How fast would the cheesiest player need to be?

[Dynam0]

Prof Dinmo is bored at work...the topic for today's discussion is manipulation:

With perfect pattern manipulation skills, what is the minimum speed a player must have in order to AAA most of the game?

Suppose there is a nerd -we'll call this player "Mr. Cheese"- who is super slow and simply cannot increase the speed of his fingers, no matter what he does! He is capable of incredible feats of manipulation, such as the long one-handed trill in Flight of The Bumblebee. Most of the time one would need to trill it but every few notes there are some frames that you could hit a [12] jump and still get perfects. The result would be some 1212 [12] 1212 [12] sort of motion which is probably way harder than actually trilling it but still...this is Mr. Cheese we're talking about and he's a goddamn legend! Just how fast is Mr. Cheese though?

FFR's current framerate is 30 frames per second, or 1800 frames per minute. This means you could cram a stream of 1800 notes all separated by one frame in the span of a minute. This translates to a 450bpm 16th note 'stream'. Any streams faster than 450bpm will convert improperly and start to have jumps scattered in them.

FFR's perfect window is 3 frames wide. This is important to Mr. Cheese and he will use this window to his advantage whenever possible.

Let's look at a notoriously demanding section of the infamous chart RATO and see how Mr. Cheese handles it:


The sequence [12]22 is a killer! But not for Mr. Cheese, below is what he sees:


0000 (start of perfect window for Note 1)
1200
0000
0200
0200
0000 (end of perfect window for Note 3)

The first two notes are separated by two frames and the last two are separated by one frame. (2-framer connected to a 1-framer)

The range of the perfect window for this sequence is 6 frames. Mr. Cheese, playing as slowly as possible, needs to hit a note on Frame 1, Frame 3 or 4, and then Frame 6. In essence, Mr. Cheese will need to make 3 inputs separated by 2.5 frames each. We've already established that one frame is equivalent to the space between 16th notes in a 450bpm stream. 2.5 frames would be the space between notes in a 450 / 2.5 = 180bpm 16th stream, meaning Mr. Cheese has to execute a 3-note 16th jack at 180bpm that starts and ends frame perfectly in order to cheese this pattern.

That doesn't sound all that bad does it?

In terms of one-hand trilling, any notes separated by more than 2 frames must be trilled correctly. This means Mr. Cheese's one-hand trill speed must be at least 450 / 3 = 150bpm.

Homework: Find a section that is more demanding than the above example and describe how Mr. Cheese would attack it. You cannot use Whimper Wall, vRofl, Crowdpleaser v1, P4U v1, etc.

[SubaruPoptart]

Re: How fast would the cheesiest player need to be?

I'm fully expecting an account named "Mr. Cheese" to pop in here at any point with the answer.

[CammyGoesRawr]

Re: How fast would the cheesiest player need to be?

A thread about how Hakulyte can AAA every chart

[gold stinger]

Re: How fast would the cheesiest player need to be?

the implication here is that mr.cheese can do 150bpm 16th quadglut/handglut/jumpglut flawlessly. That in itself, deserves D8-worthiness, regardless of the manip.

[Wiosna]

Re: How fast would the cheesiest player need to be?

The range of the perfect window for this sequence is 6 frames. Mr. Cheese, playing as slowly as possible, needs to hit a note on Frame 1, Frame 3 or 4, and then Frame 6. In essence, Mr. Cheese will need to make 3 inputs separated by 2.5 frames each. We've already established that one frame is equivalent to the space between 16th notes in a 450bpm stream. 2.5 frames would be the space between notes in a 450 / 2.5 = 180bpm 16th stream, meaning Mr. Cheese has to execute a 3-note 16th jack at 180bpm that starts and ends frame perfectly in order to cheese this pattern.

Are you sure that this is correct? I'm very sure that you can hit the 2-framer pattern that you screenshotted at 150bpm 16ths.

When jacking at 180bpm, each note is separated by 87.5ms.
The earliest possible input where Mr. Cheese can get a perfect is -50ms. 
The notes are placed at 0ms, 66.6ms and 100ms respectively 
(frame 2, frame 4, frame 5 in your example).
 
Time	Input	Difference
0ms	-50ms	-50ms	
66.6ms	37.5ms	-29.16ms
100ms	125ms	+25ms

Given that there is an additional +25ms of leeway for 2 inputs (2nd and 3rd note of the jack), 
Mr. Cheese can jack 12.5ms slower per note and would still obtain perfects.

