John Richardson

[FISHY1118]

If we dismiss Cadosch though George then there’s absolutely no reason that Long couldn’t have been correct. 5.35 TOD still within range.

The fact that you claim in your senario that he somehow brushed up against the fence which like everthing else in this incident, cant be proven as fact. There for its just speculation.

Your entitled to think that , However my suggestion worked then and it still works now .

[GBinOz]

I know this theory has been posited before, but perhaps someone encountered Annie's corpse. They tried to move the body, check for signs of life, saw the ghastly mutilations and let go in horror. That would explain the "No!" in disbelief, and the slump against the fence.

Why did this person(s) never come forward? Well, they might have been venturing into the yard for unsavory reasons themselves and didn't want to be questioned by police.

Hi Harry D,

That is quite a good speculation. The reservation is usually that Cadosch should have seen the door opening and closing as the person(s) arrived and left. But perhaps they arrived while he was in the toilet the first time and left after he went inside the second time. It also solves the open front door conundrum.

Cheers, George

[GBinOz]

If we dismiss Cadosch though George then there’s absolutely no reason that Long couldn’t have been correct. 5.35 TOD still within range.

Herlock, pleeeeze,

This is all you need to see to realise the lack of value of her testimony:

Was it not an unusual thing to see a man and a woman standing there talking? - Oh no. I see lots of them standing there in the morning.
[Coroner] At that hour of the day? - Yes; that is why I did not take much notice of them.

Three days after the event she convinces herself that she had seen Chapman—a woman she’d never seen before and glimpsed for a few seconds in unremarkable circumstances.

Cheers, George

[FISHY1118]

Hi George,

Your first question pointing out the reliability of ToD estimates in the case of Nichols, Stride and Eddowes is simplicity to answer. In the Eddowes case, Inspector Collard reported, "the blood was in a liquid state, not congealed". In the Stride case PC Lamb mentioned that some blood was congealed, and some liquid. In the Nichols case, PC Neil reported, "blood was oozing from a wound in the throat". These virtually didn't require a police surgeon's opinion at all, just how many minutes! These murders were very recent. The Chapman murder bore no resemblance whatever to these, and required a genuine, and inevitably approximate calculation. And an estimate on site which Phillips himself voluntarily chose to put a large question mark against at the inquest.

Although I personally admit to reservations about the testimony of both Cadosch and Long, I do not necessarily accept that Cadosch or Richardson for that matter, changed their stories. We can safely assume that both Richardson and Cadosch gave detailed statements to the police. We do not have these statements, nor do we have any statements from interviewing officers, nor any evidence from official sources, that these two said anything different in their final statements from their original ones. What I assume that you are using is garbled accounts in newspapers where journalists may have "sexed up" a weakish story to make it worthwhile, or may have got partial information from someone who overheard a conversation etc. We cannot claim that casual newspaper tittle-tattle is evidence of changing stories.

Originally posted by FISHY1118 View Post

Another question ,If Dr Phillips gave his t.o. d at 45/55 mins from when he arrived, do the same principles apply?, In other words based on the condition of the body at 45/55mins, would/could the same argument be then used against Phillips .

If not, what then makes Phillips assessment correct at 45 that makes him wrong at 3 hours ? [sorry two questions]

Now i know some will say he could have been right so as to sit on the fence , but lets face it a lot of posters think he was wrong but just want to come out and say it, thats fine . So for the sake of argument lets say he was wrong then apply my question to try and prove otherwise .

Just to be clear ill put in in another way , is it just as hard, risky, unsafe, accept,[ you know the drill by now Victorian Doctors and pretty much was guesswork and all ] for a doctor in 1888 to determind a t.o.d from 0 to 45 as it is from 2 to 3 hours ?


Hi Fishy,

The same principles apply, though I can't say for sure the same error range would apply. It is possible that the error associated with very short PIM estimates is smaller, and it is possible it is larger, and it's possible the same range applies; it's an empirical question not an intuitive one, so without seeing some studies on very short intervals I'm not sure.

So by jeffs response we can see that at the very least the same principles apply in regards to tying to determine a shorter t.o.d. by victorian doctors.

