Looksmaxxing Forum Logo

Join 70,000+ Looksmaxxing Members!

Register a FREE account today to become a member. Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox.

Register Log in
  • DISCLAIMER: DO NOT ATTEMPT TREATMENT WITHOUT LICENCED MEDICAL CONSULTATION AND SUPERVISION

    This is a public discussion forum. The owners, staff, and users of this website ARE NOT engaged in rendering professional services to the individual reader. DO NOT use the content of this website as an alternative to personal examination and advice from licenced healthcare providers. DO NOT begin, delay, or discontinue treatments and/or exercises without licenced medical supervision. Learn more

Info Super-Gaussian Curve for Harmony

genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
Mirin bro ,
Despite I hate maths , this is so good ngl.

Try to make it less complex the next time.
 
Histy said:
Mirin bro ,
Despite I hate maths , this is so good ngl.

Try to make it less complex the next time.
Click to expand...
I cant
Its quite simple really
 
Holy shit, mirin
 
genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
Ow are you certain on the base ratios, and how does the Gaussian formula deal with falios
 
TrueFakeCel said:
Ow are you certain on the base ratios, and how does the Gaussian formula deal with falios
Click to expand...
Sorry this question was retarded as shit
Base ratios are calculated from a sample set of hyper attractive faces and their average
Although they can be adjusted to the user’s opinion on what is correct, so we aren’t certain

The second formula of the thread is how failos are dealt with
You will be awarded negative points if your score is under 5/10
 
  • Like
Reactions: TFC
As one formula without the piecewise bullshit
IMG_3487.webp
 
genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
did
not
fucking
read
after
that
formula
 
genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
n***a i used this and you said it was inaccurate

then faceiq used it and now you're dickriding it jfl
 
foidslayer said:
n***a i used this and you said it was inaccurate

then faceiq used it and now you're dickriding it jfl
Click to expand...
n***a i said your shit was inaccurate because your decay was too high on specific ratios and much too low on others
And it was also inaccurate for a plethora of different reasons
 
genio said:
I think the piecewise works well though
Mayb
Click to expand...
It’s the exact same thing just piecewise is ugly and I hate them

No functional difference apart from mine is 1 formula
 
emeraldpill said:
It’s the exact same thing just piecewise is ugly and I hate them

No functional difference apart from mine is 1 formula
Click to expand...
Oh
I dnrd the formula so 🥶
 
emeraldpill said:
n****r mine mogs
Click to expand...
Probably
I dont mind piecewise functions
Not like anybody here is gonna actually take account this thread into anything they do so doesnt matter
 
genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
Pretty digestible, you formatted the thread in itself really well and explained the formula you presented in a really easy way to comprehend. Cage at the use of CA's old system as a refrence point though :ROFLMAO: I remember when that n***a made it and everyone went crazy, most of the raters at the time just repolished it by biasing it themselves

I swear y=-g+2g(v/a) has been used before though I personally dont use a "super-gaussian curve"

All in all pretty great thread and im happy to see users theorize its really refreshing.
 
dior said:
Pretty digestible, you formatted the thread in itself really well and explained the formula you presented in a really easy way. Cage at the use of CA's old system as a refrence point though :ROFLMAO: I remember when that n***a made it and everyone went crazy, most of the raters at the time just used it and repolished it by biasing it themselves

I swear y=-g+2g(v/a) has been used before though I personally dont use a "super-gaussian curve"

All in all pretty great thread and im happy to see users theorize its really refreshing.
Click to expand...
Thanks for reading big bhai
I gotta feed the greyfags the shitty formulas
Ooh i want to try out y=-g+2g(v/a)
Yeye im a pioneer of the space type shit
 
genio said:
Thanks for reading big bhai
I gotta feed the greyfags the shitty formulas
Ooh i want to try out y=-g+2g(v/a)
Yeye im a pioneer of the space type shit
Click to expand...
n***a said pioneer :ROFLMAO: :ROFLMAO: :ROFLMAO: :ROFLMAO: :ROFLMAO:
 
dior said:
ik it was ironic i still love you though
Click to expand...
If you love me leak a better formula im too lazy to adjust this one i got even though itd be easy :copium:
 
genio said:
It was a joke i know youre the real pioneer bhai
Click to expand...
No I dont use a looks penalty because I think it deflates the rating 2 much, but I actually do know someone who does he goes by GGOB he owns an analysis server on discord but that n***a is far more aspie then me we joined the rating com around the same time.

