Pawn tickets in Mitre Square

[Pierre]

QUOTE=Elamarna;388995

Pierre

Interesting conversation with Jeff I see.

However you have not attempted to address the issue of do different words, especially "Anne" rather than "Jane" and the use of "st" rather than "street" make a difference to your anagram solution.

Hi Steve,

The total amount of the letters at hand is the important issue and I will explain why. We have no hypothesis for a serial killer trying to find names or street names wich would give an exact amount of letter for his name (to make it more easy), nor do we have an hypothesis for a serial killer trying to find street names with a lot of redundant letters (to make it more difficult).

The examples I listed show just how many names can be found which can be linked, indeed that list is far from exhaustive!

Havenīt seen this, will read it.

Yes, and that is one dimension of the problem. From a set of letters you can construct a lot of names. So what is the chance that a particular name, one from your list for example, will be in that set?

The point is that if you do not use all the letters, in this case 42, it is possible to find many links, if you had found a name using all 42 it would be very suggestive of a genuine anagram.

However given that you have admitted not all 42 are used, the probability must be that any name found is purely coincidence.

I understand your thinking. You are looking for the perfect match. But the problem is that we do not know if you are thinking like a serial killer in 1888 when you say that.

Another very important issue is the characters of the different names that could be deduced from the letters. Some names are common, some names are short and leaves many letters unused, some names are long and uses nearly all the letters and some names are rare.

Depending on the character of the name, how many names it contains and so on, how many given names and how long they are, some names will be unlikely to be found in the set of letters anyway.

All I would like to know is the answer to the question David posed?

If you could be as kind as to repeat that question, I will see if it is meaningful to answer it.

Would the changes that are possible to the letters on the pawn tickets, surely you cannot argue they are not possible, make the finding of the name you propose possible or not?
Yes or No?

Excuse me but what does this sentence mean?

I am myself already certain of the answer to that question, however I am interested in your thought processes, and why you sometimes find it difficult to give clear answers, when doing so gives nothing away with regards to your view on naming a person, unless proven.

Yes, the thought processes are interesting, both yours and mine here, especially since you manage to contribute to it in a scientific way, and this is a contribution to the case.

Regards, Pierre

[Elamarna]

Now come Pierre

The question was clear

If and I say if the wording on the tickets was not as you have suggested and used, does the possible loss of the j and the possible loss of some duplicate letters affect your answer?
Can you still get the name in that case?

The rest of your post is interesting and does indeed tell me much about your thinking.

Yes it does matter if names are common or not I agree. You will see the list I came up with includes both common and uncommon. Indeed at least one very uncommon name is used.

I factored such in when producing the list to make that very point

Cheers

Steve

[Pierre]

Now come Pierre

The question was clear

If and I say if the wording on the tickets was not as you have suggested and used, does the possible loss of the j and the possible loss of some duplicate letters affect your answer?
Can you still get the name in that case?

The rest of your post is interesting and does indeed tell me much about your thinking.

Yes it does matter if names are common or not I agree. You will see the list I came up with includes both common and uncommon. Indeed at least one very uncommon name is used.

I factored such in when producing the list to make that very point

Cheers

Steve

I see what you mean, yes you would still get that name.

Me not being a native English speaker, which name in the list of yours is the unusual one according to you?

Cheers

Pierre

[Elamarna]

The second name of Drake is highly unusual.
In addition Llewellyn while a well known Welsh name was not that common; yet I have 2 occasions of it.
Finally Nott-Bower is not common. Double barrelled names rarely are.

Hope that helps.

From a the view of how many letters are used we have an interesting puzzle do we not.

We have 17 different letters used in your version. In the alternative it is 16.
Several of those are not used at all and some of the duplicates may be used a limited number of times.
This gives us a minium number of letters( if no duplicate is used more than once, which it must be said seems unlikely) used to reach the name of either 15 or 14 and as you rightly say it does not have an outcome on the result it is therefore actually not significant which version is used.

At the top end we can postulate 39 as the top figure. Two not used at all ( this assumes those are not duplicates, if the letters not used were say e and l we would have a top figure of 29 letters which may be further reduced by other not used duplicates) and at least 1 duplicate left unused.
Obviously the actual figure is somewhere between 15 and 39.

I hope that helps anyone trying to work this out.

Steve

[Mayerling]

[QUOTE=Pierre;388991]

Hi Jeff,

palimpsest is the word for the type of reused manuscript ("pali", Greek word for again).