The question remains and warrent asking if Long , Richardson and Cadoash are correct why didnt Dr Phillips just give his T.O.D starting from 5.30 ?

[GBinOz]

Why do we assume a talkative struggle if she was strangled?

That's what I was wondering. According to Cadosch there must have been at least 5 minutes between the "No" and the fall against the fence. As Fishy said, Jack worked from the body's right and there is no reason the speculate that he changed that method and tried to squeeze between the body and the fence. Jack wasn't known for being chatty once he got the victim alone. He blitz attacked and finished the job quickly.

Cheers, George

[DJA]

All evidence,with the possible exception of Mary Ann Kelly,is that JtR did NOT blitz attack.

[GBinOz]

Hi George,

Your first question pointing out the reliability of ToD estimates in the case of Nichols, Stride and Eddowes is simplicity to answer. In the Eddowes case, Inspector Collard reported, "the blood was in a liquid state, not congealed". In the Stride case PC Lamb mentioned that some blood was congealed, and some liquid. In the Nichols case, PC Neil reported, "blood was oozing from a wound in the throat". These virtually didn't require a police surgeon's opinion at all, just how many minutes! These murders were very recent. The Chapman murder bore no resemblance whatever to these, and required a genuine, and inevitably approximate calculation. .

Hi Doc,

I find myself at a loss as to how you can deduce that the Chapman murder bore no resemblance to the Eddowes murder. How were they so completely different?

Sequeira, with a PMI of not more than a quarter of an hour, assessed the ToD as not earlier than 1:40. Brown, with a PMI of 30-40 minutes, assessed the ToD as between 1:40 and 1:50. For medical methods that were so completely unreliable, I'm not seeing large margins of error here, or evidence of over estimating in either of the two estimates. Brown described Eddowes body as "quite warm - no rigor mortis". Phillips said Chapman's body was "cold, except that there was a certain remaining heat, under the intestines, in the body. Stiffness of the limbs was not marked, but it was commencing". The mutilation to the abdomens of both bodies were almost identical, but it is proposed that there is only 20-30 minutes for Chapman to loose all body heat, except for a small residual under the intestines, and for rigor to commence. I am not convinced.

The science of surveying includes the study of error theory. Small sample groups are notoriously unreliable due to accumulation of errors and compensating errors. A model equation such as Jeff produced when tested on one small sample of data can produce a markedly different result when a second small sample is tested. It is only when significant data observation sizes, three figures or better, are examined that true figures and margins of error consolidate. That's what I was taught, anyway.

Cheers, George

[JeffHamm]

Hi Jeff,

Equations have the property that they cannot be solved if there are too many unknowns. In the case of the C5 murders, were are not privy to any of the input data for the equations, and have a range of variables that may have been used. You have used Phillip's resulting PMI to work backwards and assess the input data that he would have used in the various methods of assessment that he may have adopted.....I think, have I got this wrong? Having arrived at a PMI there is a margin of error which has been shown by your examples to not conform to a bell curve, tending more towards over than under estimation. I may be misunderstanding, but I'm thinking that in a large sample the under estimate side would still be populated but to a lesser extent than the over estimate side.

What I was asking in the second part of my previous post was, using your models can a table be produced for the 8 (or 9) doctor's estimates showing the PMI's, or ToD's of the C5 murders together with their assessed error margins?

Hi George,

Ah. The thesis data set of only 12 cases is the best one I have at the moment as most of the murders cover pretty short intervals. There's only 4 cases with true PMI under 3 hours, which certainly applies for all but Kelly's murder. I would be reluctant to rely too much on only 4 cases, and beyond the fact that they do indicate that the margin of error does not appear to get smaller, and it appears to get larger (the 3 largest overestimations come from those 4 cases). But, just in case you're interested, the average overestimation for those 4 cases was 5 hours and 39 minutes, with a standard deviation of 3 hours 34 minutes. To get the +- range, multiply the standard deviation by 1.96 (two will be good enough, so the error range would be roughly +- 7). So the estimation would be expected to range anywhere from about 1.5 hours underestimation out to 12.5 hours overestimation. However, as I say, with only 4 cases to work with, these values are really best simply viewed as indicating the reliability does not improve with shorter intervals.


