> D <-> D2
> x=d[D] <-> x=d[D2]
> xx+e=2na
In order to estimate an error rate, the derivative must be understood. How do I understand it?
I'm unable to descent into madness.
I get that we need to use an optimization algorithm to calculate the squares.
From what I understand there are multiple ways of estimating it. We could create a tree and try and figure out what to add / remove from the squares.
We could estimate the squares using D and D2 and the continue moving from column to column or we can stay put in (e, 1) and (f, 1) and estimate using p of t.
Unless the examples recently shared using only squares is the same as the {n:e0 (from -f): .. }. Either way I'm trying to build some intuition about what a rate of change means in the context of squares.
Wait, are we supposed to use (1, 1) or (0, 1) (or both??) when looking for the derivatives?
The derivative x = dx. D of x? As in, when looking for the derivative of x, we look for the two d[t]'s x is in between? But is that column specific or are we specifcally looking at (1, 1) or (0, 1)?
What the hell, are you guys sleeping? You never miss out on some good fishing.
Oh boy, my heart just jumped for a second before it calmed down.
Working with one of the examples and I see
69^2 - 18^2 (d+n movement in c1 to c2) = 291.
Also 2537 - (47*47 + 37) (d from c1) = 291. I thought for a second that our estimate would be moving something like that. I'm still not 100% sure if it is or not.
Ugh correction 50^2 - 47^2 (x+n from c1 and c2), not d+n.
BUSY!? This shit is important! You're never to busy to share some sweet derivatives!
How do you go from:
c0 = 69^2 - 50^2
c1 = 18^2 - 47^2
That's some massive pull on the string to get (d+n) that far down. I could see 50 to 47, I mean we simply compute the new c' and then calculate the difference with our existing c, but (d+n)!? That is a whole 'nother level of magic.
I swear, if this was our fractal I would be so happy. Either way I don't see how I would use this here and now to figure out the error's for my estimates.
I suppose it would make sense, though. Binary search complexity and all that. You always have two options, just need to figure out how use the correct one.
A line between -f and e? the distance being 2d + 1
At (e, 1, (2d+1)/2) we have the next step in the fractal. Halfway between the sub-branch. Then we rinse and repeat?
One of the nerds. Based on your images I have an idea of who you are though. One of the artists?
Been missing your dank images on the board. Sadly I'm just trying to wrap my head around the latest development.
Figured I'd try some fishing, although "Maybe Chris Maybe Not Chris" has been quite active I still am not entirely sure how the hell the grid works.
I truly believe that. Just need to get an understanding of the grid, to get the last eureka moment. I know it's there and I believe it might be staring me in the face as I look upon the grid.
w H at could a base fractal look like?
H-tree sounds promising. T-tree too. Although in the back of my head I can't shake Mandelbrot. It's a beautiful fractal and shares too many similarities.
Escape time algorithm? Related to the grid? I've tried a bit, but got lost in the numbers. It's too hard to keep all the big numbers in my head to search efficiently.