Quick update from my side 

I have been work with statistics, and beside some hints, I manage to find some  odd to say at least

I was frustrated to the amount of convoluted in the scripts I needed to do to calculate P and Q and the sheer amount of bad explanations of the calculation of bitcoin without step by step (reminds me Butkov), so I worked backwards.

I did a script to run the statistics of each letter of the address and the probability of some letters to appear inspired in Kryptos.
(My personal theory is: if you add cryptography in top of cryptography several times, you eventually will arrive in a shortcut bypassing all the cryptography before)

So theoretically we have base58 what means we can have 58 letter and number to choose, means the probability of each letter is 1/58, so 1.7241379%

Even if I manage to find this, something quite stood out as a sore thumb:
Some letters  has more statistics appearance than others, in some cases up to 8% comparing to the average, and 13% compare with the other extreme of the tail.
It was significant more than I was looking for tbh

The really caveat is:
The top characters are always the same, like no randomization at all.
Its always the same group of letters.

And if you take a look deeper you will see in vastly majority of addresses are not 34 characters, but way less

a) you take the first number because is not relevant
b) you have a considerably amount of repetition inside of the same address. Like the vast majority has minimum of 1 pair of duplicates
c) the puzzle 66 have in reality only 20 characters if you take off the repetition.
d) this effect happening with most of the addresses

You will may say: Is just run more and the averages will pull to 1.72%, right?
Quite not, even increasing the size of the batches from 10k addresses to 100k the difference was negligible.
Of course I can go up, but my point is why this is so prominent?

Sorry but is color code for you see the relationship (https://imgur.com/a/9hpLpdi)

https://imgur.com/a/9hpLpdi

I tried to run different rules, like hexadecimals with a lot of number 3 or 4, and looks odd for some batches... kind of a fingerprint,maybe

Some letters looks significant more relevant than others...
Could be a dead end, i know, but someone looked this before?
Someone with more experience in riped160 can tell me why this effect happened?

Why we don't have literally any other letter above 1.72% beside as follow

3	C	7	N	8	A	B	D	H	6	E	M
P	J	K	2	F	4	G	5	9	L	Q

I have more other specific analyzes for some areas but I rather save to myself tho
I need go back to work, this already took 3 days non stop calculations.