forum

Performance Points feedback and suggestions (Taiko)

posted
Total Posts
645
show more
Jikson[m9]
I would like to avoid the situation that " I could perform more pp in this map, but those performances are of less score than my best score with lower pp, so I couldnt improve my pp by this map"
TimmyAkmed

Quiesce wrote:

Does a certain playstyle confer advantages to playing the game? This is the question we're asking.
In my opinion yes, if you are using the kddk set-up and playing full alt you can play very easily every 1/2 on 300bpm maps and if the map don't have lots of 1/4 it's very simple. But if you singletap everything it'll become way harder/ impossible.

Since this playstyle is the most "optimal" the difficulty should be calculated with the playstyle that makes the game easier imo.
AnFace

TimmyAkmed wrote:

Quiesce wrote:

Does a certain playstyle confer advantages to playing the game? This is the question we're asking.
In my opinion yes, if you are using the kddk set-up and playing full alt you can play very easily every 1/2 on 300bpm maps and if the map don't have lots of 1/4 it's very simple. But if you singletap everything it'll become way harder/ impossible.

Since this playstyle is the most "optimal" the difficulty should be calculated with the playstyle that makes the game easier imo.
i'd say full alt players usually have lower accuracy but orukaa would strike me down


i'm late to the party but i think rankings would be boring without converted maps since tvsize hr SSx3000
i've generally been avoiding ranked taiko maps for the last couple months because 95% of them just aren't fun

i could understand converted being worth less pp than taiko due to ranking guidelines and things but you can't ignore the fact that some converted maps take just as much if not more skill* to play
*there is no clear definition of skill, i personally see it as acc + reading/playing patterns (especially high speed)

these 2 maps took hours worth of tries and i consider them to be some of my best scores
https://osu.ppy.sh/p/beatmap?b=272454&m=1
https://osu.ppy.sh/p/beatmap?b=127846&m=1

or take this map, 1 try SS that i wouldn't even blink at if there weren't so few others
https://osu.ppy.sh/b/287463
this should not be worth more than the above maps just because it's specifically for taiko
lolcubes
That's because most of the ranked maps are OD5. Longer maps with OD6 and OD7 aren't really that easy to SS. :p

Oh look my map is up there. But that's actually not that easy to SS with HR ;_;
AnFace
maybe unpopular opinion but i feel like if SS is possible then OD is too low kevin doesn't count or map is too easy. od5 is always too low, od6 is fine if it's a hard map. od7+ for easier maps

these 2 are things i'd like to see more of for taiko maps
https://osu.ppy.sh/b/284037 hard
https://osu.ppy.sh/b/72460 od8
lolcubes
There is nothing wrong in map being "too easy", though. I do agree it shouldn't be valued as much obviously.

Let's see how this rating thing comes up. A proper map difficulty algorithm can really help out here.
AnFace
too easy is bad if it's also boring
doesn't have to be both!

i am super curious about how maps are valued right now and what they'll be like later. SS is way too good right now and i feel like #1 was too good before. if you have a really good score on a map and someone gets a better one, old score is still good even if it's not the best

i'd also like to see global rankings to see how wrong things are right now
-crn- rank 76???????????
lolcubes
Well, currently, the system isn't really good, but that's why this thread is here. :p

I'm #481 which makes no sense, since after I changed my playstyle I can't even FC your generic Onis haha.
WemadeFOX-solo
i was #2k2, then after doing some insanes/hards on converted maps, i started climbing back to >6k and now >10k, pretty sure this is not working as intended and this is not the finally state of pp
LunaticP
actually it is not working yet, and this is the reason we are talking here
RaneFire
Collapsed because incorrect
So I read the whole thread... and this patterns stuff seems to be going back and forth.

I should probably mention how bad I am first, so take my post as a long-winded opinion. Shoot me down if you must, but I'm interested in taiko nonetheless..

The ranking system needs to rank all players. I'm practically a newbie, despite playing occasionally in the past, only playing muzu's and trying to delve into oni's occasionally, but it's hard. The reason I can't play those oni's is because of the non-repetitive patterns which are very randomised long streams or randomised short patterns with gaps. (Random = non-repetitive) These are just too hard... But when a pattern repeats, no matter where, even in a long stream or with a gap, it is very easy to read and play, even with hand changes. It just needs to repeat only once, and reading is far easier.

Ebawer's example of how you break up reading patterns is correct... I'm noob and already do it. I think other people's examples of longer patterns involving hand changes is subjectively wrong, although their examples are correct. However this is not due to hand-changes, but due to reading complexity, determining when the pattern changes, requiring a hand-switch. You don't read 12 notes at once, you read them in pieces.