If I use all 12 cases, on the basis that the overestimation and error ranges are more uniform over PMI from 0-10 hours (which covers the C5 pretty well), then the average overestimation is 2 hours and 52 minutes, with a standard deviation of 2 hours and 58 minutes, making for a +- of 6 hours (for ease). So, again, given the variability, we're looking at a 12 hour spread in terms of confidence intervals in terms of how small or large the overestimation is.

I suppose, one could try to work out an adjusted PMI, to correct for that average overestimation and place the +-6 hours around that.

For example, Dr. Phillips suggests a PMI of 2 hours, making the ToD estimated to be at 4:30. We subtract the average overestimation, to get the time 7:22, so the ToD would be expected to fall in the range of 1:22 to 13:22. Obviously, as she was discovered at 6, we can narrow that down to between 1:22 and 6:00. That's assuming he used the /1.5 version of the equation. It just gets worse if he used the /1 version that Wickerman found (the mean and standard deviations double, so the +-6 would become +-12).

The method is simply not very precise. I suppose one could do that for all of the C5, though perhaps Kelly is the one case where it might be useful. But if I recall, the estimated ToD is around 4am, and she was supposed to be sighted at 8 or 9. So again, 4am gets adjusted to (roughly) 7am, making the estimated ToD between 1am and 1pm. Because we know the time at which she was found (around 10:45), we can then start to chip away at that window and suggest the ToD was between 1am and 10:45. So no, we can't use this to rule out an early morning sighting. As I say, it's a very crude tool, and it's not helped by the small sample of data I have to work with, but the fact the +- I get falls in the range cited in the literature (between +-3 to +-7) is something at least.

- Jeff

[GBinOz]

All evidence,with the possible exception of Mary Ann Kelly,is that JtR did NOT blitz attack.

Hi Dave,

Is your objection with the word "Blitz". I was using it to mean that once achieved a suitable location he didn't waste time with chit chat, but struck quickly with the strangulation followed by a quick non frenzied disembowelment. If this was not your objection, please elaborate.

Cheers, George

[Wickerman]

But we dont know how Phillips arrived at the estimated time of death, and you have done some sterling work on this topic but that being said no two murder victims bodies will react in the same way as far as cooling is concerned there are many factors which will be diffferent in each murder, so the figures you have quoted cannot safely be attributed to Chapman all you are in effect doing is using the figures to guess a likley time of death the same as Phillips did.

www.trevormarriott.co.uk

We have the opinion of the Lancet, who also agreed with the conclusions of Dr Phillips.

There could be little doubt that he first strangled of suffocated his victim, for not only were no cries heard, but the face, lips and hands were livid as in asphyxia, and not blanched as they would be from loss of blood. Then, with one long and deep incision he must have severed the poor woman's throat, so that almost all the blood from her body drained out of the divided vessels, accounting for the almost bloodless effect of the subsequent incisions in the abdomen and pelvis. If the evidence of Mrs. Long is to be credited, the victim was seen alive at half-past five in Hanbury-street, and about six o'clock her mangled corpse was discovered in the yard of the lodging-house. We confess to sharing Mr. Phillips' view that the coldness of the body and commencing rigidity pointed to a far longer interval between death and discovery than this; but, as he remarked the almost total draining away of the blood, added to the exposure in the cold morning air, may have hastened the cooling down of the body. Certainly the murderer must have done his work quickly; and this, again, points to the improbability of anyone but an expert performing the mutilations described in so apparently skillful a manner.
The Lancet, 29 Sept. 1888.

[GBinOz]

Hi George,

The method is simply not very precise. I suppose one could do that for all of the C5, though perhaps Kelly is the one case where it might be useful. But if I recall, the estimated ToD is around 4am, and she was supposed to be sighted at 8 or 9. So again, 4am gets adjusted to (roughly) 7am, making the estimated ToD between 1am and 1pm. Because we know the time at which she was found (around 10:45), we can then start to chip away at that window and suggest the ToD was between 1am and 10:45. So no, we can't use this to rule out an early morning sighting. As I say, it's a very crude tool, and it's not helped by the small sample of data I have to work with, but the fact the +- I get falls in the range cited in the literature (between +-3 to +-7) is something at least.

