John Richardson

[Herlock Sholmes]

I'm not interested in going round nonsensical arguments with you, even though it is obvious from your own quote in your link that they conclude rigor is useful in estimating TOD. In the event I respond to this with the same comments we will merely serve to obfuscate that which you're saying (which of course is in your interests).

What I want to ask you is this: is Dr Phillips's TOD estimate valuable?

And I’m not interested in listening to you try and twist the truth to fit your own preconception.

Is Phillips TOD valuable? No. Not in the slightest because we have no way of assessing it’s accuracy. There’s a very real, very possible, very substantial chance that he was wrong. If it was a case of there being a 90% chance of him being correct than that would tip the balance of probability in his favour but there is nowhere near that. As far as we’re concerned, with the evidence that we have can’t push us either way. And even if there was only a 10% chance of him being wrong that would still leave open a very real chance of the witnesses being correct. But we can’t talk about percentages; we can’t assess; we can’t rate; we can’t judge. Therefore Phillips gets us nowhere.

[Fleetwood Mac]

Is Phillips TOD valuable? No........Therefore Phillips gets us nowhere.

I'll give you a small amount of credit for not backtracking.

Let's be clear about this:

1) While estimated TOD can never be certain, estimating TOD is critical to crime resolution. Modern forensics commentators tell you that and the fact they actually undertake this exercise on a routine basis all over the world tells you that.

2) Dr Phillips was a well respected, experienced doctor estimating the time of death.

3) It should follow that while Dr Phillips couldn't say with certainty, he could provide a reasonable estimate (open to being inaccurate, however).

'Just out of idle curiosity, do you disagree with anything in points 1 and 2?

[Herlock Sholmes]

Here's a link to a summary report on estimating ToD. Basically, researchers in Amsterdam developed a new method for estimating ToD using thermal cameras, and by incorporating a lot of environmental factors (i.e. is the body in water, amount of clothing, ambient temperature, body structure, mass, etc).

After taking into account all sorts of specific factors about the physical size and build of the victim, of the surrounding environmental conditions, and so forth, they have been able to get the estimated ToD down to be within 45 minutes of the actual ToD (so "real time +- 45 min" is how good we can do now, but only after factoring in all of those above variables - none of which Dr. Phillips did, or knew to do).

Compared to the existing methods, which involves taking a rectal temperature reading (which Dr. Phillip's did not do), the the victim's weight (not used), and ambient temperature (as in degrees, not "it was cool that morning), one looks up on a table the estimated hours passed. But, as the article points out "Existing methods can only give a window of several hours; the new method reduces this to less than an hour."

Now look at that, bold bit: Existing methods only give a window of several hours...", and that's using modern existing methods, with proper temperature readings (from the body and the environment) and factoring in the victim's weight (the new method also factors in their "build", someone who is 130 lbs but short and stocky will show a different temperature pattern than someone long and lean; even the current methods don't factor that into their calculations).

Dr. Phillip's didn't take a temperature reading, he touched the body. He didn't record the ambient temperature, he just noted it was cool. And he didn't take into account Annie's weight. But even if he had better quality information to work with his estimate would still have a window of several hours!

I don't care how many years experience he may have with touching bodies for temperature - even if he could do better, and actually measure the body temperature, he still can't be more accurate than "real time +- several hours".

As I say, his "estimate of 4:30", once you factor in the error (+- several hours), is not at odds with a ToD at 5:25 (ish). This isn't my opinion, it is a description of how reliable the method is in determining the ToD - and frankly, in my opinion, it's pretty awful. If he had the fancy new method, then we're starting to get to the point where we might be able to separate 4:30 from 5:25 ish (but they only claim that +- 45 minutes for bodies dead at least 5 hours, so it is either more or less reliable for shorter time intervals).

- Jeff

This post alone should be game absolutely over. I thought that a bit of highlighting was in order.

[Fleetwood Mac]

Here's a link to a summary report on estimating ToD. Basically, researchers in Amsterdam developed a new method for estimating ToD using thermal cameras, and by incorporating a lot of environmental factors (i.e. is the body in water, amount of clothing, ambient temperature, body structure, mass, etc).

After taking into account all sorts of specific factors about the physical size and build of the victim, of the surrounding environmental conditions, and so forth, they have been able to get the estimated ToD down to be within 45 minutes of the actual ToD (so "real time +- 45 min" is how good we can do now, but only after factoring in all of those above variables - none of which Dr. Phillips did, or knew to do).

Compared to the existing methods, which involves taking a rectal temperature reading (which Dr. Phillip's did not do), the the victim's weight (not used), and ambient temperature (as in degrees, not "it was cool that morning), one looks up on a table the estimated hours passed. But, as the article points out "Existing methods can only give a window of several hours; the new method reduces this to less than an hour."

Now look at that, bold bit: Existing methods only give a window of several hours...", and that's using modern existing methods, with proper temperature readings (from the body and the environment) and factoring in the victim's weight (the new method also factors in their "build", someone who is 130 lbs but short and stocky will show a different temperature pattern than someone long and lean; even the current methods don't factor that into their calculations).