You are confusing academic history with the juridical system when you write:

"...not the best evidence of a court of law".

No one will ever take Jack the Ripper to court. It is now a case for history. What we have is sources from the past. They are always problematic, even if they are understood to be "originals".

We go from finding sources, to hypothesizing, to interpreting, to establishing historical facts, and back to dismissing hypotheses, making new interpretations and establishing new historical facts.

That is the common process in all research and it is not a matter of evidence of a court of law, but a matter of scientific evidence.

We must have a laid back and relaxed approach to this fact.

But we must have a scientifically rigorous approach when doing history.

Donīt you agree with all of this?

Best wishes, Pierre

Hi Pierre,

By the way, thanks for correcting me on the word "palimpsest". I was totally at sea about how to spell it.

You are right that, short of some time machne approach in the the world of never-never, no one will ever take the Ripper to court. And I actually agree on the laid back approach (how else to do it, really!). But if (as you intend) to fully prove the identity of the Ripper, it has to so pass muster that there can be no question about it - so that, despite the laid back approach, you'd have to approach the idea of proof of guilt "beyond a reasonable doubt" which is a legal concept. You can't escape legal concepts of proof in this matter.

So one is still stuck with this version of the issue of the "best evidence" rule in law - and the need for something to look at that shows writing, be it the original, the first copy, or all the way down to a photographic record.

Jeff

[Mayerling]

Pierre

Interesting conversation with Jeff I see.

Steve

Hi Steve,

Yeah it was interesting - but I feel it should have been in "Pub Talk" thread or a similar one. I really wanted to get into the matter of the use of photographs more than anything else, but fell into thinking of writing, and lost writings and records. The result may have been interesting (and certainly I'm gratified in finding it so), but it was somewhat side-tripping the matter of the thread here - which is the pawn tickets.

I may though reopen the issue of lost writings in another section.

Jeff

[jerryd]

Hi Jerry,

I'm not sure this suggestion has any legs. David raised some good points about the probability/coincidence of using the particular pawn shop, especially if the Emily Burel found by Paddy in post #106 is the same Emily Birrell.
The big problem, for me, is that Polly was murdered in the early hours of 31st Aug long before any pawn shop would have been open, so the ticket couldn't have been taken from her body.

Thanks Joshua,

This is true, too. So much for that.

[David Orsam]

Hi David,

Here is a tutorial for you.

http://www.youtube.com/watch?v=rWFDXt-MlNs

You can also go to your library or preferably to the universtity library and read the stats books. All of them include hypothesis testing.

A significance level of 20 means that if you have a random sample and find a correlation in the sample, the correlation is due to chance in 20 percent of all cases.

If you want to use hypothesis testing in history you need data sources. What we have is specific type of historical material. It is not suitable for pure hypothesis tests. But the thinking is similar. Small number of "hits": low probability that the result is due to chance. High number, high probability that the result is due to chance.

I'm not after tutorials Pierre. Nor do I need you to explain what probability is. What I am asking for is the justification of this sentence:

"If you do traditional hypothesis tests for correlations you want to have a significance level of 5 or 10 at the most."

I also want to know the actual significance level required for your experiment. Because 5 percent is not the same as 10 percent.

What if your result is 6 percent of positive hits? Would that be a result of pure chance with no correlation?

What about 8 percent? or 9 percent?

And what about 5 percent?

Then I want to know the reason for your mention of 20 percent? Why not 15 percent? Or 30 percent?

We all know that 10 percent is more than 5 percent and 20 percent is more than 10 percent. Do you think we are all stupid?

What I want to know is the scientific basis for your exercise and how we judge the results.

Telling me to go to a library and read stats books clearly means that you are unable to provide any kind of source for what you are saying and the fact that you have linked to that particular You Tube video reveals to me that you have no grasp of statistical analysis whatsoever.

[David Orsam]

Methodology

Pierre, I do hope you are aware that if this is the question:

"I am asking what the probability is that you will find a postulated serial killerīs name in a mustard tin on a murder site."

then you need to consider whether any other postulated serial killers, in addition to your postulated serial killer, can be found within the 42 characters you have identified from the pawn ticket.

