes much more effort than just spamming out a bunch of low PP scores (which would happen if the weighting were removed with no other changes).

On the other hand, everyone receives the same incentive, so "zero out".

I admit I also spam the song to get a lucky personal best, but its not like there will be a 3% difference between the spam in a single day. And like it or not, spamming is practicing.

It is practice, but not all practice is equal. Spamming is practice on the specifics of that particular map, not on maps as a whole, so it will not help you learn as quickly for other maps as randomized practice does. It can help if that map contains specific patterns you have trouble with, but it still results in slower general skill improvement. Especially if it's grinding for SS scores. All that does is make you marginally more consistent on easy songs. In the 3 months I focused on that kind of improvement in stepmania I found that I actually got worse at harder maps because of lack of practice on them.

Ideally though I'd prefer the system to track every score a player makes, so it can see how consistent a player is.

you can't ask consistency in players because the next day they could be drunk while playing. or ate a lot of carbo and meat. or had an injury. or anything.

heck I'd even wager that a change in room temperature affects the player's stamina. most people I know irl would feel their wrist/hands like "frozen" when its cold or getting hit by your standard electric fan.

kidlat020 wrote:

its not like there will be a 3% difference between the spam in a single day. And like it or not, spamming is practicing.

I sometimes get even 5% diffrence IN A SINGLE DAY, without any weird stuff going on. Also spamming is training, but it only trains your speed and you lose a lot of you accuracy in the process. I know that because I've experienced it myself (having lower 300g/300 ratios than 5 months ago on Utakata for example)

-Kamikaze- wrote:

kidlat020 wrote:

its not like there will be a 3% difference between the spam in a single day. And like it or not, spamming is practicing.

I sometimes get even 5% diffrence IN A SINGLE DAY, without any weird stuff going on. Also spamming is training, but it only trains your speed and you lose a lot of you accuracy in the process. I know that because I've experienced it myself (having lower 300g/300 ratios than 5 months ago on Utakata for example)

We were talking about spam playing the same song over and over, not anmitsu/spamming the keys.

And yeah, I've had random runs where I get like 5% better acc later the same day too.

I've had random runs where I get like 5% better acc later the same day too.

that only means that's your real play.

You don't make sense :s

that 5% increase in a single day wasn't luck. that was his true ability [at that day].

this had always been a problem in statistical data. flip a coin 100 times. same principle.

I don't get'cha, most of my top pp scores are pure luck with spamming and I can't get remotely close to beating them (with exception of Utakata, but I've played it like 100 times)

-Kamikaze- wrote:

I don't get'cha, most of my top pp scores are pure luck with spamming and I can't get remotely close to beating them (with exception of Utakata, but I've played it like 100 times)

This is exactly what I was talking about. Random lucky scores happen. I've had scores that took 9 months before I could beat them because of that. If you spam play songs, you have an advantage because you can play each song until you get a lucky score, then more of your top scores are likely to be lucky scores.

I've had scores that took 9 months before I could beat them because of that.

you already improved.

I don't consider those lucky scores as my real level.

Taking 9 months to catch up to a lucky spammed score isn't really improvement following your logic.

I can beat every my old spam/not spam score (around 3months) every try, learn to play instead of spam lol.

Tidek wrote:

I can beat every my old spam/not spam score (around 3months) every try, learn to play instead of spam lol.

To be honest, if it wasn't for kind-of broken o!m mechanics and my attitude of clearing hard stuff>getting good accuracy, it would probably be easier for me to do so. Still, lucky scores happen and it's not easy to beat them.

Tidek wrote:

I can beat every my old spam/not spam score (around 3months) every try, learn to play instead of spam lol.

Not everyone learns at the same pace. I've been playing for over a decade, and there are plenty of players with far less time who are far better than me. So just because you can beat your old scores in some amount of time doesn't mean anything for anyone else.

it doesn't disprove anything either. because, records say that you've improved even though you yourself don't feel it is so.

even if its a lucky run.

