Btk

[Roy Corduroy]

Good morming Jeff, thank you for your input on the Podcast thread about profiling and I have taken the liberty of extracting your discussion and map of Wichita and Park City Kansas you prepared and your discussion of Dennis Rader, the BTK serial killer sentenced to life in prison for his henious crimes.

Now, despite my pointing out that specific cases are not a good basis for evaluating whether or not geographcal profiles are useful, and that's not my intent here, the podcast did mention Dennis Radar, and in a way that suggested his case was unlikely to result in a profile that would be helpful because of his tendancy to search and stalk. Below is the output from the routines I've been working on. The red squares are the offense locations, the bright blue square is Dennis Radar's home location. His home falls in "zone one" (the yellow and magenta areas combined). In his case we see zone 1 (which reduces the search space to 2.5% of the total) is split between the north (Park City) and south areas. His wife worked at the VA Hospital (zone 4), and Radar used to drive her to work when he was unemployed (just before the start of his crimes) and he later was employed by ADT (I think just slightly to the west of the southern zone 1 in zone 7 or 8, but I've never been able to confirm the address). Anything in zone 20 or less is better than what a random (chance) search would produce on average (chance of course could range from zone 1 to zone 40, with zone 1 and 40 only happening 2.5% of the time each; basically, randomly placing zones should end up with 2.5% of offenders in each of the zones) In other words, the output is picking up on a couple of anchor points.

- Jeff

I have a question please, Jeff. You say the bright blue square is Dennis Rader's home. I don't see a bright blue square. Can you point it out to me at your convenience.

Roy

[DJA]

It is attempting to hide behind the second from top red square in Park City.

[JeffHamm]

It is attempting to hide behind the second from top red square in Park City.

Yes, that's it. Sorry, he's proximity to a crime location masks the home marker somewhat.

- Jeff

[Trapperologist]

His last victim was a neighbor. A decade later he left out one of many communiques in his neighborhood.

Without those, you wouldn't have a locus anywhere near his home, if you use the geoprofiling formula, as it stands. And look at how the grouping in the center looks remarkably like the Ripper map!

[JeffHamm]

His last victim was a neighbor. A decade later he left out one of many communiques in his neighborhood.

Without those, you wouldn't have a locus anywhere near his home, if you use the geoprofiling formula, as it stands. And look at how the grouping in the center looks remarkably like the Ripper map!

Oh, that is interesting, but you're right those locations look like the JtR series, but "upside down". Out of curiosity I've limited the analysis to those cases alone, and the result is, not surprising, a pattern very similar to the JtR outputs I've posted elsewhere.

And while the crime range gets greatly reduced, and nowhere near his residence, Radar did have anchor points in that area. When he was out of work (which he was at the time of the Otero murders), he drove his wife to work at the VA hospital, which is in that area. Prior to that he worked at Cessna,a bit further south. He also had ties to the University (a bit north). I've marked those three known locations with yellow squares. At some point he started working for ADT securities (but I don't know where that was located at the time and have never been able to track it down). There are a couple of potential locations that are in that part of town, which I've marked with darkish blue squares. The one in the middle (in the green area, zones 21-40) ends up as a peak area of interest once the other offenses are included, and in fact, all of them end up in zones 1-20, as do the University and his former place of work. Now, to some extent, that's not too surprising because as the crime range expands the profile too will both change shape, and of course, it also projects over a larger area. With the reduced set, they fall outside the main area of interest, but still close enough that they fall in regions not designated as "nothing to see here". As for the ADT locations, I have my suspicions, though, that it is one of the two to the west that he might have worked at, though the "green located one" is another possibility (based upon the fact that he made a phone call while out for coffee with work mates I believe, and that was made not far from those two ADT locations, but not all that far from the centre green one either, so they could have gone to a coffee shop from work starting from either of those. I really wish I knew the address where he worked, so if you ever come across it, please do let me know!)

But what is of interest is that the VA Hospital, which is the one anchor point in that area that was "in existence" from the start of the series, is right in zone 1 (and very close to the peak, the pink spots, that indicate a 0.5% search required). Now, how useful would that be? On it's own, not very, even if we allow for the idea that we concluded it was the VA Hospital that was importnat - what will that indicate? Someone with ties to the VA Hospital might be of interest. That's a lot of people! These don't solve cases. What they can do, though, is help with suggesting a lead that comes up through actual investigation is worth looking at a bit more. If, for example, Rader had turned up somehow in the investigation as some random person being questioned, if it was noted his wife did work there, it might suggest he was worth a bit more looking into, just not to the exclusion of all others. The larger, more complete profile, shows he had a number of anchor points all in that region.

anyway, below is the output from that reduced set of offenses, with the above locations marked as I've indicated.

- Jeff

[Trapperologist]

Thank you for that, Jeff. It does look like the Ripper map upside down.

You've also given me a better appreciation of geographic profiling. I was disappointed when I read Gordon Kerr's illustrated book on geoprofiling and only one crime map had the known offender's residence mapped, and that was the unorganized Vampire Killer who killed someone outside his home (BTK was not included), much less a comparison to the schematic, choropleth map.

