The So-Called Serial Killer Math Formula

on January 24, 2012 in Psychology, Serial Killers by

Two electrical engineers from UCLA, Mikhail Simkin and Vwani Roychowdhury, believe that they have shown that the erratic behavior of serial killers actually conforms to a mathematical power law known as the Devil’s Staircase.

The two looked at the crimes of Soviet serial murderer Andrei Chikatilo, they found that when the number of days between murders was plotted against the number of times that number of days was waited, a nearly straight line was formed by the relationship. This line closely matched one formed by neurons firing in the brain.

To be fair, their mathematical analysis is way over my head. I’m a marketing guy by training.

But, I have some issues with what we know about their findings so far.

Firstly, I question whether they’ve done enough to warrant publishing a paper. Really, they need to look at multiple killers, and see if the pattern holds up (I’d be willing to bet it doesn’t).

Secondly, their dataset is only as clean as the info on Chikatilo’s murders. It wouldn’t take many undisclosed murders to throw things off substantially. I suppose this could be addressed by error rate, but it’s potentially a high rate.

Lastly, this doesn’t become truly interesting until it can be used predictively. It’s one thing to look back at a dataset and find a pattern. We humans are very good at that.

If you want to impress me, take a series of linked, but unsolved murders, and tell me when the next one might happen.

Actually, if I understand their findings correctly, it would be more of an issue of saying that another murder wouldn’t happen prior to a certain amount of time passing, but the point still stands.

BTW, to the best of my knowledge, the actual paper hasn’t been published yet, so everything we know is based upon news coverage of the pre-release info. That leaves lots of potential for misinformation.

Hat tip to podcast listener Chris C. for sending in an email that inspired me to look into this more.

There are no comments yet, add one below.

Leave a Comment

Notify me of followup comments via e-mail. You can also subscribe without commenting.