"I did a fluke 700k run on that map and I won't be able to beat it in 3 months, but I've improved"
So you're saying that lucky run shows our true skill and then we go backwards?

kidlat020 wrote:

it doesn't disprove anything either. because, records say that you've improved even though you yourself don't feel it is so.

"I've run 100m in 11 seconds once, but now I'm running it in 15 seconds. Still, I'm an olympic class runner."
Well in o!m lucky spams don't show ANYTHING. You can literally spam a map to 700k without even looking. If you call that an improvement, then I'm outta here.

"I've run 100m in 11 seconds once, but now I'm running it in 15 seconds. Still, I'm an olympic class runner."

yeah sure. who's to prove (or disprove) that said athlete ate too much sugar or carbs the next day that would interfere with his performance?

or that he wasn't really "lucky" but actually in top shape (correct sugar intake and carbs)?

this doesn't apply to mania only. even std and taiko would.

seriously, if you guys think practicing is the only factor playing here, then I'm outta here.

good emotional and physical health plays an important role in anything, but for something like osu, i dont see how anything beyond keeping away from harmful illegal drugs and eating/sleeping right would affect your performance positively [aside from practicing].

I'm overweight (despite being 3k rank) so I'd know that even the smallest excess of fat in my fingers is interfering with my hitting the keys.

anything that's double note hit is my greatest enemy. jackhammers are my death.

#healthissues

Overall health will certainly help. It's well known that a healthy body will have a positive effect on the brain, and that things like eating breakfast can have a positive effect on outcomes later that day. However, I have a feeling that these factors are very small, and have more of a cumulative effect than any immediate effect on any specific day.

I currently have a score on https://osu.ppy.sh/s/239943 which is 100k higher than my other scores (roughly 9% accuracy above my average score). It will likely take me months before any one of my average scores will match that level. It's certainly possible that I could beat that score next time I play that map, but the chances are extremely small.

Unless you're entozer, everyone finds jackhammers hard.

*Cough Jxyden 2000 note per minute 1/8 + 1/16 jack master (song is plasma strike btw, its an Ftb chart/map)

Seriously though practice is still what matters most. What dictates the skill improvement rate of a player depends in the type of charts he plays (pattern specific, ranked, or comfort zone stuff) + (or times, its up to the reader) his actual talent or learning speed.

And about best plays, they DO NOT always (sightreads taken into consideration) show a player's actual skill level. It just shows his potential and that improvement is possible.

Overall body condition does help. Boredom, sleepiness, hunger, weak body, and any oher impediment does affect a players gameplay. The magnitude varies from player to player. Same goes for positive conditions.

Pls fix this!This map gives more pp with 8k mod then the original one on 7k 8k-> http://puu.sh/enkVN/ba74a51de6.png ( pp amount http://puu.sh/enlmH/0d4700d736.png ) 7k -> http://puu.sh/enkYj/f12e732f9f.png ( pp amount http://puu.sh/enlzN/40a08c1f90.png )

^ Is that a bug?

If so, please remove it as quickly as possible or better yet just make auto-converts unrankable

If not, then there needs each difficulty of each converted beatmap needs to be treated as a separate entity depending on key-mod, which would be difficult to implent so just unrank the auto-converts please.

Tristan97 wrote:

^ Is that a bug?

If so, please remove it as quickly as possible or better yet just make auto-converts unrankable

If not, then there needs each difficulty of each converted beatmap needs to be treated as a separate entity depending on key-mod, which would be difficult to implent so just unrank the auto-converts please.

nope it isnt this map on 8k has 6,XX* and on 7k it has 4,XX* and the pp ranking depends on how much * the map has so the 8k version is obviously more pp worth then the 7k version

Pipe Dream's 8K autoconvert is rated a lot higher than the 7k because it has this


not really a bug, but the way hitsounds are converted