Instead of determining which patterns are easier, we should give all patterns a static difficulty based only on their length, up to 5 or 6 notes (including streams)... and rather than rank their difficulty, instead create an algorithm for identifying these patterns (and lengths) and only weight the randomness of the patterns used, i.e. repetition vs non-repetition. You could get more into the specifics of breaking-up the easy ddddkkkk patterns or dkdkdkdkdk, but that's not hard to do.

I think including 62 patterns into some difficulty hierarchy algorithm is going to be far too much work to balance correctly between play styles. The randomness between these patterns is the only thing you can work with. Pattern groups with a random/non-repetitive nature that go on for longer would be worth more "complexity" score. It should also carefully consider odd and even pattern groups though, like having 1 pattern repeated in between a different pattern every time.

This would also solve your auto-convert problem... I think, because they're objectively broken. The hard ones would stay hard, and the easy ones with 1/8 spam would be worth nothing. Ignore rolls and spinners too from calculation... I don't think those have anything to do with skill... just mashing speed.

I'm wearing my bulls-eye T-shirt today.
LunaticP
SPOILER

RaneFire wrote:

So I read the whole thread... and this patterns stuff seems to be going back and forth.

I should probably mention how bad I am first, so take my post as a long-winded opinion. Shoot me down if you must, but I'm interested in taiko nonetheless..

The ranking system needs to rank all players. I'm practically a newbie, despite playing occasionally in the past, only playing muzu's and trying to delve into oni's occasionally, but it's hard. The reason I can't play those oni's is because of the non-repetitive patterns which are very randomised long streams or randomised short patterns with gaps. (Random = non-repetitive) These are just too hard... But when a pattern repeats, no matter where, even in a long stream or with a gap, it is very easy to read and play, even with hand changes. It just needs to repeat only once, and reading is far easier.

Ebawer's example of how you break up reading patterns is correct... I'm noob and already do it. I think other people's examples of longer patterns involving hand changes is subjectively wrong, although their examples are correct. However this is not due to hand-changes, but due to reading complexity, determining when the pattern changes, requiring a hand-switch. You don't read 12 notes at once, you read them in pieces.

Instead of determining which patterns are easier, we should give all patterns a static difficulty based only on their length, up to 5 or 6 notes (including streams)... and rather than rank their difficulty, instead create an algorithm for identifying these patterns (and lengths) and only weight the randomness of the patterns used, i.e. repetition vs non-repetition. You could get more into the specifics of breaking-up the easy ddddkkkk patterns or dkdkdkdkdk, but that's not hard to do.

I think including 62 patterns into some difficulty hierarchy algorithm is going to be far too much work to balance correctly between play styles. The randomness between these patterns is the only thing you can work with. Pattern groups with a random/non-repetitive nature that go on for longer would be worth more "complexity" score. It should also carefully consider odd and even pattern groups though, like having 1 pattern repeated in between a different pattern every time.

This would also solve your auto-convert problem... I think, because they're objectively broken. The hard ones would stay hard, and the easy ones with 1/8 spam would be worth nothing. Ignore rolls and spinners too from calculation... I don't think those have anything to do with skill... just mashing speed.

I'm wearing my bulls-eye T-shirt today.

Totally disagree sorry.

First don't call a patten random, no patten is random only if you map in random mood and make it shit. Every patten have its meaning and effect in the map. If the map continue to repeat a patten it looks shit already. And I assume mapper talking in here are not mapping shit.
If you really think a patten is random you should play more and you will not think in that way.

Second in a player's angle, I actually don't know how many times the patten(e.g ddk) have appeared. I play when I see the patten, no matter if that patten appear before. The most extreme case is this.

And on your logic, you think that the program should only check how many times a patten appeared? Then the example map given is really easy right? Please explain on your point on this.
Luna

RaneFire wrote:

I think other people's examples of longer patterns involving hand changes is subjectively wrong, although their examples are correct. However this is not due to hand-changes, but due to reading complexity, determining when the pattern changes, requiring a hand-switch. You don't read 12 notes at once, you read them in pieces.
Nope, this is not true.
Patterns that don't require leading hand switches are automatically more regular and repetitive. And I don't read streams as a single unit either, I do break them up into parts (albeit using a different "algorithm" than EBAWER); when playing long Firce-style deathstreams I hardly ever miss because I misread, I miss because I mechanically mess up a leading hand switch.
I'd also like to repeat that I don't think every stream with hand switches is extremely hard; I already mentioned before that something like kddkddkddkddkddkddkddkddk is simple due to repetition. It's the irregularity of pattern / hand switches that makes a stream truly hard.
RaneFire
LunaticP

LunaticP wrote:

Totally disagree sorry.