But now I see that I was being too demanding like everyone else who expect this mathematical formula to give us an exact address, and not some bar where he goes to with thousands of other men. The scope of that book dealing with 30 different cases I guess could not hope to search out and incorporate all the anchor points of all 30+ killers, and see how the formula does. It was a deficiency in a compendium book that doesn't necessarily reflect on the scientific value of geoprofiling.

[JeffHamm]

Thank you for that, Jeff. It does look like the Ripper map upside down.

You've also given me a better appreciation of geographic profiling. I was disappointed when I read Gordon Kerr's illustrated book on geoprofiling and only one crime map had the known offender's residence mapped, and that was the unorganized Vampire Killer who killed someone outside his home (BTK was not included), much less a comparison to the schematic, choropleth map.

But now I see that I was being too demanding like everyone else who expect this mathematical formula to give us an exact address, and not some bar where he goes to with thousands of other men. The scope of that book dealing with 30 different cases I guess could not hope to search out and incorporate all the anchor points of all 30+ killers, and see how the formula does. It was a deficiency in a compendium book that doesn't necessarily reflect on the scientific value of geoprofiling.

Yah, unfortunately these kind of tools don't sound as cool when they are simply reported as reflecting a spatial pattern analysis to provide a statistical probability distribution. And because it is such a complicated behaviour that one is trying to model, with so many factors that may (or may not) be important, there are a lot of different possible ways one could approach the problem. The routines I'm working on (which admittedly is a sideline area for me, but one which has captured my interest and curiosity due to the challenges it poses) are fundamentally different from that used by Rossmo. However, both approaches result in very similar performance on cases as a whole. While Dr. Watson might outperform Rigel on some cases, Rigel will outperform Dr. Watson on others. I've tried to see if it is possible to combine the two, but so far that hasn't shown any improvement. It might be possible to detect when a series is better suited to one analysis than the other, but I've not been exploring that just yet.

It's unfortunate that it is the most impressive outputs that tend to get shown (i.e. the BTK output from Dr. W. is impressive in my view; I'm always impressed whenever a particular case ends up in Zone 1). But really, those are just the extremes, and like most things in life, the everyday performance is well below the very best possible performance. Analyse enough cases and, I'm sure, sooner or later one will be "right at the very highest point", as if it identified their house. But those are cherry picked examples, and in the end, doing what these do is actually useful enough. Hopefully some improvements can be made, and hopefully I'll be able to continue to increase the sample of cases I can work with.

I don't want people to get the wrong impression, and end up thinking these are absolutely correct, but at the same time, people shouldn't dismiss them entirely either. They are well informed opinions, and they are objective evaluations of the spatial pattern. But like all opinions, well informed or not, they are not "facts that JtR lived here" type things.

- Jeff

[jmenges]

Marine Hedge was his neighbor and 8th victim. Dolores Davis was his last victim. She also lived in Park City though a few miles away.
I have a video of myself driving these Park City locations a few years ago if anyone is interested in their proximity to his house.
Excuse my general unpreparedness and my wife’s occasional finger.

https://vimeo.com/158707206

JM

[Trapperologist]

Thanks, Jonathan. I had a feeling the close neighbor was earlier in the series.

But isn't it significant that his first 7 were in the city while he lived in the suburbs? I think that's the case with all "spectacular" crimes. They don't seem to make a spectacle of themselves in their own area. My thinking is that he ended up diverting from his plans and killed close to home because he lost control or got to old to troll.

[Trapperologist]

Jeff,

do you know how Rossmo got two hot spots for the Ripper crimes? I can understand how you get two for BTK but not JTR.

[JeffHamm]

Thanks, Jonathan. I had a feeling the close neighbor was earlier in the series.

But isn't it significant that his first 7 were in the city while he lived in the suburbs? I think that's the case with all "spectacular" crimes. They don't seem to make a spectacle of themselves in their own area. My thinking is that he ended up diverting from his plans and killed close to home because he lost control or got to old to troll.

His crimes in the city started when he was unemployed (but drove his wife to and from work each day; and I believe it was during that he first spotted Mrs. Otero), and continued while he was working in the city for ADT I believe. So, much of his daily routine and travels would be concentrated in that area, which is reflected in the "area of interest" in the city. He later ended up working in Park City itself, so again, he was more likely to spot potential victims in that area. BTK is, in many ways, an ideal case for geographical profiling because of how he spotted victims, during his day to day routines, so the areas he found them was very much related to his daily spatial activities. Those who go tolling at night time, become spatially aware of areas they might not otherwise go. But again, the analysis will still pick up on areas of interest, which might suggest it would be worth checking that area, maybe routine checks, parking tickets, etc. And if someone shows up in searches like that, particularly if they have no obvious reason for being in that area, that would produce a lead to explore. It might not lead to anything, of course, but again, this sort of information is helpful to suggest where might be good places to look. It doesn't pinpoint people, it provides an analysis of areas in which an investigation might have a greater chance of finding a lead to investigate (and it is through investigation that cases are solved).

