I think you're mixing up the concept of entropy. The entropy is the measure of randomness in the data and with more entropy, the harder cryptographic schemes are to break. Going back to your comment, the asserted 130 bits of entropy in the key would be harder to break than 65 bits.

I'm also unclear on where you got the 'multiple of 5' bit about. It seems the keys corresponding to numbers divisible by 5 were used in a spend transaction by the puzzle creator. Using those addresses in spend transactions reveals the public key and saves compute that would be wasted hashing. It also enables direct attacks using Pollard's rho (which someone already posted a link for above).

Src: https://bitcointalk.org/index.php?topic=1306983.msg51466379#... https://en.wikipedia.org/wiki/Pollard%27s_rho_algorithm_for_...

Another interesting discussion on bitcointalk about using Pollard's kangaroo: https://bitcointalk.org/index.php?topic=5244940.0