First don't call a patten random, no patten is random only if you map in random mood and make it shit. Every patten have its meaning and effect in the map. If the map continue to repeat a patten it looks shit already. And I assume mapper talking in here are not mapping shit.
If you really think a patten is random you should play more and you will not think in that way.

Second in a player's angle, I actually don't know how many times the patten(e.g ddk) have appeared. I play when I see the patten, no matter if that patten appear before. The most extreme case is this.

And on your logic, you think that the program should only check how many times a patten appeared? Then the example map given is really easy right? Please explain on your point on this.
Thanks. I guess I do need to play more taiko to recognise the mapped rhythm better.

Anyway, for the example:

Is not the hardest part executing those patterns because of the speed? It's hard to sync two hands that fast for many people, because that skill is mental, so if the map were 180bpm, would you call it easy then? (Not saying I can play it though)

Also my point was on the adjacent repetitive nature of patterns because of group reading, like in the start. Sure the "ddk" pattern appears a lot, but the sections where it is divided between 1/2 notes makes it harder doesn't it, because the repetition is broken? Or are you saying it's the other way around?

Luna

Luna wrote:

Nope, this is not true.
Patterns that don't require leading hand switches are automatically more regular and repetitive. And I don't read streams as a single unit either, I do break them up into parts (albeit using a different "algorithm" than EBAWER); when playing long Firce-style deathstreams I hardly ever miss because I misread, I miss because I mechanically mess up a leading hand switch.
I'd also like to repeat that I don't think every stream with hand switches is extremely hard; I already mentioned before that something like kddkddkddkddkddkddkddkddk is simple due to repetition. It's the irregularity of pattern / hand switches that makes a stream truly hard.
Yes, I'm trying to pinpoint that irregularity with an algorithm. You make a good point, but it could also be a playstyle difference, and that's where I was trying to stem my argument from... because no one wants to agree. There's arguably more main-hand players though and alternaters have an easier time with hand-switches. Oh well... I tried.

IMO I treat the idea of reading and execution as one and the same. If you can't do it consistently, you still have trouble reading it. "Mechanical" may just be the word I was looking for. In osu!std I have this problem, and there's only notes and sliders there... much easier by comparison, but there are still problems because of "mechanical" issues with odd/even numbers of notes.
Luna

RaneFire wrote:

IMO I treat the idea of reading and execution as one and the same. If you can't do it consistently, you still have trouble reading it.
By that logic I can't read 400 BPM monocolor streams
RaneFire

Luna wrote:

RaneFire wrote:

IMO I treat the idea of reading and execution as one and the same. If you can't do it consistently, you still have trouble reading it.
By that logic I can't read 400 BPM monocolor streams
You know what I mean. The definition of the word was the problem. Stop being stupid about this just because I'm not a pro taiko'er. It's a game, just because I'm not good at it does not mean I'm stupid.

It was also a clarification that did not embody the message I wanted you to read.
Luna
I'm not trying to make fun of you or anything, I just disagree on that point.
There are many things that I can read perfectly fine, but can't do mechanically. That's just how it is.
LunaticP
Just answer this , which one you think is easier?

1. ddk ddk ddk ddk ddk

2. ddk d k ddk d k ddk

3. ddk ddk d k ddk ddk
RaneFire

LunaticP wrote:

Just answer this , which one you think is easier?

1. ddk ddk ddk ddk ddk

2. ddk d k ddk d k ddk

3. ddk ddk d k ddk ddk
1. ddk ddk ddk ddk ddk

Not counting speed limits. Also note that this only has to do with complexity score, nothing else. If you know osu!tp, you'll know std has aim/speed/acc. In Taiko, for example, it would be complexity/speed/acc. If you fill the whole map with ddk, the complexity may be lower, but the speed score will be higher because of density.

I'd like your explanation though, because throwing examples around just clutters the thread. Or at least shoot a hole in my logic that I can see, and let's be done with it.
Luna
Actually, for me ddk d k ddk d k ddk... is the easiest out of those patterns. Can't really explain why though, it just flows better than ddk ddk ddk ddk...
ddk ddk d k ddk ddk... would probably feel awkward with a 4/4 signature
AnFace

Luna wrote:

Actually, for me ddk d k ddk d k ddk... is the easiest out of those patterns. Can't really explain why though, it just flows better than ddk ddk ddk ddk...
yea this
i think just because the single notes are easier to watch timing on
or 5 note pattern repeating instead of 3
lolcubes

LunaticP wrote:

Just answer this , which one you think is easier?