Dr. Phillip's didn't take a temperature reading, he touched the body. He didn't record the ambient temperature, he just noted it was cool. And he didn't take into account Annie's weight. But even if he had better quality information to work with his estimate would still have a window of several hours!

I don't care how many years experience he may have with touching bodies for temperature - even if he could do better, and actually measure the body temperature, he still can't be more accurate than "real time +- several hours".

As I say, his "estimate of 4:30", once you factor in the error (+- several hours), is not at odds with a ToD at 5:25 (ish). This isn't my opinion, it is a description of how reliable the method is in determining the ToD - and frankly, in my opinion, it's pretty awful. If he had the fancy new method, then we're starting to get to the point where we might be able to separate 4:30 from 5:25 ish (but they only claim that +- 45 minutes for bodies dead at least 5 hours, so it is either more or less reliable for shorter time intervals).

- Jeff

There is no link.

Can you post the link so that anyone interested can have a look, assess its content and determine whether or not this is peer reviewed.

[JeffHamm]

That wasn't the estimate of Dr Phillips. He specifically stated: "at least two hours and probably more".

Hi FM,

Yah, I shouldn't have used quotes. His 2 hours, though, would be from the time he did his touching, which would be some time after his arrival at 6:00, but we don't know how long. I tend to think, based upon his testimony of things he observed at the crime scene, time required to just observe the positioning of the body, visually assess and record his observations, and the fact he's likely to have an initial discussion with the police about circumstances, whether or not anything has been touched, and so forth, that his touching for temperature was probably not until 6:30 ish (give or take), and so his 2 hours suggesting 4:30. But, if one envisions him immediately going over and touching the body upon his arrival, then one could argue for his 2 hours indicating 4:00 I suppose.

Regardless, given the fact that even if he took proper internal temperature readings rather than subjective surface temperature readings the best he could do is be within several hours, it's a bit of a moot point.

- Jeff

[JeffHamm]

There is no link.

Can you post the link so that anyone interested can have a look, assess its content and determine whether or not this is peer reviewed.

Click on the red words "summary report", that's the link. The report is summarizing the findings of their peer reviewed article, which can be found here. I have access to it at least, but that may be through work, so I'm not sure if it's available to everyone or if it is behind a paywall.

- Jeff

[JeffHamm]

Hi all,

I've located the article that is summarized above (found here, but I'm not sure if it will be available to everyone and I may just have access through work).

Just to give more details on the reliability of their new approach, they indicate "We found the maximum and average error of these reconstructed PMIs to be as low as ±3.2 hours and ±38 min, respectively. Moreover, 83.3% of the reconstructed PMIs deviate no more than ±1 hour from their corresponding true PMIs.", so even with their improved technique the error can still be +- 3.2 hours (that was their maximum, so such large deviations are rare), with the average being +- 38 minutes (which is pretty impressive). They report an 83.3% Confidence interval of +-1 hour, which is a non-standard thing to do, usually one reports a 95% range, which will be some amount more than +-1 hour, but it's not clear by how much.

They also report "Errors of the reconstructed PMIs induced by parameter-specific uncertainties remain within ±2.5 hours for true PMIs ranging from 6 to 40 hours.", so once they factor in some other unknowns, their estimates are +-2.5 hours (so even their method might not be able to distinguish Dr. P's estimated ToD from the witness based one). For example, in their study they know the initial body temperature of the deceased, but once one admits that they don't know the initial body temperature of a murder victim and has to include that as an "unknown parameter", their calculations produce estimates with error ranges much more than the +-1 hour and we're looking at +-2.5 hours.

But, as they say, "Notwithstanding, these results represent a notable improvement over the current gold standard (Henssge’s nomogram) where uncertainties range from ±3 to ±7 hours." - That's the current "gold standard", which is far more advanced than Dr. Phillips had available to him. And reports on the accuracy of the current "gold standard" mean at a minimum one has to consider the estimate "ToD +-3 hours", though others will argue for as much as "ToD +- 7 hours".

Basically, Dr. Phillips' estimate in many ways is rather impressive. Using nothing more than a touch to obtain a subjective impression of her surface temperature, and noting the coolness of the day, he produces an estimate that even the gold standard of today (and even the "new improved method" when having to deal with some unknown values, like Annie's body temperature at the time she was killed), once coupled with it's margin of error, fits with the witness statements. Meaning, he seems to have done as well as a modern set of measurements could do, so hats off to him for that. But when we look at his estimate, we have to recognize that his estimate has a very very large window around it, and it is one that encompasses the witnesses.

- Jeff

[JeffHamm]

Hi all,

I just want to emphasize what I'm saying here "...But when we look at his estimate, we have to recognize that his estimate has a very very large window around it, and it is one that encompasses the witnesses", which is that Dr. Phillips' estimate of the ToD is not actually in conflict with the witness based ToD. Basically, all of the evidence we have, from the witnesses and from Dr. Phillips is consistent with Annie being killed after Richardson's visit. Nothing is actually in conflict here.

