Attempting any precision from the age analogy is fraught with issues but here goes:
The chance of living to 105 when born (group eating meal) is .074%. Almost impossible just like not starting to empty in 2.5 or 3 hours. However if one lives to 100 the chance of making it to 105 is 14.6%.
You didn't notice the same thing in LJ's chart? Or the height analogy I used?
Look again at what happens to the percentages when everything left of 9:00 is eliminated or you remove the possibility they're shorter than 6'2". What you're looking for is there. Just as the chance of living to 105 goes from .074% at birth to 14.6% if you make it to 100, the chance of her having died between 9:00 and 9:20 goes from 1.9% to 95% simply by the fact we know she was alive at 9:00. The same factor increases the probability she died between 9:20-10:00 PM from 0.09% to 4.5% based on the results of that study which had a normal distribution (unskewed bell curve).
The linked "analysis" seems to be based purely on the bell curve and not the science of digestion but even so it says much of what I've been saying
No, it says the
opposite. If you're talking about the caveat, then the answer to that is to look for more studies and greater samples. This has been done, but please, feel free to look for more. The best of course are the ones which take into account the starch from the pizza, are for a small to moderate meal and include alcohol. They all have one thing in common: no or almost no data after 3.0 hours, not a single subject starting emptying that late.
As for being based on the bell curve...hmmm...lemme put it this way: have you ever wondered why every one of those studies posted includes a curve and notes whether it's a normal distribution or skewed (for these studies) to the right? Why they all start going up and then start going down and always hit the bottom? There's really no zigging, or zagging, (pretty much) no frills no mess, but sometimes a 'tail' (always to the right)?
Those curves are based on the actual data, it's the data that creates the curve and just so happens to look like a bell curve, often with a tail to the right. Something like
this.
If you're familiar with these, you can look at the figures for T1-4 (based off the actual data) and
know what the curve is going to look like, much like you could look at the figures for a bar graph with 4 numbers such as 8-7-6-5 and know it will look like, (steps going down to the right). Look at that one I linked, which perhaps I should have done before when saying things like being on the 'far right' of the curve, and how you should [/i] know[/i] that anything to the left of the curve is going to be a higher probability. That's what the data says, those are just physical representations that illustrate what happens to the probability, often they look very similar to each other, even for wildly different subjects. A graph of the probability of rolling each number on a set of dice (2-12) looks very much like a graph of the heights of people in the US, a bell curve.
There's a reason for that, and it's not magic, proof of God or cheating, but it is one of the Secrets of the Universe and if I told you then you'd have to shoot me. That might not be fun for me and you could get in trouble, so we can't have that!
the linked "analysis"
a) Prob she died before 9pm = 0%
b) Prob she died between 9pm and 9.20pm = 95%
c) Prob she died between 9.20pm and 10pm = 4.5%
d) Prob she died later than 10pm = 0.5%
Of course, the caveat to all of this is that the bell curve isn't exceptionally precisely defined at such a far end of the curve. This of course is why I originally introduced a +/- 25% error to those numbers. However, even when such a large error factor is applied, it's patently clear that the overwhelming likelihood is that Meredith died well before 10pm, and that it's virtually impossible that she died later than 11pm (or at the very least was violently attacked any later than 11pm).
I've said 5% but 9:30 instead of 9:20 IIRC but really since we don't know when or how much she ate (I've looked and so far nothing close to a measure of how many pieces of pizza or how much apple crisp and ice cream) the ten minutes has no significance.
I thought it would be recognized that by using 9:20 that analysis
exceeded your parameters. The probabilities for 9:00 to 9:30
have to be higher than 9:00-9:20, which
reduces the probability Amanda and Raffaele could possibly be involved.
As I also have said all along the death was clearly well before the original TOD of 11:30.
I have said that I believe that the phone activity shows that the TOD was no later than 9:45 or 9:55.
This is something that strongly corroborates that, and shows that the earlier it was the more likely.
At what point did Lalli change? What did he change? Was this before or after he was removed from the case?
He changed his estimate from 3-4 hours to 2-3 hours. The first I believe was in his report, the second happened during the trial. I think both were after he'd been removed from the case, he did his work but hadn't filed his report until after Mignini cashiered him. I wish I remember exactly why that was.
Introna in his testimony has TOD between 21:30 and 22:30
pm 
p)
LOL!
Finally I get it! (I think)--You said something like that before and I couldn't figure out what you meant. I gather now you mean that military time doesn't need a 'PM' and some little bunny screwed the pooch on that one? Sorry, my mind automatically translates military time to normal time, I didn't even notice.
Read the first line or so from what you posted to me the other day.
Here the defense asks about 21:00:
DEFENSE – so that he may be regarded as the time of death at around 21.00 hours taking account of evidence that there have been offers, i.e. that the beginning of the meal would be for someone around 17.30?
Consultant-(IDREES) – Some by a reasoning that does not consider the extreme highest but the extremes that you consider the Medes and also willing to consider the hours we have more 18.30 18.30 three hours are emptying at 21.30 21.30 up is due to begin the attack.
I can't find where he has it before 21:30.
I posted it to you, citing Massei! If you don't find it in his testimony, look for his report if it's available. The British girls also had interviews with Mignini that aren't in their testimony as well, though some of it is represented in Massei.
Regarding the British girls and my 'damning with faint praise' comment regarding them being more likely to be involved than Amanda and Raffaele, I didn't say anything about Meredith being murdered before she got back to the cottage, did I?
Now, how exactly do you know it was at 9:00 PM?
Lemme go over their testimony with my Kook Mignini hat on and I'll find all sorts of 'inconsistencies.' Even you mentioned the testimony was 'vague' and they didn't know some things you expect they'd know--to some that's evidence they're
lying therefore they must be
guilty! The 'only' reason they couldn't know would be because it never happened and they're covering for their nefarious 'involvement' in Meredith's murder! They
were the last people who saw her alive (outside Rudy--their pimp

p) ) and what did they do when they got back to Britain? Started spreading malicious (but entirely irrelevant) gossip to defame Amanda--trying to implicate an innocent woman!
Obviously they must be involved!
Then I put on my Stefanoni hat: Lemme at those samples taken from the cottage, especially Meredith's clothes, and allow me to use the same parameters for the bra clasp and knife. I could take their profiles, look down to the single digit picogram level and use mixed samples with numerous contributors and I'd find their profiles
somewhere in those 200 or so samples. After all, Meredith spent the night there, a trace transfer under those conditions is probable if I look hard and long enough.
But that's damning with faint praise, of course, the British girls had nothing to do with the murder, but in a rational world they'd be (slightly) better suspects for those bound and determined to imprison additional people for Rudy's crime.