87.5ms + 12.5ms per note = 100ms per note = 10 notes per second = 150bpm 16ths.

The issue in your calculation, I believe, is that the gap between the start of frame 1 and the end of frame 6 is 6 frames, not 5. The start of frame 1 and end of frame 6 constitute virtually no frames at all, since they're specific points in time and not periods of time.

With that said...

Homework: Find a section that is more demanding than the above example and describe how Mr. Cheese would attack it. You cannot use Whimper Wall, vRofl, Crowdpleaser v1, P4U v1, etc.

Sinthasomphone has 2 5-note jacks (e.g. combo 805), with each note in the jack being separated by 2 frames (66.67ms) each. Assuming that Mr. Cheese hits the first note of the jack in the earliest part of the perfect window (-50ms), Mr. Cheese has 100ms leeway over 4 notes. Combining the leeway per note and the distance between each note, Mr. Cheese has to jack at about 91.67ms per note, or about 163.63bpm 16ths. This is strictly harder than a 150bpm 16th 3-note jack.

I'm sure there are a few other examples, but the general idea is to find charts with practically consecutive 1-frame minijacks with 2-note gaps (e.g. xx-x-xx) or a 2-frame jumpjack but the jump is separated by 1 frame each (since Mr. Cheese would have an 66.67ms window instead of a 100ms window -- the optimal position to hit these kinds of grace jumpjacks would be 16.67ms before the first note since the second note comes 33.33ms after), but I can't think of many concrete examples.

[Dynam0]

Re: How fast would the cheesiest player need to be?

Are you sure that this is correct? I'm very sure that you can hit the 2-framer pattern that you screenshotted at 150bpm 16ths.

When jacking at 180bpm, each note is separated by 87.5ms.
The earliest possible input where Mr. Cheese can get a perfect is -50ms. 
The notes are placed at 0ms, 66.6ms and 100ms respectively 
(frame 2, frame 4, frame 5 in your example).
 
Time	Input	Difference
0ms	-50ms	-50ms	
66.6ms	37.5ms	-29.16ms
100ms	125ms	+25ms

Given that there is an additional +25ms of leeway for 2 inputs (2nd and 3rd note of the jack), 
Mr. Cheese can jack 12.5ms slower per note and would still obtain perfects.

87.5ms + 12.5ms per note = 100ms per note = 10 notes per second = 150bpm 16ths.

The issue in your calculation, I believe, is that the gap between the start of frame 1 and the end of frame 6 is 6 frames, not 5. The start of frame 1 and end of frame 6 constitute virtually no frames at all, since they're specific points in time and not periods of time.

With that said...


Sinthasomphone has 2 5-note jacks (e.g. combo 805), with each note in the jack being separated by 2 frames (66.67ms) each. Assuming that Mr. Cheese hits the first note of the jack in the earliest part of the perfect window (-50ms), Mr. Cheese has 100ms leeway over 4 notes. Combining the leeway per note and the distance between each note, Mr. Cheese has to jack at about 91.67ms per note, or about 163.63bpm 16ths. This is strictly harder than a 150bpm 16th 3-note jack.

I'm sure there are a few other examples, but the general idea is to find charts with practically consecutive 1-frame minijacks with 2-note gaps (e.g. xx-x-xx) or a 2-frame jumpjack but the jump is separated by 1 frame each (since Mr. Cheese would have an 66.67ms window instead of a 100ms window -- the optimal position to hit these kinds of grace jumpjacks would be 16.67ms before the first note since the second note comes 33.33ms after), but I can't think of many concrete examples.

Looks like I was forcing Mr. Cheese to hit the middle of each frame and not use the extra half frames on either end of the sequence. A half frame is 16.6ms for a total of 33.3ms extra time.

Only thing is 180bpm 16ths = 83.3 ms between notes, but overall the logic is sound.

Why does 150bpm 16ths seem too slow to hit that pattern though? I might need to mess around in iso to get a better idea of the feel but that is def right!

edit: Props for completing your homework too

[Wiosna]

Re: How fast would the cheesiest player need to be?

Looks like I was forcing Mr. Cheese to hit the middle of each frame and not use the extra half frames on either end of the sequence. A half frame is 16.6ms for a total of 33.3ms extra time.

Only thing is 180bpm 16ths = 83.3 ms between notes, but overall the logic is sound.

Why does 150bpm 16ths seem too slow to hit that pattern though? I might need to mess around in iso to get a better idea of the feel but that is def right!

edit: Props for completing your homework too

Whoops, you're right. I wasn't using a calculator for 1000 / 12 and it showed there...

