ATTN: Scoring System Needed

[Patashu]

how indeed....

[TC_Halogen]

Originally Posted by Patashu

but as you just pointed out not all notes are created equal; most of almost there is easy ****, so why should those notes be worth as much to hit?

oh how about a GH style combo scorer but with a far far shorter limit; that is, 1.0x for a 0 combo, 1.1x for 1 combo...2.0x for 10 combo and up. that doesn't make bursts of cbs quite as bad as single cbs in varying places

You'd have to have quite a large scoring spread for grades because you'd have to remember that in order to get in the upper tiers, you'd have to be holding a constant string of notes compared to combo breaking just before the 2.0x boost.

I do agree though that if a person is struggling throughout the song though, the GH style multiplier is pretty nice of a thing--all you'd have to do is separate the DP%s a bit farther apart.

A thing about SM's machine scoring system that I'm sure everyone has noticed by now, every arrow hit is worth double the last one. So it already has some sort of multiplier, the only thing is that it doesn't reset.

With that being said, trying to incorporate GH style multipliers will probably not work because it's beyond SM's limitations scoring wise. You will not be able to define the maximum score on any given song unless you get 100% DP, and even then, it more than likely won't end up fitting the scoring system that SM has (difficulty x 10,000,000?).

The passmark idea is probably the easiest thing to control because the metrics of which the lifebar goes up and down can be changed, and even better, the grading percentage that you want can also be changed. The only thing about the passmark system that might be hard to do is to somehow incorporate the grading bonus in accordance to lifebar.

[abstractrevenge]

Originally Posted by Xiaounlimited

Osu! has a concept called Kiai Time that mappers can implement into their beatmaps. What it allows is the mapper to set a certain timeframe for Kiai Time to activate, up to a certain limit. During this timeframe, the score multiplier is increased (iirc it's 1.1 or 1.2) for the duration of it. Something similar could be done here so while notes would be of equal worth, steppers (charters, mappers, w/e) could add this into sections that would be worthy of it (i.e. Almost There's dumpstream) and therefore have a more profound effect on score.

This is a really cool idea, but I still think your number score is completely obsolete and should be done away with. Scores should be compared on DP because the only thing other than your DP/MA/CBs that should matter is your combo.

It might be cool if there was a spot like this Kiai Time where it keeps track of your specific stats during the hard parts.

This would be kind of like guitar hero scoring breaking the files into sections so you can see how you did on the hard parts compared to easy parts. This would be a better way to earn bragging rights from the hard parts rather than your score which is just pointless numbers.

[Patashu]

oh it's double? man I didn't know that haha

[stargroup]

Originally Posted by Xiaounlimited

Osu! has a concept called Kiai Time that mappers can implement into their beatmaps. What it allows is the mapper to set a certain timeframe for Kiai Time to activate, up to a certain limit. During this timeframe, the score multiplier is increased (iirc it's 1.1 or 1.2) for the duration of it. Something similar could be done here so while notes would be of equal worth, steppers (charters, mappers, w/e) could add this into sections that would be worthy of it (i.e. Almost There's dumpstream) and therefore have a more profound effect on score.

while this is a nice concept keep in mind the amount of work that needs to be put into it in terms of scripting, and also notice that since users set this timeframe, you're putting the complicated, intricate scoring matters into the hands of the amateurs

[leonid]

Actually I don't care much, but I prefer simple scoring systems.

I'd say FFR scoring system without combo bonus

[[TeRa]]

Similar to ITG, I suggest making SM % based. Change it so marv/perf are different and make holds not weigh so heavily.
For example:
Marv +4
Perf +3
Great +1
Good -1
Boo -3
Miss -5
OK +2
NG 0
Mine -2

I don't really care what you decide to do with grade letters, it's just a less accurate way to look at a score.

[TC_Halogen]

Originally Posted by [TeRa]

Similar to ITG, I suggest making SM % based. Change it so marv/perf are different and make holds not weigh so heavily.
For example:
Marv +4
Perf +3
Great +1
Good -1
Boo -3
Miss -5
OK +2
NG 0
Mine -2

I don't really care what you decide to do with grade letters, it's just a less accurate way to look at a score.

Originally Posted by stargroup

I don't want just random numbers. If you're gonna suggest something, make sure you have good reasoning or solid mathematics to show why your system would be suitable. I am willing to provide any information that I have currently.

>_>

[[TeRa]]

I didn't put much thought into those numbers anyway lul. I typed those like 2 months ago, I'm too lazy to explain right now, plus I'm about to go out. I'll actually think about values and explain them later though. Ignore that post for now, I'll most likely change most the numbers anyways.