1. ddk ddk ddk ddk ddk

2. ddk d k ddk d k ddk

3. ddk ddk d k ddk ddk
#2 for me. For my old really weird playstyle #1 would be easiest but not on really high bpm. Now it's #2 regardless of the bpm, unless I am not really able to do that bpm in general.

RaneFire wrote:

Yes, I'm trying to pinpoint that irregularity with an algorithm. You make a good point, but it could also be a playstyle difference, and that's where I was trying to stem my argument from... because no one wants to agree. There's arguably more main-hand players though and alternaters have an easier time with hand-switches. Oh well... I tried.
Don't feel offended, however you still can't grasp the true difficulty of oni maps (and higher level oni maps) just yet. :)
The thing is, seeing is one thing, hitting what you see is another. Mechanics used to hit certain things, handswitches especially are really complicated to execute. It looks simple, but it's a really pain in the ass thing to do. People who fully alternate for a long time and are really good at it (like EBAWER) don't feel this mechanic problem as much because it's just natural thing to do, but there are not many people on that level.
Let me put it this way, it's not hard after an insane amount of practice, however that just says how hard it actually is, by seeing how much practice you need with it.

Even on low bpm maps, when I see something like [dddkkkd] at 1/6 inside or at the end of a stream, or even isolated, I still can't hit it. It's really complicated to adjust your rhythm and to perform a handswitch at the same time, at my level. Will take a lot of practice for me to perform it well.
RaneFire
Yeah, I figured my inability to read oni's properly would probably be a flaw. So there's more going on than meets the eye. It's gonna be hard to create system around the subjectivity of each pattern though. Thanks for the replies, some responses are always better than none.

lolcubes wrote:

#2 for me. For my old really weird playstyle #1 would be easiest but not on really high bpm. Now it's #2 regardless of the bpm, unless I am not really able to do that bpm in general.
I'd be interested to know what, or why, you changed. Was there something "wrong" with your previous style? Is there a superior playstyle overall?

I do play full-alternating in Taiko, so maybe this takes a little longer to get into oni's, but full-alternating always felt more natural to play. Watching some of Orukaa's DT plays was even more of an inspiration to go for it since he was so fast and accurate. I was pretty much looking for an excuse to play it and he gave me one.
lolcubes

RaneFire wrote:

lolcubes wrote:

#2 for me. For my old really weird playstyle #1 would be easiest but not on really high bpm. Now it's #2 regardless of the bpm, unless I am not really able to do that bpm in general.
I'd be interested to know what, or why, you changed. Was there something "wrong" with your previous style? Is there a superior playstyle overall?
I swapped to full alternate because I was unable to do streams longer than 9~ notes (well, monocolors or streams which had 2 same colors if you divide a stream on 2x2x2x2 etc were actually fine) or even most simple chains on bpm higher than 150 (kdddkdddkdddkdddkdddk etc) because my playstyle was to cheese every pattern combination in a way I find it easier to play, so every 3~5 note (except for dkkkd, kkkdd and dkdkk, those were really impossible) was really easy to perform for me in any situation, but put them inside a stream and I'm done for.
Cheesing patterns required a certain way to play them, and streams are just ongoing notes, you can't really "think" on which pattern you have to play while actually playing it, you need some buffer time (I did at least). With training you can get around it I guess, but because of my 2nd problem (chaining simple patterns in a stream) I chose to swap, because the way I learned to play kdd really destroyed my ability to chain kdddkdddkdddk, making my hands entangle and spend insane amounts of strength just to do it badly. You can see it in some of my vids I posted how I cheese the patterns. :p

Now with full alternate, I just hit them naturally and when I see an even amount of dons, I know the next kat is a hand switch.

RaneFire wrote:

I do play full-alternating in Taiko, so maybe this takes a little longer to get into oni's, but full-alternating always felt more natural to play. Watching some of Orukaa's DT plays was even more of an inspiration to go for it since he was so fast and accurate. I was pretty much looking for an excuse to play it and he gave me one.
Yeah, I think Orukaa is an inspiration to everyone. :p

While I can still use my old style to play some maps really nicely, I choose not to because full alt is infinitely more fun to me.
LunaticP
Orukaa is not a human at all,no need to kill your hand because of this....