- Jeff

[Abby Normal]

Hi all,

I just want to emphasize what I'm saying here "...But when we look at his estimate, we have to recognize that his estimate has a very very large window around it, and it is one that encompasses the witnesses", which is that Dr. Phillips' estimate of the ToD is not actually in conflict with the witness based ToD. Basically, all of the evidence we have, from the witnesses and from Dr. Phillips is consistent with Annie being killed after Richardson's visit. Nothing is actually in conflict here.

- Jeff

hi jeff
your link works and has from your original posting of it. thanks very informative! and yes once we take into consideration the window of tje estimate and when phillips actually did his examination we can see that actually he his not even in conflict with all the other witnesses re a later time of death around 5:30.

debate over.

[JeffHamm]

hi jeff
your link works and has from your original posting of it. thanks very informative! and yes once we take into consideration the window of tje estimate and when phillips actually did his examination we can see that actually he his not even in conflict with all the other witnesses re a later time of death around 5:30.

debate over.

Thanks for confirming the link is working. It worked for me, but I was worried I may have done something wrong (something that happens far too often I'm afraid )

Anyway, glad that we agree. That's always nice, but we shouldn't get used to it! ha ha

- Jeff

[harry]

I mentioned Spilsbury,Herlock,because you hadn't,not because you had.So read properly and stop your accusations.
For good measure it might help the cause of those using lack of knowledge,if they quoted word for word,what it is that modern day people are saying regarding lack of knowledge,especially in the case of estimation of how long a person had been dead.You know how to use quotes,and the information is there you claim.
I do not question Phillip's times as I have no qualification to do so.I do know he was a surgeon,and had years of experience.others have proved that.That there were teaching hospitals and universities ,in Victorian times,where doctors gained knowledge,I have no doubt.That Phillips could estimate,with some degree of acuracy how long Chapman had been dead I do not question.
That he could be fifty per cent,or an hour out I do not believe.

[FISHY1118]

His level of experience is irrelevant. The level of forensic medical knowledge in the Victorian era is relevant though. Nothing has been overlooked although some have ‘overlooked’ the fact that modern day medical experts are 100% consistent on the fact that TOD estimation is unreliable. Phillips might have been the best Doctor in London for all that we know but it doesn’t matter. There is no tipping of the scales. Phillips should be ignored.

In your opinion , not mine. or others for that matter, he could, as others people have accepted been correct. .

Ill leave the Phillips debate to others as this is more a richardson thread, which going by the number of post back and fourth the conflicting evidence also leads people to a number of different possibilites . All based on the evidence.

[GBinOz]

Hi all,

I just want to emphasize what I'm saying here "...But when we look at his estimate, we have to recognize that his estimate has a very very large window around it, and it is one that encompasses the witnesses", which is that Dr. Phillips' estimate of the ToD is not actually in conflict with the witness based ToD. Basically, all of the evidence we have, from the witnesses and from Dr. Phillips is consistent with Annie being killed after Richardson's visit. Nothing is actually in conflict here.

- Jeff

Hi Jeff,

You have presented data on the unreliability of ToD's in the Victorian era. Do you have any similar studies into the unreliability of witness testimony in the Victorian era, and in modern times with the advent of CCTV footage?

Best regards, George

[GBinOz]

My objective for participation in this forum is to learn from the views of others without any compulsion to conform to those views. Nothing in my objective persuades me that I am entitled to belittle the opinions of others. I have observed that some posters do not seem to feel constrained by that basic courtesy. I am now wondering if, as suggested by Scott Nelson on a different thread, my time may be better dedicated to reading back issues of Ripperologist.

Cheers, George

[JeffHamm]

Hi Jeff,

You have presented data on the unreliability of ToD's in the Victorian era. Do you have any similar studies into the unreliability of witness testimony in the Victorian era, and in modern times with the advent of CCTV footage?

Best regards, George

Hi George,

I posted some information on witness testimony with regards to the duration of events a while back on the Stride thread. And as we saw there, estimates tend to be slightly inflated until you reach about an hour, afterwards they tend to underestimate. And, like the ToD estimations, just how wide those errors can be is far more than most of us would find believable. But that's just the problem, we think we're more accurate at things than we are.

I've been trying to find studies on the accuracy of eye witnesses statement of the time (it was at x o'clock type thing), but so far no luck. Basically, eye witness testimony is fraught with problems, there are almost always details incorrect, or things out of sequence, and so forth. And of course there are times when they are just way off base. I've not found anything that let's me put numbers to any of those statements, though (as in 5% of the time we can expect the witness to be completely wrong, while 80% of the time the overall story is correct, with only details out of place, etc; I just made those up to illustrate).

But, as Trevor points out, eye witness testimony is error prone, particularly in fine details. We need to remember that, and not be too concerned in detail changes between tellings (because we can't go back to verify which, if any, version is the right one), and just look at the "gist" of their stories. I think that's the best we can get, but that's just my view.

- Jeff