I have already found:

1. James Kelly (who escaped from Broadmoor in January 1888)

and

2. Joe Barnett

Both of these men are postulated serial killers (for the JTR killings). So, in addition to your postulated serial killer, in one supposedly random group of characters in a mustard tin, we find at least three postulated serial killers for the same set of serial killings! Wow, what are the chances of that do you think?

And, furthermore, going on the basis of the C5, what are the chances of the name of the fifth victim's boyfriend being found in a mustard tin of the fourth victim? Astronomical surely!

But I just wanted to point out to you that if you are going to attempt to produce 100 sets of random names and addresses (and I'd love to know how you propose to do that) then for each single group of 2 names and addresses you will need to search for the name of every single postulated serial killer in existence before you can accept or reject your hypothesis.

If we take the concept of the postulated serial killer literally, to include any set of serial killings, then it wouldn't surprise me if you managed to find one in every set which would produce a wonderful 100% result

[Pierre]

[QUOTE=David Orsam;389026]

I'm not after tutorials Pierre. Nor do I need you to explain what probability is. What I am asking for is the justification of this sentence:

"If you do traditional hypothesis tests for correlations you want to have a significance level of 5 or 10 at the most."

I also want to know the actual significance level required for your experiment. Because 5 percent is not the same as 10 percent.

What if your result is 6 percent of positive hits? Would that be a result of pure chance with no correlation?

What about 8 percent? or 9 percent?

And what about 5 percent?

Yes. "What is a heap" and "Where starts the rich and ends the poor?"

Here you can learn about the philosophical problems you are trying to deal with on a low level by asking those questions:

http://plato.stanford.edu/entries/sorites-paradox/

For a sociologist or an historian such questions are not relevant. We always discuss the problems of definitions, the advantages and limitations. It is standard procedure.

Then I want to know the reason for your mention of 20 percent? Why not 15 percent? Or 30 percent?

We all know that 10 percent is more than 5 percent and 20 percent is more than 10 percent. Do you think we are all stupid?

Read your last question, David. It is very rude. Do not put words into my mouth, not even in the form of questions.

What I want to know is the scientific basis for your exercise and how we judge the results.

Sure, David. The scientific basis and the judging of results have been available in sociology and history at universities for decades.

Telling me to go to a library and read stats books clearly means that you are unable to provide any kind of source for what you are saying and the fact that you have linked to that particular You Tube video reveals to me that you have no grasp of statistical analysis whatsoever.

Yes, David. I am unable. And I have no grasp. It is OK: Everyone can see that you are using belittling and rude comments. Since you have no understanding of science, such comments are all you have.

The tutorial is a good source for you. Try to learn from it!

[David Orsam]

Yes. "What is a heap" and "Where starts the rich and ends the poor?"

Here you can learn about the philosophical problems you are trying to deal with on a low level by asking those questions:

http://plato.stanford.edu/entries/sorites-paradox/

For a sociologist or an historian such questions are not relevant.

I thought we were looking at the problem from the perspective of a statistician.

If you can ask "what is a heap?" and "where starts the rich and ends the poor?" you clearly don't understand the concept of statistical significance. That's what you are supposedly trying to establish, i.e. whether you can conduct an experiment which produces a statistically significant result.

If you do your experiment and we end up asking ourselves philosophical questions about whether you have achieved anything then you will have achieved...well...precisely nothing.

[David Orsam]

Read your last question, David. It is very rude. Do not put words into my mouth, not even in the form of questions.

Pierre you need to be aware that I will ask you whatever questions I feel it is appropriate to ask.

Your explanation of what 20 percent means led me to believe it was appropriate to ask if you think we are all stupid.

From your response, or rather your non-response, it looks to me like that is exactly what you must think.

[David Orsam]

Sure, David. The scientific basis and the judging of results have been available in sociology and history at universities for decades.

Vague and non-specific answers like this support my conclusion that you think we are all stupid and that you think you can get away without providing a single source for what you have said on this subject by making generalised and meaningless statements such as the above.

[David Orsam]

The big problem, for me, is that Polly was murdered in the early hours of 31st Aug long before any pawn shop would have been open, so the ticket couldn't have been taken from her body.

Are you planning on responding to Joshua's post by the way Pierre?

What he says seems pretty conclusive to me. You?

[Pierre]

Are you planning on responding to Joshua's post by the way Pierre?

What he says seems pretty conclusive to me. You?

Jolly good! Let us start with "Characterizing the posts of David", shall we? This is

CPD 1:

Trying to control other peopleīs answers in the forum.