Statisticians Uncover the Mathematics of a Serial Killer

[Celesta]

This is an abstract concerning a mathematical model being developed for the patterns of a serial killer's murders.

http://science.slashdot.org/story/12/01/17/0420232/statisticians-uncover-the-mathematics-of-a-serial-killer

[Heinrich]

Rubbishing the skill of profiling strikes me as ironic coming from a couple of mathematicians who have based their theory on only one case of serial killing. As statisticians, they ought to know better than rely on such a small data sample.

[Barnaby]

Ebbinghaus worked out most of the basic laws of memory and forgetting using a single subject - himself. Over time, of course, his research was validated and has proven to be extremely general, but that doesn't change the fact that the principles were discovered by self experimentation. There are numerous other examples in which scientific discoveries have been made by an intensive investigation of one or a few subjects.

Behavior, including murder, is something individuals do. Knowing something about what the "average" serial killer will do tells us next to nothing about what the at-large fiend will do next. The average serial killer is a construct that may not exist. No one has 2.5 kids. Studying huge populations isn't always the best way to go when ultimately we are interested in individuals.

If we assume that there are neural similarities (as well as differences) among the brain functioning of some types of serial killers, then this research has some value. Perhaps the individual differences lie in the exponent. That is, it is a power function but it must be weighted differently for different killers. Who knows, it's pure speculation, until they demonstrate some predictive validity, but the research surely has merit.

[RedBundy13]

Nicely put Barnaby. I agree, any research at all has at least some value.

[K-453]

http://arxiv.org/ftp/arxiv/papers/1201/1201.2458.pdf

[Heinrich]

Poorly put, Barnaby. This research attempts to demonstrate a pattern of interludes for all serial killers based on the time frame of one maniac. It is flawed in its design.
As for all research having value, RedBundy, some tobacco companies can prove from their own sponsored research that cigarette smoking has no causal relationship to cancer.

[K-453]

This research attempts to demonstrate a pattern of interludes for all serial killers based on the time frame of one maniac.

Do they?
They are talking of Chikatilo in their paper, everything else is called a hypothesis.

Are there some mathematicians around here? Could they apply these formulas to other serial killers? Would be interesting.

[Barnaby]

http://arxiv.org/ftp/arxiv/papers/1201/1201.2458.pdf

Thanks, I just read the paper and am not as impressed as when I just read the Abstract. Figure 4 is pretty, but all this really shows is that he killed in spurts. And actually, if you look at the raw data, with the exception of a few year period in the mid 80-s where he pretty much stopped, the pattern is fairly linear.

One major limitation of this curve fitting technique is that you need lots of points. Usually data is a good things but lots of points corresponds to lots of murders unfortunately. With respect to Jack the Ripper, you can fit all sorts of things to a data set of five (or 3, or 4, or 6) points, so this won't be very helpful. I'm interested if the data from other prolific serial killers would look similar. If I have time I will see if I can repicate some of this with the Green River Killer or Bundy.

[Barnaby]

Analyis of Ridgway and Bundy

Here are the cumulative murders of Ridgway and Bundy over time. Ridgway slowed to a trickle, and in fact I do not show his last three confessed murders over a 12 year period. Bundy stopped while incarcerated. I do not show his last three murders after his escape from his initial incarceration. To me, there is no evidence of a spurt-and-stop pattern arguably seen with Chikatilo. Ridgway did stop and pick up again, but at a very slow rate. Bundy just kiled like clockwork. If anything, their patterns appear mostly linear during their main killing sprees.

[Barnaby]

Also note that these are data based on confessed murders. The actual count is probably considerably higher, especially for Bundy. The slight leveling off you see with Bundy in the middle of his record corresponds with him changing locale.

[Barnaby]

One last observation: To me these records look virtually identical. Had Bundy not been apprehended, he probably would have continued killing at a high rate for some time. I guess that, like Ridgway, eventually as he aged he would have slowed down considerably. The records differ only because Bundy was arrested in his prime, so to speak.

I am not very familiar with Chikatilo. Do the stops in his record (Figure 1 in the publication) correspond with obvious environmental events (such as a move, a run-in with the law, etc.)? If so, it would argue against the authors' power function explanation based on neural activity. With respect to my analysis, they just seem to kill and kill and kill until they are caught (Bundy)or, for whatever reason, eventually turn it (mostly) off (Ridgway). Once they turn it off, they never really crank it up again. Can anyone think of a fairly prolific serial killer who stopped killing volitionally and, some time later, picked it back up at the same high rate (besides Chikatilo)? I cannot. BTK comes to mind but his pattern seems more like Ridgway.

[The Good Michael]

I am not very familiar with Chikatilo. Do the stops in his record (Figure 1 in the publication) correspond with obvious environmental events (such as a move, a run-in with the law, etc.)? If so, it would argue against the authors' power function explanation based on neural activity. With respect to my analysis, they just seem to kill and kill and kill until they are caught (Bundy)or, for whatever reason, eventually turn it (mostly) off (Ridgway). Once they turn it off, they never really crank it up again. Can anyone think of a fairly prolific serial killer who stopped killing volitionally and, some time later, picked it back up at the same high rate (besides Chikatilo)? I cannot. BTK comes to mind but his pattern seems more like Ridgway.

Chikatilo never stopped killing. He had a few lean years when he only killed 4 or 5, but he never stopped. I would say that being a family man, and employed, he had other obligations at times that prevented him from killing.

Mike

[RedBundy13]

Poorly put, Barnaby. This research attempts to demonstrate a pattern of interludes for all serial killers based on the time frame of one maniac. It is flawed in its design.
As for all research having value, RedBundy, some tobacco companies can prove from their own sponsored research that cigarette smoking has no causal relationship to cancer.

Then it definitely has great value to the tobacco companys

[K-453]

Barnaby, if you're looking for a prolific killer, look in South America: http://en.wikipedia.org/wiki/List_of...ber_of_victims