- Jeff

Hi Jeff,

I hope that you discerned that my post was in no way meant to be critical of your work or your results. As you say, you had available only small sample groups to work with - that's just the way it is. Would you agree that if a sample group numbering in four figures were to fall into your hands, it may become apparent that your small sample was populating a fringe area of the total, and perhaps the C5 sample populating a completely different area?

I find it ironic that some posters (not you) that support the witness testimony for Chapman as totally reliable, and relegate the medical estimate to the unreliable category, change their song sheet when it comes to the witness testimony of Maxwell that you mention above. Even though Maxwell was steadfast and unwavering in her testimony, and could not be broken by Abberline or tricked or threatened by the coroner, she is labelled as unreliable, while the medical estimate is suddenly transformed into being reliable evidence, combined with reports of a cry of murder, the caveat of which was that they were stated to be a regular occurrence in that area.

Each person has to make up their own mind. My opinion meter is at present about 15% off centre towards the earlier ToD, but the meter is in good working order and can still shift in either direction given appropriate argument. Once again, thank you for your invaluable work.

Best regards, George

[JeffHamm]

Hi Jeff,

I hope that you discerned that my post was in no way meant to be critical of your work or your results. As you say, you had available only small sample groups to work with - that's just the way it is. Would you agree that if a sample group numbering in four figures were to fall into your hands, it may become apparent that your small sample was populating a fringe area of the total, and perhaps the C5 sample populating a completely different area?

I find it ironic that some posters (not you) that support the witness testimony for Chapman as totally reliable, and relegate the medical estimate to the unreliable category, change their song sheet when it comes to the witness testimony of Maxwell that you mention above. Even though Maxwell was steadfast and unwavering in her testimony, and could not be broken by Abberline or tricked or threatened by the coroner, she is labelled as unreliable, while the medical estimate is suddenly transformed into being reliable evidence, combined with reports of a cry of murder, the caveat of which was that they were stated to be a regular occurrence in that area.

Each person has to make up their own mind. My opinion meter is at present about 15% off centre towards the earlier ToD, but the meter is in good working order and can still shift in either direction given appropriate argument. Once again, thank you for your invaluable work.

Best regards, George

Hi George,

No, I didn't think you were being critical. Small samples are not great, they have margins of error with respect to how well they represent the larger population (which is what we really want to know about). So yes, it's possible this data set is spurious, but it's also possible it's not. I don't see any reason to suggest it is spurious though, and it would take another larger data set to compare with in order to make that call.

So yes, I would most definatly go with the outcome of a much larger sample, that's always more reliable. I would be surprised if the results from the small sample ended up being bang on, and I rather suspect the overestimation averages would indeed change - certainly the estimation of the population variations are probably off.

But the data we do have is not pointing in the direction of smaller margins of error for shorter time frames. If I find more, I'll add to this, and see what comes up.

- Jeff

[DJA]

Hi Dave,

Is your objection with the word "Blitz". I was using it to mean that once achieved a suitable location he didn't waste time with chit chat, but struck quickly with the strangulation followed by a quick non frenzied disembowelment. If this was not your objection, please elaborate.

Cheers, George

Yep.

Reckon you've watch too many of Trev's videos.

[Fleetwood Mac]

You get worse.

So because we have no record of what she did from 1.45 until the time of her death we have to assume that she couldn’t have eaten again? During that period of time for Annie Chapman time simply stood still.

Please think before you say things.

Dear Sherlock,

You clearly do not understand the difference between evidential basis and supposition.

How on earth can your theory be taken seriously?

A large part of your theory is built upon supposition, invention and Long/Cadosch/Richardson 'trivial errors' (in your own words).

Conclusion:

Your case has collapsed, the jury aren't listening anymore: instead they're having a cup of tea and a slice of cake and are talking about the hitherto unheard of quality of jam making at the Sussex Women's Institute, and the judge is about to throw the case out of court and put you in the dock for contempt of court due to wasting the court's time and expending taxpayer's money on a whim.

Case dismissed. The judge has no time for this spellbinding nonsense and simply says: "get out of my court".

[DJA]