Drezi wrote:

dennischan wrote:

1. There is no pattern recognization.
2. Fast single taps are underrated.
3. High SV is underrated.

And how exactly does the current score system do a better job at evaluating these aspects?
You go on about this and that being underrated, which is a problem when comparing scores between different maps, for which pp is already being used so I don't see your point... When comparing scores on the same damn map it doesn't matter if the map is underrated in absolute value compared to other maps at all, every score on that given map is treated the same way, and their relative order is prefectly fine and more representative than score, how is that not obvious.

dennischan wrote:

4. HD and FL are not suitably considered in the program
It is not reasonable to put an arbritary multiplier on a map.

FL1,12 HD1,06 -.-

Kaeru wrote:

To whoever said HD doesn't help with pp, that's complete nonsense. HD can give a ton of pp. A lot of my top plays were worth almost nothing before I put HD on (for example, this map).

I still don't think pp should replace score. I am for the idea of replacing global scores with ones that are worth more performance points (I hate it when I set a record that's worth more pp but it doesn't get counted because my score isn't as high). I just don't feel like every mod is weighted fairly in every instance. In fact, I feel that it's impossible for the system to be so perfect that it isn't possible to set a more impressive score worth less pp. Score is a flawed system too, but it's the best system we have.

The problem with the later notes being weighted heavier (getting 100s near the end of the song hurts your score more) can be easily fixed by slightly altering the way score is calculated (though it would have to be possible to change old scores too or this would be pointless).

I don't know about other people here, but I don't want to see 85% DT scores in the top 50 on maps. I don't care how hard it is to play a song with DT.. you can just mash your way through most maps. HR takes a lot more skill than people give it credit for, and it is currently weighted far too little in my opinion. I don't even want to get into FL.. that's a whole different story in itself. I am totally against abandoning the old score system. The idea is just ridiculous.

The "standards" for score have been set a long while back, and completely messing them up now (not to mention the leader boards) is unfair to everyone.

Also, Kyou-kun calls him/herself a "mid-tier" player at rank 180..? Are you so incredibly egotistical that you fail to see that you're biased towards this idea because of how high your rank is? I don't mean that as a personal attack, but you are in the top 0.1% of everyone that plays this game.

dennischan wrote:

What I meant was that since pp is not a good calculation of actual difficulty between maps

dennischan wrote:

For one, the pp system only considers a score's max combo and accuracy when weighing scores, and that makes the pp system not suitable for weighing scores since it works with too limited information.

dennischan wrote:

Also did you not know that DT is considered a map changing mod and thus the pp calculator treats the DT-ed map as a new map? This is why what I said do matter, in fact, since DT just boosts up the speed of the map, and does not make the complexity of the map increase. Without pattern recognization, it is inevitable that the DT-ed map would be treated as a very hard map which in reality its not so hard as it seems, since the pattern keeps being simple

dennischan wrote:

Actually although map difficulty is calculated by looking at the whole map, the method of map difficulty calculation is rather simple and cannot be a fair basis for score calculation.

1. There is no pattern recognization.
The pp calculating method simply takes note of the difference between notes and treats it as the so called strain on a map.
Sadly, in the process there is only a simple weighing of the strain and simple addition.
Therfore, DT wiuld be overvalued in the sense that it simply makes easy patterns faster. Hard patterns would be at a disadvantage on this point.
This is why skystar maps and chloe maps are of relatively low rating since their difficulty lies in reading difficulty, not in so called aim.

2. Fast single taps are underrated.
There is no way for our simple pp calculator to distinguish between single taps and streams. Therfore fast single taps are underrated.

3. High SV is underrated.
pp calculations do not take sliders in account. Thus maps which are difficult because of sliders are underrated.
such as Talent Shredder, Strange Program

4. HD and FL are not suitably considered in the program
This ties with what I have said about no pattern recognization. The main difficulty of HD and FL is that they make the map harder to read.
It is not reasonable to put an arbritary multiplier on a map. After pattern recognization is finished, the diifculty of HD and FL can be gauged in a better way.
In this state of pp calculation, using pp as score will be unfair.

I dont deny that pp calculation can be in time, a fair substitue for the score system, but currently it is still a work in process.
I urge people in this thread to see that pp is currently not a suitable system for weighing any scores.
But I am not remotely concerned about this being implemented since it is impossible for our osu community to agree on this issue.
I know that this message might be ignored by people in this thread who only care about DTing Hards, but I hope that you people can check out my points before mindlessly dismissing it.

xasuma wrote:

jesus1412 wrote:

wow. this guy. you're so wrong. it hurts me to read on. every post just hurts me more.

