I have disassembled the password checker for bloodlines. The password contains 32 bits of data (4 types of picture choices, which is 2 bits per picture and 16 pictures). Instead of using a lookup table like I thought it might be, it is just a complex algorithm. There are 14 bits of data (2 bits for difficulty, 3 levels of difficulty and one combination is illegal, 4 bits for rest but it caps any values over nine to be nine, 1 bit for character john or eric, 4 bits for stage which I will detail further and 3 bits for continues allowing between 0 and 7 continues). There are 8 bits of checksum (two different checksums) and 10 bits unused. That means every valid password has up to 1024 variations to get the same result. There are 4320 valid password scenarios and 4826060 valid passwords. That means a randomly generated password has around 1/1000 of a chance of being valid.
The stages selectable using the password are 1-1, 2-1, 3-1, 4-1, 5-1, 6-1, 6-6, 6-8, and 6-11.
For further details, you can either look at the actual coding algorithm by looking at my commented disassembly of the routine or my perl code that decodes and encodes passwords.
I wrote a script to find the shortest password (most blanks in all the variations) for each character, difficulty and stage with 9 rest and 7 continues. I converted these into easy to read images like the one above, just look at the file names to know what the password does.
Password images, 68k code, and perl script are located at:
https://www.dropbox.com/sh/791lx5seaklk ... rIKumsOa#/
