geometry, watson

[MayBea]

Here's the link for the Victorian Murder Map 1888.

You can see the difference between random murders and a cluster that does not appear to be random.

https://www.google.com/maps/d/viewer?mid=zwbz2aRvX7CU.kiDtOQLr_R80&ie=UTF8&source=embed&oe=UTF8&msa=0

Martha Tabram is in the wrong place. She should be in the middle of the cluster, not outside.

There's a large equilateral triangle formed from the random murders but two of the murders in the triangle are questionable, as I would expect.

[GUT]

Here's the link for the Victorian Murder Map 1888.

You can see the difference between random murders and a cluster that does not appear to be random.

https://www.google.com/maps/d/viewer?mid=zwbz2aRvX7CU.kiDtOQLr_R80&ie=UTF8&source=embed&oe=UTF8&msa=0

Martha Tabram is in the wrong place. She should be in the middle of the cluster, not outside.

There's a large equilateral triangle formed from the random murders but two of the murders in the triangle are questionable, as I would expect.

Did anyone else notice that the points between Annie Smith and Lucy Clark form a near straight line, that must prove something,

[MayBea]

The six murders in a straight line looks strange but its a aberration caused by error and exaggeration.

Annie Smith went into the river and her death, the Lea Mystery, is questionable as a homicide. The second 'murder' is an attempted murder and suicide, and the third, the Canonbury Murder, is really a death by heart attack during a robbery.

The last three ending at Lucy Clark are all along or by the A501, so it's a typical random line of murders you'd expect around thoroughfares.

[GUT]

The six murders in a straight line looks strange but its a aberration caused by error and exaggeration.

Annie Smith went into the river and her death, the Lea Mystery, is questionable as a homicide. The second 'murder' is an attempted murder and suicide, and the third, the Canonbury Murder, is really a death by heart attack during a robbery.

The last three ending at Lucy Clark are all along or by the A501, so it's a typical random line of murders you'd expect around thoroughfares.

As, in my opinion, are the other shapes and groupings people see, the grouping around the C5 is what you would expect to see if someone is killing in a fairly confined area.

[Errata]

If you ever go searching through Google Earth, especially over the States, you can see thousands of examples of unfortunate patterns that just sort of crop up. I was playing this game recently in Boston, and if you zoom out enough over one area, you see several office type buildings laid out looking like a hand giving you the finger. A hospital in some place I visited and now don't remember accidentally built a hospital in the shape of a squatting man with a giant penis as the entrance (that enough people saw that they now are changing it), and my old hometown Nashville apparently has a completely random placement of buildings downtown, but it was actually very carefully planned and executed. Nashville is half clay half granite. The big buildings had to go on the granite. So downtown Nashville is in a stupid place given how the city was forced to adapt around it, but it is really the only place it can be. And my friend is an architect who says that there are more swastika shaped buildings than anyone would reasonably expect, but there was no choice and so they just had to try and hide the shape.

Patterns are usually not significant. And a lack of a pattern may or may not be significant. They occur bloody everywhere. And they show you something, just not perhaps something sensational. I mean, what if his pattern is based on taking care of his sick mother and going home late at night. Or running errands for her like getting soup or brandy? The pattern is not significant but it is explained.

It's like statistics showing that Great White Sharks are immortal. If you don't know what the data means, the conclusion gets muddled.

[somerset]

There's the bunny. Although, what if you could trace apparently useless facts about the areas of the crime scenes, and piece together the habits of someone in the area (at the time)? There's a lot of possibilities and factors in the air, I'll give you that - he could have lived there, or he could have been a businessman or a drifter or some fellow seeing a friend, for all we know. Still, perhaps there's at least one name someone jotted down somewhere that would fit some pattern. I guess the odds are around one in the entire population of the East End, registered or not . . .

Sounds risky, men. Any volunteers?
somerset

[Batman]

Chapman, Kozminski, Tumblety plus a few more that are a good fit for it.

This is a 9-12 km2 area. Its very small, much smaller than many other geographical areas that serial lust murderers have covered I think. Tiny by comparison.

[John G]

Chapman, Kozminski, Tumblety plus a few more that are a good fit for it.

This is a 9-12 km2 area. Its very small, much smaller than many other geographical areas that serial lust murderers have covered I think. Tiny by comparison.

Hi Batman,

What about Frances Thompson?

[Batman]

Why not? I'd put Cutbush also. Behavioural there is something very interesting about all of them.

[John G]

Why not? I'd put Cutbush also. Behavioural there is something very interesting about all of them.

Thanks Batman, I actually think Thompson is a very strong suspect. I've also found some interesting literature on lust murders: see Kocsis, Serial Murder and the Psychology of Violent Crimes, available from www.reasearchgate.net
Also: http://books.google.co.uk/books?hl=e...0model&f=false

[MayBea]

This is a 9-12 km2 area. Its very small, much smaller than many other geographical areas that serial lust murderers have covered I think. Tiny by comparison.

Are you going with suspects that might be religiously orientated and therefore picked the Whitechapel area and locations within it for religious/occultic reasons?

If Stride is a victim, then he had 45 minutes to go to Mitre Square so I don't think that would be a random choice, which changes the whole complexion of the murder pattern.

