hashcat Forum
statsprocessor v0.04 with per-position markov chains - Printable Version

+- hashcat Forum (https://hashcat.net/forum)
+-- Forum: Support (https://hashcat.net/forum/forum-3.html)
+--- Forum: hashcat-utils, maskprocessor, statsprocessor, md5stress, wikistrip (https://hashcat.net/forum/forum-28.html)
+--- Thread: statsprocessor v0.04 with per-position markov chains (/thread-1285.html)

statsprocessor v0.04 with per-position markov chains - atom - 06-13-2012

Hey Guys,

today I had the following idea: What if we can link the markov-chain to the password-position, in other words, create a per-position markov-table?

I'm not a guy who thinks to long, so I modified the 0.03 statsprocessor and tried it. Then I did a little experiment you find out if its more efficient or not.

You will see the new per-position enhanced markov-attack is around 29% more efficient than default markov-attack.

I created a hcstat file with both versions. The v0.3 which is default markov-attack and v0.4 which uses the per-position markov-attack.

Code:
root@sf:~/hashcat-utils-0.6# ./hcstatgen.bin rockyou.hcstat < /root/dict/untouched/rockyou.txt
Reading input...
Sorting tables...
Writing stats...

root@sf:~/hashcat-utils-0.7# ./hcstatgen.bin rockyou.hcstat < /root/dict/untouched/rockyou.txt
Reading input...
Generating root...
Sorting tables...
Writing stats...

I took the rockyou.txt which is plain-password, and hashed them as MD5 so that I can crack them later.

Code:
root@sf:~# perl dict2hash.pl < /root/dict/untouched/rockyou.txt > rockyou.md5

For both versions I created a simple dictionary. These dictionaries will be used to crack the hashed rockyou list later. When generating, I took a very low threshold value to force only the potential best password candidates are generated, since this is what I want to compare.

Code:
root@sf:~/statsprocessor-0.03# ./sp64.bin rockyou.hcstat ?l?l?l?l?l?l -t 16 -o /root/dict3.txt
root@sf:~/statsprocessor-0.03# wc -l /root/dict3.txt
17895696 /root/dict3.txt

root@sf:~/statsprocessor-0.04# ./sp64.bin rockyou.hcstat ?l?l?l?l?l?l -t 16 -o /root/dict4.txt
root@sf:~/statsprocessor-0.04# wc -l /root/dict4.txt  
17895696 /root/dict4.txt

The dictionary which cracks more passwords is more efficient.

Code:
root@sf:~/hashcat-0.40# ./hashcat-cli64.bin /root/rockyou.md5 /root/dict3.txt -o /dev/null
.
.
.
Recovered.: 171780/14344391 hashes, 0/1 salts

root@sf:~/hashcat-0.40# ./hashcat-cli64.bin /root/rockyou.md5 /root/dict4.txt -o /dev/null
.
.
.
Recovered.: 221877/14344391 hashes, 0/1 salts

You can download the updates here:

http://hashcat.net/files/statsprocessor-0.04.7z

http://hashcat.net/files/hashcat-utils-0.7-32.7z
http://hashcat.net/files/hashcat-utils-0.7-64.7z

RE: statsprocessor v0.04 with per-position markov chains - Pixel - 06-13-2012

Good job Atom, this is very interesting. Keep up the good work

RE: statsprocessor v0.04 with per-position markov chains - fizikalac - 06-13-2012

Cool! You can call this the atom-chains attack.

RE: statsprocessor v0.04 with per-position markov chains - atom - 09-07-2012

Done with oclHashcat-plus v0.09: https://hashcat.net/forum/thread-1541.html

Thread closed