[stargroup]

I'm thinking of taking out marvelous timing altogether

there's no point in it and it just makes scores revolve around ma

then again, a lot of files do have lots of colored notes. but now the question is: do we need marvelous timing to force players to distinguish the colored notes?

[gnr61]

perhaps not for that exact reason, but i believe we do still need marvelous timing; singles is a game in which scores will literally be all but indistinguishable from one another on the vast majority of files without it. in solo there appears to be a looot more variation in peoples' great and/or cb counts on nearly every chart, whereas in singles most commonly accepted "good" charts are the ones almost everyone AAAs.

[fysx]

Ok yeah lets see whats goin on in here

Originally Posted by stargroup

also this should be geared more towards cbs, not ma, because it doesn't matter how accurately you're hitting the note as long as it's in time and you're feeling the chart. that's how iidx and pop'n work

I support this wholeheartedly, though balance is very important in this situation. We don't want to end up with a system like stepmania, where accuracy is a little too ridiculous, and we don't want to go the way of o2jam

if we focus mainly on cbs, we run the risk of charts getting AAAA'd or whatever too easily, and this only leads to denser and denser charts to keep up with increasing skill levels.

I personally think that popn's timing system is WAY too lenient, but I don't really know what the criteria should be for a good timing system.

Naturally any sort of progressive scoring system is needlessly complicated, and not worth it. Keep it simple, and all that.

[Kraezymann]

I'm thinking of taking out marvelous timing altogether

there's no point in it and it just makes scores revolve around ma

then again, a lot of files do have lots of colored notes.

Which is where I think the cb bar or penalty bar would be handy. Because if you have something like A Crimson Rose and you mess up with the rainbow notes even at the beginning, you could lose the score if you aren't hitting the notes properly.

[Patashu]

definitely do not take out marvelous timing. the reason why you get so much better at stepmania than you do at ffr in the same period of time is partially because of how finely graduated the timing windows are; it gives you more and faster feedback on how well you're doing and allows you to learn quicker
even if it's just a slight visual effect as you play that doesn't add to score keep it in somehow. in fact I wouldn't mind a protiming variant where the colour bursts/judgement text were coloured over a gradient depending on exactly how well it was

[stargroup]

alright I talked to fysx

we're not going to take out marvelous for at least 4 key. it's also possible the marvelous timing won't even affect score, that it's only there to distinguish the great scores from the amazing scores and nothing else

we need marvelous timing because it makes it so that the players can feel the difference of the colored notes (otherwise what's the point in using greens rofl) and also because otherwise almost all the files will be too easy for some people (the hardest files we'll include in the game are probably easily AAA'ed by good players)

[Patashu]

OK I was thinking
firstly ever notice how as the judgements get better the words get longer?
marvelous
perfect
great
good
boo

well it made me think; the more feedback a game gives back to you and the less you have to do to process it the faster you can get better at it (think stuff like tetris the grand master 3 playing a unique sound depending on what piece is next), so what I propose is a pro-timing based system that makes the judgement as detailed as possible.
ideally it would be applied per column, but one for the whole thing should be fine.
it would have two components; one is that the length of the judgement would go up linearly proportionally to how far your hit was from the true note placement, the other is that the colour would change along a gradient at the same time, again linearly, but it would go in a different direction depending on whether you hit early or late.
early would go like this: white for spot on to aqua green for about perf to aqua for about great to lightish blue for about good to deep blue or purple for about boo
late would go like this: white for spot on to lime green for about perf to lime for about great to yellow for about good to orange for about boo
there might not even need to be a distinct set of timing windows any more; all you need is a 'best' timing cutoff beyond which it would still give the visual effects but not grant any more points, a combo breaker cutoff if it still exists (alternatively, just a point at which a hit stops gaining life and starts losing it) and a 'worst' timing cutoff, akin to the far end of boos now. again, within that we don't necessarily NEED to have timing windows; we could just display the results as a frequency graph and show the average path through it. alternatively we could just subdivide it finely or something

anyway the problem with perfs, greats, etc now is that if you just get one you don't immediately know if it's early or late, if it's in the middle of the window or almost one better or almost one worse. the boundaries between timing windows are essentially arbitrary so this makes more sense to do and provides better feedback. for instance sometimes I find myself cb rushing a trill, but I cannot just at a glance determine if I am a note ahead or a note behind on it, and on which column or both; with column specific judgments that showed different colours or some other indication of being early or late I could very easily adjust myself to hit on time again

also for some reason I play better when the 'star' type target flashes turn on, so maybe have something like them turned on at all times

also turn off O.K.s they're just spammy and annoying lol