But amending that,

Time	Input	Difference
0ms	-50ms	-50ms	
66.6ms	[B]33.33ms	-33.33ms[/B]
100ms	[B]116.67	+16.67ms[/B]

Given that there is an additional [B]+33.33ms[/B] of leeway for 2 inputs (2nd and 3rd note of the jack), 
Mr. Cheese can jack 16.67ms slower per note and would still obtain perfects.

83.33ms + 16.67ms per note = 100ms per note = 10 notes per second = 150bpm 16ths.

We still have the same answer. You'll obtain the same answer through this method as long as you use any BPM as some baseline, since the gap between notes while jacking at a certain speed + the amount of leeway that Mr. Cheese obtains from jacking at that speed is always a constant.

[mi40]

Re: How fast would the cheesiest player need to be?

mos def - mathematics

[xXOpkillerXx]

Re: How fast would the cheesiest player need to be?

This is quite an important topic for FFR difficulty computation. I've spent a lot of time thinking about it. That being said, I think a major factor component to not forget is what exactly is meant by "cheesing" a pattern. It seems in this context you're purely defining it as:

"Maximizing average note distance (in time) + maximizing the number of jumps (or minimizing the number of hits per hand)"

This definition sadly will never apply in practice, since it doesn't consider what I believe is probably the most important thing: one-handed complexity. Your FotBB is a Great example of how that applies; I would almost always argue that on a one hand basis only (spread playstyle implied), 1212 [12] 1212 [12] is harder than 121212121212 and [12] [12] [12] [12] [12] [12]. The change from trill motion to jack motion is, imo, not something that can be left out if we're gonna talk "cheese".

When going down the rabbit hole of making the above definition as accurate as possible, I think it often boils down to setting some baseline of what the majority of people think is easier to hit legit and what isn't. From there, you can start building a more robust definition. The hard part of course is deciding what the smallest building block of your system is, and their possible relationships (mathematically). For example, your smallest block could be a simple tuple of consecutive notes on the same hand (1 [12], 2 1, [12] [12], etc), a more complex n-tuple on same hand, a tuple that includes both hands (12, 24, [13][23], [123]3, etc), and so on. That decision I think will likely need to be backed up with stats on what the body can do at what speeds on average (compared to enough statistical data from the vsrg population), and that's just a full on research project.

Fwiw, I tend to believe that one hand side 4-tuples are pretty robust blocks as they include most hand motions and aren't a massive list (2^4=16 possible blocks). There's more to it than just that, but just keepin it short here.

So, to answer your question:

With perfect pattern manipulation skills, what is the minimum speed a player must have in order to AAA most of the game?

I think there's no answer to this as long as "perfect manipulation skills" isn't well defined.

[Hakulyte]

Re: How fast would the cheesiest player need to be?

Just a quick glance at RATO.

Cheese report says that colors cannot be trusted and arrow conversion is strange.

I feel like the 24th in this song are not supposed to transform into jumps all over the place, but I'm not even completely sure.

This is an irregular pattern that needs several speed adjustments.



Not the hardest pattern, but conversion.



More weird [14] conversion.



More strange 24th jumps.



Rolls have odd spacing and you may need to integrate two delays in your jumptrilling.



Spooky [123][14] out of nowhere that makes no sense with colors.



Difficult pattern to take advantage of without having enough speed.



Besides that there's Dynam0's screenshot with the framer.
That being said, if you have the speed/control requirement for this song, it's a must play for its difficulty.

OP basically wants to find a pattern that's more annoying to hit than the RATO mini jack ?
I'm a bit lost basically on what you're looking for because we will either end up on the weirdest poorly converted rapid burst pattern or the highest bpm song with a 1 framer mini-jack.

If you look up songs that haven't been AAA'd yet, you will probably find what you're looking for.

With perfect pattern manipulation skills, what is the minimum speed a player must have in order to AAA most of the game?

"most of the game" = ?
There's a huge difference between 95%, 99% and 99.5% here.

[mi40]

Re: How fast would the cheesiest player need to be?

Mr. Cheese wouldn't need to be Mr. Cheese if we didn't have frame based timing

[xXOpkillerXx]

Re: How fast would the cheesiest player need to be?

Mr. Cheese wouldn't need to be Mr. Cheese if we didn't have frame based timing

It could still apply to ms timing, it's dependent on what the goal is. A ms timing FC would still allow for a fair amount of cheese, but a AAA definitely wouldn't.