[Batman]

This is my thinking on Stride/Eddowes timing.

[Hakeswill]

Another possibility: after the aborted Stride attempt/possible interruption, the killer was returning to the safety of a lodging rather than directly looking for another victim. If the route took him via the known soliciting area of the church then KE may have been a random meeting that JTR exploited.

On an aside to the geoprofiling, I wonder what the effect of psychological barriers may be - I'm thinking about various roads or boundaries to districts that may mark a territory in someone's mind. Can geoprofiles can account for such issues?

[Rosemary]

Another possibility: after the aborted Stride attempt/possible interruption, the killer was returning to the safety of a lodging rather than directly looking for another victim. If the route took him via the known soliciting area of the church then KE may have been a random meeting that JTR exploited.

On an aside to the geoprofiling, I wonder what the effect of psychological barriers may be - I'm thinking about various roads or boundaries to districts that may mark a territory in someone's mind. Can geoprofiles can account for such issues?

Could you be thinking of spatial analysis in geography? A geographer once explained the concept to me regarding 3 dismemberment murders occurring in the French Quarter in New Orleans. I was ready to attribute them to a Palo Mayombe cult, but her analysis convinced me that I was just searching for anthropological patterns.

[JeffHamm]

Geographical profile is based upon the idea that crimes are committed in areas where the criminal searches for crime opportunity (hardly surprising). However, we become familiar with different areas, even within our own city of residence through every day activities, and so there are locations (called anchor points) that determine where we become familiar. The most obvious is our place of residence and our place of work. We spend the majority of our time within a relatively short distance of these locations. Canter & Larkin proposed the "circle theory", which is just "take the two most distant crime locations" (so for the C5 that's Eddowes and Nichols) and use this as the diameter of a circle, and this defines the "crime zone". What they found was that something like 80% of offenders lived within the crime zone and these offenders are referred to as "marauders", the other 20%, those who lived outside the crime zone, are called "commuters" - they travel to the crime zone. Now, it also turns out (I forget who found this relationship), but the maximum distance offenders travel to a crime location from their anchor point (usually their residence), is roughly 80% of the diameter of the crime zone. That means, they tend to live near the edge of the crime zone, rather than near the middle of it (tend, there are exceptions). They also tend to travel a minimum distance, leaving what is called a "buffer zone" around their residence, but this relationship is weaker, so there are lots of exceptions.

Kim Rossmo, whose PhD work really got geographical profiling going, plots probability distributions around each crime scene (which have a buffer zone, and then a distance decay function which is based upon solved crimes, etc) and uses these to find locations that have high probabilities because these functions interact with each other based upon where each location is. Based upon his approach, areas of interest for the JtR highlight Flower & Dean, and two minor hot spots (one near the Goul. Street, and one that is located half way between Kelly and Chapman's location then shifted north west to form an equilateral triangle, and that's the other minor location of interest).

Anyway, there are a number of different approaches, but basically the idea is to try and extract information from the pattern of crime locations that can be used to then predict the offender's anchor point. And by that, I just mean, reduce the search space to be better than chance. The offender may not be right at the peak of the geographical profile, but what you're hoping is that if you search the area guided by the output you will come across the anchor point sooner than if you just do a random search.

Some do try and factor in psychological components, like boundaries or other things that might influence how somebody conceptualizes "space".

Like any analysis based upon extracting group trends, it won't always predict every individual case, and it will get it wrong sometimes (commuters, for example, will fall outside the hotspots because of the way these analyses work - so right there you've got about a 20% failure rate. It's not entirely clear how useful the approach is, and there is a lot of room for improvement, but it is quite an interesting topic. Here is the output from my own work, not published, on geographical profiling with regards to the JTR case, if I input the locations for Tabram + the C5 (based upon Keppel's linkage analysis). The pinkish areas are the top suggested search zones, and there are contours falling away from that. I'm quite partial to the zone around Hanbury street, simply because that crime occurred so close to morning, when the streets would be busy, I've always thought that JtR must have figured he could get home really quick after that. Also, you could see how going from Eddowe's to Goulston street could reflect someone heading in that direction. Mind you, the lower left top zone works too. However, we see what we want to see. And no, I'm not saying this is correct and Jack must have lived in one of those pink areas. Rather, this would simply be used to prioritize the search. In my testing of this with solved crimes, the anchor points tend to fall within 4 or 5 zones (pink is zone one, the next red ring would be zone 2, etc), so it's still a pretty big search space, but it is smaller than searching the "crime zone" at random. Is Jack in zone 1-5? Have no idea. If he's a commuter, then no. I will tell you that BTK's house falls in zone 1, and Ted Bundy and the Golden State Killer both were in zone 2 (none of those series were used to develop the analysis, so they are independent tests of it; McArthur in Toronto was a commuter, but the house where he was hiding the bodies (where he worked/kept his supplies, so is an anchor point) falls in zone 7 (which isn't bad given the only locations I had to work with were press reports indicating where the victims were last reported as being seen). However, Desalvo (Boston Strangler) ends up in zone 20 (he's much more centrally located than most), so if Jack is like that, well, ....


- Jeff