08-02-2015, 06:51 PM
Have you looked into PACK? If not, you will find rulegen.py to be of interest. If you benchmark your own PACK-generated rulesets (Ordered by popularity, of course) and swap out rules that don't work in your real-world testing with rules that performed well in your already-released rule files, you might get some slightly better crack percentages.
To give you a frame of reference, my testing of rule1.rule with a specific hash list/dictionary combination yielded 65.71% cracked. My personal best with a PACK-generated ruleset of a similar size under the same conditions is 64.37%. Frankensteining the two together by incorporating the best rules from both sets could probably lead to some gains worth talking about.
To give you a frame of reference, my testing of rule1.rule with a specific hash list/dictionary combination yielded 65.71% cracked. My personal best with a PACK-generated ruleset of a similar size under the same conditions is 64.37%. Frankensteining the two together by incorporating the best rules from both sets could probably lead to some gains worth talking about.