I am seeing us going through the same questions and patterns. For example, serial killer/not serial killer/pretending to be serial killer. From Delphi/outside, but close to Delphi/spent childhood in Delphi/had a 2-hour trip to MHB because it is really enough to know the area/accidental truck driver. Etc, etc.
I don't know if there is a way to write all these questions/patterns down and feed them to a computer, that would make several models. Different ones. And see where it overlaps with other situations.
Inexperienced as they were, I see that ISP has collected a lot of material. And it was probably fed into a computer. What came out, obviously, was not enough, but I wonder if different scenarios could match some other cases.
Because this is what they do with DNA. "Bad" or "good". I have read about DNA familial searches, it is human work but it uses certain algorithms. Human logic is still more interesting than AI, but computer can search databases of different crime versions.
I once spoke with a guy who was writing computer algorithms for search of SK, for a police department in another country. He said it was, in a way, straightforward, because the mentality of SKs was very rigid. But this situation is somewhat different, even if he is a SK (or a budding one). But really, they need good programmers, it seems, at this stage. MOO
P.S. The search should be not only through found victims, but the disappeared ones, too. For the pre-disappearance similarities. It is possible that this killer does not bury, but hides better when he has the time.