I know max and a couple other users are working on training an ai model to make a large enough dataset to produce a far better accurate version of the weights in which features hold and I believe they are using some rough form of a "looks" penalty when quantifying it into a formula
 
genio said:
If you love me leak a better formula im too lazy to adjust this one i got even though itd be easy :copium:
Click to expand...
woahhhhhhhh bucky BOYO..... i usually get wined and dined before I get FUCKED.
 
dior said:
No I dont use a looks penalty because I think it deflates the rating 2 much, but I actually do know someone who does he goes by GGOB he owns an analysis server on discord but that n***a is far more aspie then me we joined the rating com around the same time.

I know max and a couple other users are working on training an ai model to make a large enough dataset to make a far better accurate version of the weights in which features hold and I believe they are using some rough form of a "looks" penalty when quantifying it into a formula
Click to expand...
Oou shit
Tbh ive never been in an analysis server ive always found information by myself
That sounds great though
Itd be awesome if the model could creative a bell-curve too ive always wanted an accurate one
I feel like percentiles are more fair and based in reality than “htn” “mtn” blah blah
 
Only reason all the previous formulas were dogshit was because no1 has created a large enough dataset to pinpoint an accurate mean
 
dior said:
woahhhhhhhh bucky BOYO..... i usually get wined and dined before I get FUCKED.
Click to expand...
Yeye whats ur definition of wined and dined
 
dior said:
Only reason all the previous formulas were dogshit was because no1 has created a large enough dataset to pinpoint an accurate mean
Click to expand...
Dont yall js take 1000 chads and measure the fuck out of them
 
genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
Not a single molecule erlite
 
genio said:
Oou shit
Tbh ive never been in an analysis server ive always found information by myself
That sounds great though
Itd be awesome if the model could creative a bell-curve too ive always wanted an accurate one
I feel like percentiles are more fair and based in reality than “htn” “mtn” blah blah
Click to expand...
yeah percentiles will always be far more accurate I can send u over CA's old server and a couple other servers

but ive only ever taken rating as a hobby even though that was the primary community I was influential in, I would say im far prouder with my achievements within looksmaxxing ( actually taking the steps to improve myself ). Hopefully once I go thru my DIY procedures I'll make some guides that mog orkas dogshit (mediocre not dogshit sorry orka) guide that lowk almost botched him :ROFLMAO:
 
genio said:
Dont yall js take 1000 chads and measure the fuck out of them
Click to expand...
lowk kind of but not really, like i said this ai model is going thru fucking thousands of people to try and create the most accurate mean
 
anyways i REALLY need to go to sleep im waking up incredibly soon and I have the beach to slay at
 
dior said:
yeah percentiles will always be far more accurate I can send u over CA's old server and a couple other servers

but ive only ever taken rating as a hobby even though that was the primary community I was influential in, I would say im far prouder with my achievements within actually taking the steps to improve myself. Hopefully once I go thru my DIY procedures I'll make some guides that mog orkas dogshit (mediocre not dogshit sorry orka) guide that lowk almost botched him :ROFLMAO:
Click to expand...
Should i join i feel like a greyfag to the analysis community
tbf i also find rating as js fun even though its my main expertise
Im too lazy to even make my own formulas i got bored after the first line and then i pmd surgerymax got distracted by super gaussian bell curves instead of making my formula
I havent even maxxed much bruh i js grew up a little and my bones got better
@Orka thoughts?
 
dior said:
anyways i REALLY need to go to sleep im waking up incredibly soon and I have the beach to slay at
Click to expand...
Buh bye goat
My discord is @genioplastic btw when ur free
 
dior said:
lowk kind of but not really, like i said this ai model is going thru fucking thousands of people to try and create the most accurate mean
Click to expand...
Js like regular people? I thought it was only mega attractive faces
 
genio said:
90% of you n*****s won't understand this
I didn't write well and it's quite complex

I'd recommend reading this (the Harmony portion of my thread The 4 Pillars of Attractiveness (Revised) beforehand:

1. Harmony
In short, facial harmony is how balanced and proportionate your face is. It uses ratios and math to measure how each feature on your face interacts with other features.

