Hello,
any one here please can explain to me what mean by Wild and Tame ? , I know that they are hops of kangaroo, but how can I use them.
for example I want to know if i run pollard for #105 and get a Wild hop, but I shutdown the pollar code .. I don't want to start from the beginning so how can I import the Wild that I find in first ? ,and the same for Tame.
the second question is, what is tame and wild ? they are collision ? if yes and we can import them ,so we can just everyone poste the wild and tame that he find and we will find it quicly !!
Thanks
Even the inventor John Pollard regrets this name, this tech is some 40+ years old. Originally is was called Pollards's Rho/Lamba method, where Lamda is an upside down Y
The notion was that two different searches converge on a common point, leading to the end goal
Then historically Pollard was reading 'national geographic' on a story about kangaroo breeding in Austrailia, and decided to call lambda-method the Kangaroo method, it is complete bullshit, there is no abstraction or moral equivalence, or equivalent thinking here
Long ago, they used to call this crap baby-step/giant-step which is better, it all come out of the same people
The principal IDEA of Kangaroo is they HOP, they HOP BIG ( giant step ), they hop small ( baby step ), you jump around the finite-field ( your prime space, a big number )
Originally they kept all the hop history in memory, but of course that restricted the search, think 2^256 is 10^77, which is all the electrons in the universe, impossible to find this memory
So they came up with a method that didn't require lots of memory, which is this method the 'kangaroo' or 'lambda' method.
Jump around the field space, and if you find a Disctinct-Point you record it in memory, and you have 100's if not 1,000's of processes doing the same, if later another process hits that special-point, then all the kangaroos follow that point. So a DP is the 'Y' where the two top lines converge to make one.
Now the problem with DP is its not the real Lambda Point its just a special point, defined as you wish, sort of like mining bitcoin, where difficulty is set by the number of leading zeros on a hash, here the DP is the special point defined much the same way, its just a unique point that all processes can agree on; it doesn't mean that this is the true point of convergence, just means that its a "Point of Interest", randomly defined by humans. Under their rules.
So kangaroos hop around looking for a DP in giant-hops, and if a DP is found, then they incrementally search that point forward;
The problem with the Lambda method is its restricted to a tiny subspace of the field, if your doing btc which is 2^256, the jean-luc kangaroo algo only lets you choose a 2^20 field of view, so unless its a 'made up puzzle' problem in that 2^40 space, the likelyhood of solving the problem is zilch
Most success in ECDLP has been in Pollard-Rho, like a '6' character you following the tail until find the point of entry into a circle, but this method requires memory storage, but this method only works on TOY problems because there is not enough memory on earth to store the history.