DES and markov

[Mem5]

Hi,

I have 20 DES hashes.
I already recovered 3 of them. I noticed these 3 passwords were created with the same pattern :
8 chars, upper + digits, seem to be random.

I first think about brute-forcing all the charset ?u?d len=8, but this will take 4 days.
Then, I thought about using hcstatgen and statsprocessor with ocl-plus and markov chains to reduce the charset.

But the passwords seem to be randomly created, so is it a good idea ? My dictionnary for this crack is only composed of 3 lines (the 3 recovered passwords).
Do you think the statistics will be good enough ?

If no, any other idea ?

Thanks.

[epixoip]

it would help you crack the three you already found faster but that's about it.

the only way markov would help with this is if you had a very large sample of passwords that were all generated from the same source, and that source had very poor entropy.

[ati6990]

maybee the bruteforce will reduce some hours if you do a minleng=8 so you dont need to walk the 1-7 ?u?l shit, maybee....

[eljolot]

Use markov, in my experiencie with big list I use at first I use -t 10, -t 20 , -t 30 and -t 40 , more markov more sucess, you can try with the all list using one of this markov or better use oclhashcatlite bruteforcing hash by hash, you have an advantage that you know the number of chars, the use markov -t 30.

[epixoip]

eljolot seems to be lacking a fundamental understanding of the markov attack.

[Mem5]

Ok thank you. So markovv won't help me here.
I have already used the --pw-min/max=8.
So i will bruteforce it using sessions to restore when I want