And why would that be, Mr. Jesus ?

You can't just say, "wow so wrong, I can't believe this guy" in an argument. That's not a valid argument.

jesus1412 wrote:

I do indeed like to read dennischan's posts they're well formed and researched. I honestly can't find many flaws in his arguments other than "the scoring system does that one aspect even worse".

Ziggo wrote:

That'd be awful. I hope this will never happen.

simplyparanoid wrote:

Ziggo wrote:

That'd be awful. I hope this will never happen.

I agree with this.

Tom94 wrote:

I feel like keeping score as a stable base measure is not such a bad thing. Especially with the possibility of mod-specific highscores (no more losing pp by beating your score with mods but worse accuracy) and optionally sorting the top50 pp-wise I don't see any reason why this wouldn't suffice.

nocipher wrote:

It is a disservice to advocate for the separation of scoring and pp. They should be very interrelated. In particular, a higher score should always mean more pp. The current implementation is silly. There are many forum posts asking why people lost pp after improving their score. Most of the time the response is to chastise them for not searching before asking such a common question.

Many of these people will be players just starting to be invested enough in osu! to pay attention to ranking. According to Peppy, one of the major goals of the pp overhaul was to encourage progression, but the current system runs counter to that. After you set a decent score on a map, it is in your interest to actually not play that map again until you have improved significantly. Otherwise, you risk disappointment after a few missteps on on otherwise excellent play (for you) leave you a few thousand ranks below your starting point. This does not encourage to keep people playing osu! and fixing it should be made more of a priority.

jesus1412 wrote:

nocipher wrote:

Goodwork.

Making pp the scoring system would fix everything you posted and your idea that a higher score should always mean higher pp is the worst thing I have ever read and shouldn't be read by anyone else.

TheVileOne wrote:

It's essentially sorting the scores by accuracy with a secondary weight of combo. This will practically reverse the importance of score.

TheVileOne wrote:

If you get a few misses, might as well retry because your pp gain will be next to nothing. Your modded play with misses will be worth as much pp as No mod scores or even less.

TheVileOne wrote:

Double time scores can make tons of mistakes and still get higher pp than SS with mods. SS ranks will be booted off leaderboards for B rank DT scores.

TheVileOne wrote:

Don't even try to go for leaderboard unless you can high accuracy with a given mod or use DT, which tends to be overrated.

TheVileOne wrote:

Also I'm not sure how you're going to calculate pp as you play in real time. Leaderboards are in game as well. PP would have to replace score as an ingame scoring system.

TheVileOne wrote:

Your modded play with misses will be worth as much pp as No mod scores or even less. Don't even try to go for leaderboard unless you can high accuracy with a given mod or use DT, which tends to be overrated.

Double time scores can make tons of mistakes and still get higher pp than SS with mods. SS ranks will be booted off leaderboards for B rank DT scores.

NixXSkate wrote:

Oh my god no, this won't work at all. ppv2 is sooo bad at accurately measuring the skill required for FL.

The current score system needs change but ppv2 has not evolved enough for it to be the answer.

Drezi wrote:

Uhm guys, ppv2 is already the global ranking system? Hello? Why are you talking about it as if it was some kind of half-assed attempt, and not a pretty reliable system? It has flaws but is generally accepted as being a lot more accurate than score. Do you really belive that a DT SS is exactly 1,12x times harder than a nomod SS? And that a score with plenty of 100s at the beginning is worth more than one with a single 100 at the end?

It had been accepted as the global ranking sysytem because it's pretty good already, and pp works especially well when we're talking about comparing scores on the same map, since the algorythm underweighting certain map types has no effect here, unlike on global ranking. Those B rank DT scores already give more PP, and thus contribute more to your global rank than some SS ranks, why should they have a lower place on map leaderboards in that case? If a B rank DT score gave more pp, than more probably than not they were harder to obtain, especially since the pp system doesn't like rewarding "bad performances" such as passes on hard songs with with lots of misses, combobreaks etc.

You keep pointing out it's flaws not realizing how infinitely more flawed the score system is.

NixXSkate wrote:

You can't just change the scoring system that has been used since the beginning just because a system that may be a bit better just came out, it's still flawed.

NixXSkate wrote:

Ppv2 is great for measuring physical skill, but not mental skill or spinning. Hidden has no effect toward accuracy (only aim), and flashlight gives a similar percentage bonus for aim to all maps, there is no intelligence given in them to reflect accurately the skill they require for different maps.

NixXSkate wrote:

Also, did you just ignore the point I made about Dungeon's score? Do you have no argument for that?

Drezi wrote:

I don't think it's off too much, FL bonus needs only a minor buff, and the pp ranking order would be the same as the current one.

But I think the main point of this game is to have fun being able to sightread and play a huge variety of maps, not to sit down and memorize one single map for hours and days. Everyone is better off focusing on improving their universal, overall skill, rather than spending a lot of time to memorise each map in order to rank high on it. So I don't think FL should be the best way of earning top ranks, even if memorisation was actually more demanding (it is not), than the physical skill required for a DT play on the same map.

I guess everyone supporting this idea would be fine with that compromise of having separate leaderboards for both pp and score though.

TheVileOne wrote:

Also I'm not sure how you're going to calculate pp as you play in real time.

DT-sama wrote:

TheVileOne wrote:

Also I'm not sure how you're going to calculate pp as you play in real time.

You assume you're going to SS the map, and start with the maximum pp the map can give you with the active mods. As your accuracy gets worse, it scales down the pp score accordingly, same for misses.

BRBP wrote:

CalignoBot wrote:

Playing FL doesn't require any more skill than no-mod. All it requires is more patience.

I'm dying to hear your definition of skill.

2. If pp were ever to update its calculations, it could affect (possibly significantly) the placement of scores on the leaderboards. However, as far as I know, no significant changes have been made to either tp or pp in a long time, so I don't think this would be too much of an issue.

You assume you're going to SS the map, and start with the maximum pp the map can give you with the active mods. As your accuracy gets worse, it scales down the pp score accordingly, same for misses. As for combo breaks, you subtract your combo from the current highest achievable combo (which is FC at the start of the map), and the pp score considers the highest between the two. Missed slider ends just subtract one from the current highest achievable combo.

NixXSkate wrote:

Drezi wrote:

I don't think it's off too much, FL bonus needs only a minor buff, and the pp ranking order would be the same as the current one.

But I think the main point of this game is to have fun being able to sightread and play a huge variety of maps, not to sit down and memorize one single map for hours and days. Everyone is better off focusing on improving their universal, overall skill, rather than spending a lot of time to memorise each map in order to rank high on it. So I don't think FL should be the best way of earning top ranks, even if memorisation was actually more demanding (it is not), than the physical skill required for a DT play on the same map.

I guess everyone supporting this idea would be fine with that compromise of having separate leaderboards for both pp and score though.

That picture proves that it's significantly flawed, it's off by a truckload. People who rarely play flashlight fail to understand that flashlight isn't just memorization, it's also practice because it's just such a mental overload. Even if YOU don't think flashlight takes much skill, the fact is it DOES, ESPECIALLY at higher BPMs, fast jumpy maps, and with DT. No player can just casually FL a map even if they know where every note and slider is in a map. We don't think in four dimensions, playing FL takes a lot fast recall and rote memory, which makes focusing on accuracy significantly harder because of the focus you need for aim. Don't even get me started about how calm and collected you have to be with FL as well since it takes so much focus, sometimes I thought I was going to have a heart attack, and none of my FL scores are even that great. I'm sure everyone could agree that with the amount of plays Dungeon put in World's End, he could have HD+DT SSd it multiple times easily. But, with FL he was challenged so much to just combo the thing that he couldn't possibly dedicate so much focus to the accuracy. In terms of maps that are actually challenging in aim and speed, FL is more of a PENALTY or DISADVANTAGE in ppv2.

Also, if a player is playing DT on something, they know what to expect from the leaderboard, nothing is stopping them from adding hidden to beat the HR+HD players. It's not like someone has to rely on some calculation to determine how good they are on a leaderboard, they know exactly what they expect based on how they play and who's on the leaderboard, no surprises.

I honestly think a better solution to the problems of the current scoring system is to lower the hidden multiplier to 1.03x or 1.04x, and flashlight to 1.09x or 1.10x, rather than use ppv2 for scoring.

Tom94 wrote:

There is still the one issue of overwriting a score with another one that gives less pp, but that would be mostly eliminated by per-mod highscores. The cases where you beat your score but lose pp would almost completely cease to exist,

Ace3DF wrote:

HRHD would become more pointless and DT mashing would be happening everywhere.

CalignoBot wrote:

Ace3DF wrote:

HRHD would become more pointless and DT mashing would be happening everywhere.

This hasn't happened with the current scoring system so it won't happen in the new one either.

Kaeru wrote:

Why do people have a problem with the score system? It's the best system. It's a constant system. It never changes. Performance points will constantly vary as the calculations become more accurate. Score never has to change, nor should it ever change. I really would like to see a way to replace plays with higher pp plays (or delete records, maybe a supporter only feature) but this is really not a solution.