There are many ratios depending on what formula you use as it could up to as high as 40 or more purely off the front profile. Every ratio has a different level of importance.

There are two profiles measured for harmony; front profile and side profile. The frontal profile should always be more important than the side profile no matter the formula. Personally, I weigh frontal harmony as 66.7% of your total harmony and side harmony as 33.3%.

Important ratios include:
ESR, FWHR, Facial Thirds, Canthal Tilt, TFWHR Midface Ratio, Bigonial-Bizygo, etc.

Here's CA's old formula. It's outdated and will probably give you an inflated score, but it serves a little use nonetheless. I'd subtract 10% off the final score personally. Anyways, you just add up all your values depending on what value is assigned to that tier and ratio, then divide it by either 2.625 for the frontal or 1.365 for the side.

View attachment 358959
View attachment 358961

View attachment 358962
View attachment 358963

In harmony, we usually score ratios based off a tier system. It's simple and convenient.
For example:
View attachment 358925

However, in recent times, raters like FaceIQ have used Super-Gaussian bell curves that score your ratios out of 10 based on a decay system.
But how do you do this?
THROUGH THIS FUCKING RETARDED ASS FORMULA THAT TOOK ME WAY TOO FUCKING LONG TO FIGURE OUT!!!
View attachment 358924
a = Plateau Height (I like to use 10, so that you get a value out of 10 based off your measurement)
p = Center Position
w = Plateau Half-Width (the distance from the center (p) to the edge where the flat top ends and the curve begins)
d = Decay Spread (Standard Deviation of the Tails)
x = Your Measurement of Your Specific Ratio

Basically, you set this up, then you put x equal whatever your measurement is (say if you're measuring ESR, you'd put your ESR measurement here).
Then, you should get a score out of the plateau height value (I usually put 10, so I'd get a value out of 10).

Here's an example I used for ESR (Eye Separation Ratio) with an ideal of between 45.6-47.7
View attachment 358926
Let's say your ESR is 42.77422 here, it would get 5/10

But there's still an issue even with this Super-Gaussian curve; there were negative points in the tier system, so how do we replicate this with the new system?
ANOTHER FUCK ASS FORMULA n*****s
View attachment 358930
g = Max Points (on the specific ratio in your formula (check example for better understanding))
v = Value on Curve (earlier, in the ESR example, a 42.77422 gets 5/10, the 5 would be used here)
a = Plateau Height (previously mentioned)

Here's an example of a formula where ESR's maximum value is +20.
View attachment 358942
If you got 5/10 on the Super-Gaussian after inputting your measurement (x) and your Plateau Height (a) as 10, you'd get 0 as your score to go towards your formula.
If you got 9/10 with the same conditions mentioned above, you'd get 16 as your score to go towards your formula. 3/10 would get -8 and 0/10 would get -20.

This shit was some of the most effort I've ever put into a thread. Just for 90% of you n*****s to not understand. Thank you for reading.
Click to expand...
I kind understand it you basically fixed the aspect that overated people by not deducting harmony like tier system
 
Ldar_till_htn said:
I kind understand it you basically fixed the aspect that overated people by not deducting harmony like tier system
Click to expand...
I didnt fix it cuz i didnt use it first
Surgerymax and faceiq use it
 
genio said:
I didnt fix it cuz i didnt use it first
Surgerymax and faceiq use it
Click to expand...
But every analyzer gatekeeps so i had to figure oht a fucking formula
 
genio said:
But every analyzer gatekeeps so i had to figure oht a fucking formula
Click to expand...
Yeah but this is the idea for an optimal harmony formula ?
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 8 (Looksmaxxers: 0, Guests: 8)
Back
Top