Now the true question come,and this may brings problem because of two common playing style

Select the one you think the hardest

1. ddk k k ddk k k

2. ddk d d ddk d d

3. ddk d k ddk d k

4. ddk k d ddk k d

This result will be put into my program as well
WemadeFOX-solo
#3 and #4, but they are the same just like #1 and #2
Luna
#4 is the hardest for me, then #2, then #3, #1 is the easiest
LunaticP
I have a reason to ask this....

This is to show not only a stream, but the notes after the stream actually make the map looks different. We should not only focus on stream only

Anyway please answer that, because I really want to know the main trend of playing styles
Nashmun
For me it's the same. Maybe #2 would be harder at very high bpm
lolcubes
4->1->2->3 for me, but thats cause i find kat heavy patterns harder than don heavy ones.
EBAWER123

Luna wrote:

By that logic I can't read 400 BPM monocolor streams
ofc you can't
WemadeFOX-solo
So i keep losing ranking by playing normals and fcing it with 99~100%, is this part of the new pp system which try to promote to play harder beatmaps or it wont be part of the new pp system and is just being not functional?
EBAWER123

WemadeFOX-solo wrote:

So i keep losing ranking by playing normals and fcing it with 99~100%, is this part of the new pp system which try to promote to play harder beatmaps or it wont be part of the new pp system and is just being not functional?

not functional, play for fun!
WemadeFOX-solo
yeah, thats why i keep playing normals and hards, because they are fun :>
Jikson[m9]

LunaticP wrote:

Orukaa is not a human at all,no need to kill your hand because of this....

Now the true question come,and this may brings problem because of two common playing style

Select the one you think the hardest

1. ddk k k ddk k k

2. ddk d d ddk d d

3. ddk d k ddk d k

4. ddk k d ddk k d

This result will be put into my program as well
2>1=4>3
TimmyAkmed
I agree with the discussion about "HD = HR" in term of difficulty but if it means that "DT+HD = DT+HR"
That sounds not correct to me because as said earlier HR is harder on High BPM and HD harder on Slow BPM most of time.
When you add DT the BPM is almost everytime more than 200 and it makes DT+HR way harder than DT+HD.
imo (and sorry if this has been already said) DTHR should be way more rewarding than DTHD.

Anyway, what I say would have no sense if you already works on HR high bpm more rewarding than HD high bpm...
Yuzeyun
Not telling that 95% of taiko maps have their OD >=5, which in DTHR gives you 10 at least, up to 12.
Dolphin
I've been reading through this whole thread and I find myself agreeing with OnosakiHito more than anyone else.

Put to put it simple from me:

BPM, the higher the BPM is, the harder it usually gets.
Patterns, if there is more pattern variety in a single stream the harder it is (complexity)
OD, as it requires more accuracy the higher it is.
HP shouldn't be counted too much, but higher HP drain makes you lose more and gain less HP per note.
Mods. DT highest. HR and HD same. FL disabled due to difficulties with DS players and cheaters.
I wouldn't want to disable HDHR since I find it very playable on BPMs lower than
Standard converts should also be disabled since they're not made to be played on Taiko mode.
No, I'm not saying that they can't be hard, I've played a handful of difficult converts that were fun too, but it doesn't change the fact they were made for Standard, not Taiko. If you think disabling it is too drastic, then at least they should count as FAR less than Taiko specifics.

I put my thoughts into this very lazily so please yell at me if you want me to explain something more or you disagree with me.
Pretty sure all these points have been stated so dont mind me \o/
BrokenArrow
I still haven't seen a reasonable argument on why standard converted difficulties should not be weighted equally under the same circumstances as taiko specific difficulties should.
"They are not meant to be played in taiko mode" is just stupid. Might as well remove rankings on converted maps completely then. If they are so bad that they shouldn't get taken into account for the pp ranking what is the point in them existing? Besides, those opinions are always based on subjectivity, which shouldn't influence the ranking at all.
For example, if a converted map basically just goes d d k d d d k [...] just treat it the same way you would treat a taiko specific map going like that (even though that's unlikely to happen but you get the point) instead of straight up removing them entirely from the ranking.

Sorry if I repeated anything what has already been said, I haven't followed this thread at all.
lolcubes
There are so many Normal diffs which get converted into something so insanely dumb that I don't even want to start describing how they look. And yes they look harder than insanes, the only easier things are OD and HP.

"Not meant to be played in taiko", how is that not reasonable?
As long as standard converts don't meet the ranking criteria (and never will), I vote against them being counted towards pp. While some may convert well, most of them don't. Imagine if any retarded taiko specific maps are allowed to get ranked. It's kinda the same.
show more
Please sign in to reply.

New reply