Author

Topic: Reducing variables for a wallet attack... (I am a newb, go easy) (Read 390 times)

member
Activity: 84
Merit: 10
Hello all,

OK, so I have a problem. My wife has 199 bitcoins in her wallet, but the machine, wallet and private key are LOOOOOOONG gone. She did have the presence of mind to keep the wallet key. She generated the wallet using a brain-wallet type system where she picked a 12-word password which was then converted to a wallet key.

My wife has a VERY large vocabulary Smiley And I think she went out of her way to make this difficult to crack too.

Words remembered (and I quote): "Urm....I think one of them was 2,4 dinitrophenylhydrazone...dunno the others. I picked weird ones. Actually that could have been 2."

So I am trying to crack her wallet as a kind of private project.

So far I have been running dictionary attacks using a quite large english dictionary as well as chemical compound lists, latin legal terms etc. But the issue is that to cover all possible 12 word sentences (and the missus is not even sure that she used 12 words) is.....a big job Wink

I think we both might get quite old before we get the phrase.

On the other hand, I was thinking the following:

1) A version of vanitygen which runs on simplest settings (i.e. 1*) where it discards all keys generated which don't fit the profile and outputs to another function
2) The second function simply compares with the wife's address

The thing is, I am not sure I am actually reducing the problem here. Obviously, knowing the address, I can reduce the first 2 or 3 or 4 characters so that it's only the rest we need to weed out, but with each character the difficulty to find a match goes up, and my list of "interestings" grows more slowly.

Can anyone advise, given the FORMAT (i.e. 1-number-CAP-LOWER-number-CAP-CAP) etc a faster way to break this down?

Thanks, and sorry for absolute newbness - I am neither a cryptographer nor a mathematician.

Rit./
Jump to: