| WARNING: This tool is created for entertainment and fun purposes only. It should not be used to crack a real life locks! |
|---|
Combination locks usually have C = 10^N (where N is a number of rotating discs) combinations. In case of average 3 disks lock one would have to try up to 1000 combinations to open it.
But what if we take human factor into account and assume when owner scrambles the lock, the wrong (or closed) combinations are usually within ± range.
With this assumption we can potentially learn a code by M numbers of the wrong combinations where M is much smaller than C.
Example:
Suppose we have a lock with the passcode 309.
We open/close it multiple times and write down resulting combinations:
Now let's feed it to the tool:
$ go run github.com/leoleovich/glock 527 117 581 117
Number 1 is probably 3 (4/4). If not try: 6 (2/4) or 4 (2/4)
Number 2 is probably 0 (4/4). If not try: 9 (3/4) or 2 (3/4)
Number 3 is probably 9 (4/4). If not try: 8 (3/4) or 6 (3/4)
Glock operates on intersections of sets which means the more diverse numbers we get the faster and more reliable the result is.