You fibber you. The compressed address leads to the same wallet and it's still a red zero, you just changed the source of the page.  
My proof.  
http://directory.io/61426087740424605227689896346135129414548009642951912170648466395300309182#5Hwwc74Mz8gKoRiQMJkRDDL2izyHpWpDbbuB6zhrq3yUZh2Q7iN  

The point is not to use the same thing as RSA, but that there is a pattern that once was not known and when it was discovered it was a big break through. It's math, there's a pattern, I guarantee it, especially since there is a finite number of combinations, making it a rational namespace not an irrational one like the number pi.

Hey, reviving old topic...I know. The value in this is that someone like me that's creating a table of key pairs that are sorted by public address, with enough, can derive patterns of offset that the private keys are, assuming they wrap around. 


On the subject of RSA's:
Source: https://blog.cloudflare.com/a-relatively-easy-to-understand-primer-on-elliptic-curve-cryptography/ 

Eventually there will be a breakthrough when it comes to priv, pub keypairs with bitcoin. This idea can be used as a bitcoin recovery business for what was once forever lost bitcoins. This will also mean bitcoin will be broken unless it adapts. I for one would love to create this recovery service for the bitcoin community. If I get close to the breakthrough, I'll be sure to quitly tell whichever devs are working on the main bitcoin development. What good would it be to crash the price of bitcoin to nothing and hold all the coins, but there's value in being the first to make it to recover any current lost coins to re-enter into circulation (a recovery miner), and any new wallets of course would be migrated into newer wallet formats as long as the owner is still alive, thus revealing lost coins.

What do you guys think about this.

Using Python 
I'm gathering a list of private and public addresses into two separate files. 
I keep the private and public addresses mapped together throughout the whole process btw. 
Sort the list of public addresses. 
Convert the addresses into an integer, rebase into regular base 10.
Compare how close the public addresses are numerically. 

The goal is to get enough public addresses that are close enough together , or similar distances apart, to see if there is a pattern in numerical difference in private keys.  I.e. - every priv addrs resorted by pub_addrs is some integer/pattern apart
Finding an approximate step distance, assuming that the it wraps around. 
For instance. 

step 3 starting at 0
0 1 2 3 4 5 6 7 8 9 A B C D E F 

0 3 6 9 C F 2 5 8 B E 1 4 7 A D 0

Obviously pattern could come as more in depth like first step 3 the step 5, repeat. But the only way figure that out is to get enough public address neighbors in different regions of the address space. 

Then when I think I have a pattern that may work, create a test address pair, proceed as if I don't know the private address. 
Determine my test ranges of private addresses based on the public address. 
Decode my base58 guess range addresses into long integers. 
Integer // 128 + 1 for page number of these "Directory.io" like sites and scrub in either direction looking for the public address to match. 




The problem really comes down to that I periodically get a shorter number, unless, that's a key to the pattern.  Check it out. 

Ha. Already found three that were emptied at some point. Does that mean I broke the odds of a lifetime burnt it on empty addresses, or should I go buy a lotto ticket.

Great novelty game. LOL 
How can I set it up to get the information from another location. I have a setup that allows me to get the information from, which is accessible via the browser and I would like to point the 'http://blockchain.info' CORS request there (while keeping the address in link) for retrieval instead of blockchain.info  ( http://replace_this_part_to_my_local/q/address/#keep_this_the_same)