Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - daveK

Pages: [1] 2 3 ... 9
1
As I and Mr Schweitzer agreed to discuss this topic in private correspondence (as he prefers that way), there will be no more posting regarding this subject. Just so you guys know…..

2
-Mr. Richard Schweitzer:

I am the one who posted the Nature Genetic’s article. After I posted it, I read the book “Romanovs” by ROBERT K. MASSIE in which he pointed out that Nature Genetics’ editorial had error. As you suggested, I have no reason to doubt the authenticity of your wife’s family history.    

I would be greatly appreciative if you answer the question someone asked: Did you change your mind for past 10 years? Or do you still think Ms. Anna Anderson was indeed  GD Anastasia? As far as I know, even Peter Kurth or Greg King no longer believe it, although they do not believe that she was a Polish peasant Franziska Schanzkowka.

I will respect anyone's belief, whether it is based on science or religion or pure instinct. The point of my postings was to eradicate disingenuous pseudoscience trick that is prevalent in this board to deceive scientifically lay audiences. I hope you understand it.

DaveK

3
Bear, your conjecture has a point. My numbers were just a guess. And it's up to the jury to decide which conjecture sounds more plausible.

I'll be busy this and next week, so won't be able to check the site for a while. I should come back with more new result of my current research though.

4
I was always curious about Anna Anderson's foreign accent. As far as I know and I heard from people, her English on BBC and  PBS program had Polish and/or German accent, but definitely NOT Russian. Is there anyone who speaks three languages German/Polish/Russian at native level in this AP site? And can they cofirm this??



5
Quote
Since Giller was unable to stop all the clocks then, as most magicians do, there is either a slight of hand, which couldn't occur in homes he couldn't reach, or there was a percentage of clocks that would stop within the range of his show.

Since clocks stop all the time for a variety of reasons, Giller, I assumed, was relying on "coincidence" and using it in his favor.

The reason I mentioned dust, is because dust would have stopped at least 50% or more of the clocks because that is major reason behind clocks stopping in 1981.

Science is often the best way to discover the truth.  The only problem with science, science can only be as good as the data presented.

daveK gave us very little data and since we lacked information such as:  Could Giller stop my clock at exactly 2:00 pm on the the 27 of Feb 2005, the same clock on the 28th, 29th....? Ten times out of ten?   If he can present this kind of information, then we have to deal with a new problem which could  rule out coincidence.

AGRBear 

No, no Bear. That's not the point. I used the Geller's anecdote as AN EXAMPLE to demonstrate how you use the scientific approach to explain a seemingly odd coincidence. Specific detail in Geller's trick itself has no relevance to the Grossmann case.  

6
Quote
When we talk about other reasons over on the other threads, we are constantly told that mtDNA is 99.99999% proof that anything we say is useless data.  
AGRBear


Bear, I specifically said that I admit the importantce of historical research.  I quote:

Quote
I also support the view that there is a possibility that FS was killed by Grossmann and even encourage people to investigate the event from historical perspective.


The whole point in this thread is "Let's not mix Apple and Orange".

Scientific research and Historical research are fundamentally different. If we mix them, we will waste our time forever, without producing any constructive thought, just ending in the bitter feud between us.

7
The Myth and Legends of Survivors / BTK Killer and CODIS system
« on: February 26, 2005, 05:35:52 PM »
This is also a good opportunity to talk about CODIS, as every time I use the term, this jargon confuse people.

This is from a news explaining BTK killer’s DNA sample was searched against CODIS system.

“Landwehr said all DNA samples in the Wegerle case are being sent to the FBI's Combined DNA Indexing System, which is known in law enforcement circles as CODIS. As of November, CODIS contained more than 1.5 million offender profiles, most of which were taken from inmates entering prison.”


--------Wichita, Kan., Police Chief Norman Williams receives an ovation after announcing the capture of a suspect in the BTK serial killings. ---------------



But think this way. Even if you have a extremely rare DNA type (1 in 100,000), if you scan 1.5 millions DNA, you would find 15 people whose DNA matches. right?

How do you solve this problem?

This is why CODIS uses 13 combination of STR. By using 13 combinations of STR, you can distinguish 7.7x10e15 (7.7 quadrillion) Caucasian person!! Even if you search 1.5 million people, the odds you find the random match is one in 5 billions.

So, if they find the BTK killer by CODIS system, even if you had sample size of 1.5 millions, he is likely an owner of sample DNA. Although, it doens’ guarantee he was the killer. He might have been having a sex with victim before BTK killer killed her.

Now, let’s go back to AA case. Some criticizes that the Gill used “ONLY” 5 STR, compared to 13 STR, so his method was not good enough. No, this logic is fallacious. Gill didn’t search a 1.5 million database, or 5,000 middle aged guy in Kansas. He only sequenced ONE person, Anna Anderson. So he didn’t have to use 13 STR. 5 STR was enough for paternal testing, especially as the result was non-matching.



8
The Myth and Legends of Survivors / BTK killer arrested?
« on: February 26, 2005, 04:54:05 PM »
Today, Kansas authorities said that they finally caught BTK killer.

