when looking over and over all address from 1 to 65 i couldn't see anything that would help, obviously no pattern as we know,
So I simply asked my brother to look at me and yes! he managed to see the obvious, see how simple it is...
if you generate a private key with a reasonable size in the most random way you know, you will see that it will almost never, or even never, have standard characteristics as the friend mentioned on the previous page, such as:
0000, 1234, 4321, ffff.... right?
in the same way, it will be very difficult to have a few or many 1 bits, let's analyze the keys from 1 to 65:
Code:
private key bits zeros bits ones
1 0 1
3 0 2
7 0 3
8 3 1
15 3 2
31 3 3
4c 4 3
e0 5 3
1d3 3 6
202 8 2
483 7 4
a7b 4 8
1460 9 4
2930 9 5
68f3 6 9
c936 8 8
1764f 6 11
3080d 12 6
5749f 7 12
d2c55 10 10
1ba534 10 11
2de40f 10 12
556e52 11 12
dc2a04 15 9
1fa5ee5 8 17
340326e 15 11
6ac3875 13 14
d916ce8 14 14
17e2551e 13 16
3d94cd64 14 16
7d4fe747 10 21
b862a62e 17 15
1a96ca8d8 17 16
34a65911d 18 16
4aed21170 20 15
9de820a7c 19 17
1757756a93 15 22
22382facd0 21 17
4b5f8303e9 19 20
e9ae4933d6 18 22
153869acc5b 20 21
2a221c58d8f 23 19
6bd3b27c591 19 24
e02b35a358f 22 22
122fca143c05 26 19
2ec18388d544 27 19
6cd610b53cba 23 24
ade6d7ce3b9b 17 31
174176b015f4d 24 25
22bd43c2e9354 26 24
75070a1a009d4 32 19
efae164cb9e3c 22 30
180788e47e326c 29 24
236fb6d5ad1f43 22 32
6abe1f9b67e114 24 31
9d18b63ac4ffdf 22 34
1eb25c90795d61c 28 29
2c675b852189a21 33 25
7496cbb87cab44f 26 33
fc07a1825367bbe 28 32
13c96a3742f64906 32 29
363d541eb611abee 28 34
7cce5efdaccf6808 27 36
f7051f27b09112d4 34 30
1a838b13505b26867 36 29
it is very clear that the larger the interval the less chance that a key has too many or too few bits (if they are randomly generated, of course)
some keys even have the same number of bit 1 and bit 0 having an exact 50% rate.
let's say we're going to try 65, first we'd try ~50% right?
as it is odd let's assume it is 33 bits 0 and 32 bits 1, in the first interval would not be found, since we know the numbers of bit 1 and bit 0.
with a multi process we can search for the following ranges:
0 = 33 and 1 = 32
0 = 31 and 1 = 34
0 = 30 and 1 = 35
0 = 29 and 1 = 36
0 = 28 and 1 = 37
and their inverses
0 = 32 and 1 = 33
0 = 34 and 1 = 31
0 = 35 and 1 = 30
0 = 36 and 1 = 29
0 = 37 and 1 = 28
eliminating all other intervals that are very unlikely to be our much sought after key
66 is not too far from 65 with possible 1 and 0 bit intervals like:
0 = 33 and 1 = 33
0 = 32 and 1 = 34
0 = 31 and 1 = 35
0 = 30 and 1 = 36
0 = 29 and 1 = 37
and their inverses
0 = 33 and 1 = 33
0 = 34 and 1 = 32
0 = 35 and 1 = 31
0 = 36 and 1 = 30
0 = 37 and 1 = 29
It makes sense?
this is just a small start of what I got in about 2 months, I identified extra patterns, improved and adapted operations on the elliptic curve secp256k1 to speed up the search, but as I said I'm stuck at 250,000 key/s, I believe memory issues
help me to continue
1JamesJ2H2myei94NswaBATqEsBhATENSU
, I'm improving to have at least 1Mk/s, a dream would be 300Mk/s 
