I wondered about a related problem: how many of the codes are "easy"? Easy as in they are composed of "simple" patterns, such as going to adjacent digits in a n

I wondered about a related problem: how many of the codes are "easy"? Easy as in they are composed of "simple" patterns, such as going to adjacent digits in a number pad. Often it seems that if you are given such a sequence, there's a pattern you can use to recall it. Seems like it would make the randomness suspect, perhaps? But maybe all the sequences have easy rules, the rules just differ? So during a short hackfest I created this to check it out: https://github.com/eras/reco https://github.com/eras/reco . Sorry, no binaries and the font size is hardcoded for presentation, and actually the whole UI stuff is just for that reason there.. By default it scans the whole 6-digit sequence space, but you can also give it a sequence and it will show the rules it finds. Given the rules it uses, it turns out 50% of 6-digit sequences are "easy". Because it is based on the rules I just thought would apply there are probably other "easy" rules that could cover a lot of the remaining 50%.. It also cheats in a way by trying to apply the codes to all* shapes and sizes of numpads (1x10, 2x5, 3x3+1): match in any numpad is accepted for a sequence to be "easy". It may also be some of the rules for the sequences it finds are not "easy" after all :).