This is a good example to explain the DNA matching issue.

In this case, police has been conducting a massive scale of DNA testing, sequencing almost all the middle aged man in neighborhood.

News doesn’t say if they had mtDNA or nuclear DNA from BTK killer. But as their DNA samples are from 70’s, it could be mtDNA.

If this mtDNA type is as rare as 1 in 9000, you would expect you find one match after testing 9000, statistically speaking. So we have to be careful.

But, boy, he looks like just a regular guy. My boss looks like him.



9
The Myth and Legends of Survivors / About raw data
« on: February 26, 2005, 04:32:28 PM »
I understand there are confusing parts about the DNA testing. I am now preparing all-new article to entangle all the confusions. Many people complained that my previous one was way too complicated and not easy to read, so the next one should be more accessible to everyone. I will also address Michael GD case as I mentioned before. Stay tuned.

Also, I do not expect everyone to read all the raw data I post in this thread. The point is that there are many audiences at AP site, and each has his/her own expertise. If I expose many data, there would be likely that at least one person who knows the subject would notice my mistake (or any disinformation). So, if you are not interested in the data, just skip it and they can read my “interpretation”.

10
Quote

I wonder if there is a way we can come up with a concrete statistical number of the probability of this happening (a random mtDNA match in a situation like AA/FS)? I am sure this number would be very very low (probability extremely low), but I still would like to see what it is. This way it would be easier for everyone to understand how unlilkely it is.

Denise, maybe you can ask your husband (I wasn't kidding before ;)). I wish I was better at stats myself! :-/


No, no Helen.
The prolem is NOT STATISTICAL one. Statistics is extremely easy. One mtDNA type was found in 9000 samples. That's it.
What must be confirmed is that "How do you know that one mtDNA type in 9000? Why 9000? Are you sure you didn't make a mistake?". This question is related to population genetics and bioinformatics, but not statistical one.

How did I know? People are underestimating my obsesseive efforts I put into my investigation. I listed the list of literatures before. I checked all the articles, page by page, table by table, to see if there is FS/AA's mtDNA type. I also went to FBI's US caucasian mtDNA database, and NIH's gene bank. However, some part of them overlaps, and it is difficult to estimate how large my sample pool is. My number 9000, is a conservative estimate. There are probably more than 2 million mtDNA sequences in US and Europe, but I don't know how many are actually open to public, or how many are put into online database. I was able to manually count about 9000 samples from my papers. That's my best estimate.

Am I 100% sure that I didn't miss AA/FS's mtDNA? I did my best, but I admit that I am not 100% sure. Human makes a mistake. That's why I listed the list of literature (list contained only part of entire list though) so that other people (if any) can start their own investigation to see if I missed any AA/FS's mtDNA.

11
The Myth and Legends of Survivors / To Bear's question
« on: February 26, 2005, 04:06:25 PM »
Quote


AA isn't Polish?
AA isn't Russian?
AA is an Icelander?

No, I couldn't be reading this right.

Oh dear....  Please, explain.

AGRBear
.


OK, this part might have caused confusion.
AA's mtDNA are rare, probably around 1/9000.
But it's EVERYWHERE in Europe.
It has been mixed up for past 15,000 to 40,000 years.
It was found that one person had AA's mtDNA type in Iceland, but it doesn't mean anything. It could be found in anywhere.

12
Anastasia le jour de son mariage!  Je ne puis pas attendre pour la voir.  Beaucoup sont en désaccord, mais je pense qu'elle est la plus jolie parmi des soeurs.  Je m'inquičte cependant.  Ce qui si elle ressemble ŕ Anna Anderson...

13
Thanks for all the response to the quiz. Some of them are better than my analysis here. The whole point was that I wanted people to think by themself. Now here is my answer. Bear, you may have misunderstood my answer. I was challenging you to see if you reach the same answer as mine. Because the answer is that it was by coincidence. But how do you know? is the next question.


----Warning-----
Compared to my other argument, this posting contains some ambiguous conjecture, and I admit that. I also support the view that there is a possibility that FS was killed by Grossmann and even encourage people to investigate the event from historical perspective.

However, the purpose of this posting is to reconcile the two contradictory hypothesis, “AA=FS” and “Grossman killed FS” by applying a rational approach. (Remember, two hypothesis are mutually exclusive. If one event has 99% chance, another has 1%; This is called “addition law of probability”, expressed as P(X or Y) = P(X) + P(Y).)

[size=18]------ANSWER to the URI GELLER QUIZ--------[/size]


This trick was one of the favorite Geller’s tricks. He repeated in the TV in many countries. This case was actually listed in a statistic textbook. Here is my calculation.
Uri Geller was very popular during 70’s and this program in which Geller showed the trick had large audience. If I remember correctly, the viewers (not household) was something like 50,000,000. Household is probably something like 20,000,000.

Denise guessed 1 million. But for example, Nielsen reported that last week of CSI show was watched by 30,000,000. In 70’s, there was no cable TV, no internet, there was only a few broadcasting program. 50,000,000 for a popular program was not uncommon.