- Jeff

[jmenges]

It should be noted that Rader stalked -or as he would say, made “projects” of- many more women than he ended up murdering. Some, rumoredly, up to 150 miles away from Wichita. Their identities and locations have been kept confidential, so without that data it’s difficult to say for certain how much his ‘day to day routine’ influenced who and where he killed. I think with Rader, as with most every serial killer, we’re working with only a portion of the puzzle pieces and aren’t able to truly have a complete picture.

JM

[JeffHamm]

Jeff,

do you know how Rossmo got two hot spots for the Ripper crimes? I can understand how you get two for BTK but not JTR.

It's a function of how the underlying analysis works. Without going into too much detail, the basic approach is that you take a location, and then look to see if there are a bunch of offenses at a particular distance, with fewer as you get closer to the location you're testing (as this reflects the "buffer zone"), then a high concentration of offenses, and then a decreasing density as you go "too far". You can get local "hot spots" to areas with a few offenses even though the majority are further away because those that are further away have less influence. The ones at "the right distance" sort of signal "this looks good", and even a large number of far away ones produce less of a negative impact on that assessment.

Rossmo's routines produce 3 areas of interest. Most times what is shown is just the very peak area, but a full presentation shows areas of interest in 3 locations, with the peak in the Flower and Dean area (yellow area with the pink centre peak), and the two other yellow areas. Those would be zone 1, split over 3 locations. They also produce the two areas, one around Park City, and one in the Main City. it also locates his wife's place of work in zone 1 (in the peak even). In this case, Rader's house is in zone 4, but again, you can't really compare two analyses based upon comparing one case, so while Dr. Watson had Rader's house in Zone 1, etc, both of these analyses (which are based upon very different ways of doing the analyses) end up with quite similar global patterns. Sometimes the patterns do look quite different though, but for the most part the regions of highest interest tend to more or less correspond, it's in the less confident areas that the patterns can tend to diverge, and again, that's probably not overly surprising either. (the top image is the JtR C5 profile generated by Rossmo's routines, the bottom being the BTK profile). It's possible Rossmo has updated things of course, so it may be that the most current version would produce something a bit different.

- Jeff

[Trapperologist]

So we actually had 3 areas of interest with Jack the Ripper. I thought it was only two. Thanks for explaining. I thought he might have run the program twice, once with and once without the GSG or something.

Is the problem (of the possibility of an outsider) with JtR being one or with his being an "organized" killer and therefore possibly an outsider who didn't want to draw the whole city's attention to the place he was living in, versus a disorganized killer who basically couldn't go a half mile in any direction, or something else? BTK has to be considered a suburbanite who lived miles away from the majority of his crimes and, as Jonathan said, he was ready to go 150 miles away, although he didn't.

[JeffHamm]

So we actually had 3 areas of interest with Jack the Ripper. I thought it was only two. Thanks for explaining. I thought he might have run the program twice, once with and once without the GSG or something.

Is the problem (of the possibility of an outsider) with JtR being one or with his being an "organized" killer and therefore possibly an outsider who didn't want to draw the whole city's attention to the place he was living in, versus a disorganized killer who basically couldn't go a half mile in any direction, or something else? BTK has to be considered a suburbanite who lived miles away from the majority of his crimes and, as Jonathan said, he was ready to go 150 miles away, although he didn't.

Organized vs Disorganized needs to be viewed as two ends of a continuum, and not as "either/or" classification. JtR has some organized aspects (he brings a weapon with him, he appears to pose as a customer, he doesn't leave things at the crime scene, etc), but he has disorganized aspects too (he attacks on the spot, he doesn't seem to care how risky the area is, he engages in a lot of time consuming extraneous behaviours). Few people are at the extremes of any dimension, and JtR is not a "pure" example of either organized or disorganized. He's probably more towards the disorganized end, but some of that may reflect the time period (behavioural profiling isn't something I'm well versed in).

That dimension, though, doesn't really influence the spatial pattern all that much, as both spend time in certain areas and so those will be where they end up committing crimes. Disorganized offenders can travel great distances, and organized can be concentrated, or vice versa. The areas we are most familiar with tend to be near our "anchor points" in our every day lives, hence, hot spots will tend to indicate areas where the offender has some connection, be it work, home, or whatever. What it does for an investigation is try to suggest areas that are worth putting their limited resources into.

BTK had lots of anchor points around the bulk of his crimes. His house was miles away, but much of his daily life was in that area. It's not about locating the residence per se, but the locations where the offender has some regular activity (i.e. where you'll find them a lot of the time). Generally, though, the residence is often a major anchor point that emerges from these analyses. And many "commuters" have normal anchor points in their crime zone, such as work, or a club, or a pub, or church, or family member that they visit, etc. It's much more rare to find a "true commuter", meaning someone who commits offenses in an area for which they have no other connection in their lives. Still, some offenders will travel a fair distance, and commit crimes spanning very large distances. The profiles, while they might narrow the search space down to a small percentage of that space, will still be suggesting large areas, and by no means will those correspond to someone's backyard!

- Jeff