Now, each household had multiple clocks and watches. Let’s say, on average, each household had three of them that were battery-operated.

Also, during 70’s, people didn’t use Alkaline batteries with “bunny inside”. Suppose that each clock or watch’s battery died in a year.

Now, let’s calculate the chance that each household would observe the clock/watch dies for a certain 30 minutes.

Numbers of 30min in 1 year: 60min x 24h x 365 h x 1/30 = 17520 (As Denise calculated)
20,000,000 x 3 = 60,000,000 (numbers of clocks/watches)

So the number of clocks/watches dies:
60,000,000 x 1/17520 = 3424

3424 watches and clocks dies during his meditation.

Of course, not all of them actually calls. In fact, TV station received only a few hundreds call, but it still reported as “many audiences”.

My point is that human innate perceptions can be easily deceived unless you rely on the solid scientific approaches. This Geller case is the typical fallacy of so called “Law of very large numbers”. If you have a large number of input, you will have a coincidence. Just as a common sense says.

Now,  let’s apply some rational approach to the Grossmann case. But first, I have to explain the difference between prospective and retrospective analysis (Denise, let me know if your husband find a mistake).

[size=18]
---------PROSPECTIVE and RETROSPECTIVE Statistics------------[/size]
Many people does not consciously take this into account. AA’s mtDNA matched to FS’s relative, NOT AFTER they sequenced hundreds of AA candidates or hundreads of FS’s candidate’s family. They had only two people’s sample. Anna Anderson and C Maucher. And they matched. This fact, they had a “small window” of two person’s sample is very important. This is so called prospective approach, as they did not know the answer before the test.

In contrast, why are we paying our attention to the name “Saznovsky” contained in the Grossman’s diary, which is not even Schanzkowsky? Grossman was not arrested because of the detective who investigated F Schanzkowky murder case. He was arrested because of other case. We already knew the name Schanzkowsky, and look for any name “sounds like” Schanzkowsky in murder case in Berlin region in those era. This is so called "retrospective analysis", because we knew the answer (the name Schanzkowsky) before our test. Comaparing two results by two different approachs trick you to think they are "contradictory", but they are actually not.

[size=18]--------GROSSMANN CASE--------------[/size]
Now let’s analyze Grossmann case in a way I did with Uri Geller case.

Let’s say there are 10,000 last names in Berlin region. If you group them into “sounds-like” name subgroup, my estimate would be 300 such group. Many Polish names has “sky” at the end of last name, so if you compare “Saznov” and “Schanzkow”, they don’t sound THAT similar. If you apply same degree of stringency to the similarity test, I would say at least 30 names can be grouped into one name group.

Next, how many names were mentioned or written in the testimonies or written evidences in potential murderer case that occurred during1920-22 in Berlin region. Remember, we are not talking about actual murder, but “potential” murder, as the body was never recovered for Grossman-Saznovsky case. Berlin was one of the biggest city and my low estimate say around 2000 names. Using this rationale, [size=16]I would be surprised if you don’t find any name “sounds like” Schanzkowsky.[/size] [/color]

Then, why was it found in such a well-known case as the Grossmann case? It’s the same reason why a lottery winner is usually a person who buys multiple lottery tickets. Buying multiple lottery tickets increase the odds. Serial killers who killed multiple victims also increase the odds. Also, a well known case has been under a lot of scrutiny compared to unknown case.  

And remember, even if it was the correct name Schanzkowsky, first name could  be different among hundreds other name. Even if first name and last name is same, it could be a person who has same name. For example, if I enter the name “Peter Kurth” at ancestry.com, it finds 6495 matches in US census records. And even if the diary was true, it doesn’t necessary guarantee that he actually killed FS.

Considering all this, I would say the probability of FS was killed by Grossmann was around 0.0001%.

The number i used here was pure estimate. I may be wrong, and I admit that there is a chance that she was killed. But as an amateur forensic scientist, I had to reconcile two mutually exclusive events by rational scientific approach.

Again, this is my conjecture, an educated guess. If you don’t find it plausible, that’s perfectly fine with me. So don’t attack me on this.


14
Quote

It's actually both Polish and Russian, which is an important point because the conjecture is that AA was Russian and not Polish.
Perfect chart to show your husband, Denise!  ;)


No Helen, Russian/Polish doesn’t matter. [size=16]mtDNA group was generated for past 50,000 years, and they have been mixed [/size]to the degree you don't observer these differences between neighboring countries.

In addition, my point was that there is no AA mtDNA in any population in the all world (except one case in Iceland).

This is called population structure, which is the absence of random mating within a population leading to allele frequency differences among subpopulations. It does exist between Asian and European population. But it is statistically negligible between Polish and Russian, or German and French.

15
As you see, there IS a common mtDNA haplotype. A certain type of CRS is shared by many Polish.

Some people argue that because of this fact, AA/FS's mtDNA may be prevalent.

However this is WRONG. The fact that CRS is prevalent is nothing to with AA/FS's rare mtDNA type, as I showed in above figure.



Pages: [1] 2 3 ... 9