Many people are wrong interpreting statistics ...

[Heriberto Janosch]

I have read the following in Internet:

"If a person is missing for more than 48 hours, there is a 90 per cent chance that they won’t be coming back. A lot of the [Scotland Yard review] team will be thinking that she simply isn’t alive."

It is based upon a reasoning like this: "in the 90% of the last 2,000 solved past cases of missing people who were missed for more than 48 hours the person was found dead".

But you can not say from this that Madeleine has a likelihood (see note below) of 90% to be dead.

What you can do, for example, is estimate that in the following future 400 similar cases there will be approximately 360 found dead, and 40 found alive.

But Madeleine has not any likelihood of 90% of being dead. You can not applied statistics to a particular case.

Madeleine is now 100% dead or 100% alive. And given that nobody knows it is worth the searching for her.


Note: I write "likelihood" because "probability" is a mathematical formal science concept which can not be used in a science of facts statement.

Heri.

[C.Edwards]

So 40 in 400 will be found alive (approximately) and that's 10 in 100, yes?  So 90 of those 100 will be found dead. Therefore there's a 90% chance that Madeleine, if found, would be found dead then?

So... ?

[Heriberto Janosch]

I have read the following in Internet:

"If a person is missing for more than 48 hours, there is a 90 per cent chance that they won’t be coming back. A lot of the [Scotland Yard review] team will be thinking that she simply isn’t alive."

It is based upon a reasoning like this: "in the 90% of the last 2,000 solved past cases of missing people who were missed for more than 48 hours the person was found dead".

But you can not say from this that Madeleine has a likelihood (see note below) of 90% to be dead.

What you can do, for example, is estimate that in the following future 400 similar cases there will be approximately 360 found dead, and 40 found alive.

But Madeleine has not any likelihood of 90% of being dead. You can not applied statistics to a particular case.

Madeleine is now 100% dead or 100% alive. And given that nobody knows it is worth the searching for her.


Note: I write "likelihood" because "probability" is a mathematical formal science concept which can not be used in a science of facts statement.

Heri.

C.Edwards: You can not applied previous statistics to a particular case. It is wrong. You have to applied previous statistics to a set of many new cases, and only to estimate ... This error is one of the causes of Criminal Profiling to be a false science.

If 60% of the persons of your age have illness X, have you 60% of illness X?

No. you have 100% of illness X, or 0% of illness X.

Heri.

[C.Edwards]

I'm not applying it to a specific case though. I'm saying that if Madeleine is found in the next 400 "finds" then there's a 90% chance she'll be dead, correct?  She's just potentially 1 in 400 to be found, therefore it's irrelevant to say statistics don't apply.

Otherwise car insurance wouldn't work, would it?  Insurers look at statistical chances of things happening that they'll need to pay out on. And bookmakers.  If I were to place a bet on Madeleine McCann being found alive and well within the next 5 years, would I get odds reflecting a 1 in 10 chance of her being found alive or 50/50 do you think?

[Eleanor]

Statistics are pointless and mean absolutely nothing.  If one child in a thousand is found alive that completely puts paid to all statistics.  Madeleine could be that one in a thousand.
Ninety percent of children are dead, blah blah blah.  So what are the other ten percent?  Should be dead?  How very thoughtless of them to b....r up statistics.

[Heriberto Janosch]

I'm not applying it to a specific case though. I'm saying that if Madeleine is found in the next 400 "finds" then there's a 90% chance she'll be dead, correct?  She's just potentially 1 in 400 to be found, therefore it's irrelevant to say statistics don't apply.

Otherwise car insurance wouldn't work, would it?  Insurers look at statistical chances of things happening that they'll need to pay out on. And bookmakers.  If I were to place a bet on Madeleine McCann being found alive and well within the next 5 years, would I get odds reflecting a 1 in 10 chance of her being found alive or 50/50 do you think?

I repeat:

If 60% of the persons of your age have illness X, have you 60% of illness X?

No. you have 100% of illness X, or 0% of illness X.

You can not applied statistics collected on a aggregate level to a single case level.

Madeleine has not neither "likelihood" nor "chance" of anything. She is alive 100% or dead 100%. Nobody knows (well, the abductor maybe).

Going to the extreme with the example, you can have 2,000 abductor cases with the abductor a man, age between 16 and 35, white. The next case the abductor could be a woman, age 60, black. There is no way of applying "probabilities" here, because crimes are not aleatory events.

[C.Edwards]

OK. You obviously are struggling to understand my point. I will simply refer you to my earlier question.  If I was to go to a bookmaker and ask to place odds on her being found alive and well, would I get odds reflecting a 1 in 10 chance of her being found alive or a 50/50 as you suggest?

[debunker]

The statement
"If a person is missing for more than 48 hours, there is a 90 per cent chance that they won’t be coming back. A lot of the [Scotland Yard review] team will be thinking that she simply isn’t alive."

Does NOT say thar 90% of people will be dead, merely that they won't come back. Includes adults and teens going absent for reason.

[debunker]

OK. You obviously are struggling to understand my point. I will simply refer you to my earlier question.  If I was to go to a bookmaker and ask to place odds on her being found alive and well, would I get odds reflecting a 1 in 10 chance of her being found alive or a 50/50 as you suggest?

It is intellectual density like this that causes people so many problems.

What Herberto says is true; you just do not understand the math.

[Heriberto Janosch]

The statement
"If a person is missing for more than 48 hours, there is a 90 per cent chance that they won’t be coming back. A lot of the [Scotland Yard review] team will be thinking that she simply isn’t alive."

Does NOT say thar 90% of people will be dead, merely that they won't come back. Includes adults and teens going absent for reason.

Yes, it is ambiguous, that is why I rewrote as an example "in the 90% of the last 2,000 solved past cases of missing people who were missed for more than 48 hours the person was found dead", which is that the person wants to say.

[C.Edwards]

Ignoring disputes on mathematics from those that haven't got a clue, what, sort of odds would a professional bookmaker offer on finding Madeleine McCann alive and well after all this time?  Just roughly?

In your opinion, it would have to be 50/50 - therefore evens - to be fair.  I reckon I'd get 100 to 1 against as a bare minimum.

[Joanne]

C.E.Edwards is right, you cannot have it both ways.
She has been missing for more than 48 hours which then puts her in the 'dead' group (as those statistic's go), you then can't say it doesn't apply to her. However, I accept that until she is found, we cannot tell either way but statistically, she's dead. Its like the Mccann's saying "There is no evidence to suggest she wasn't alive when she was kidnapped". Of course there isn't because nobody would kidnap a corpse.

http://www.mccannfiles.com/id260.html
"There's nothing to say she's not out there alive"

By Dr Martin Roberts
27 June 2009
 
Gerry McCann: "She's out there or she's not, and there's nothing to say that she's not out there alive. So it's simple. She's out there until proven otherwise."
 
The expectation appears to be that, in order to prove she's not 'out there' one must establish her alternative whereabouts. That this is extremely unlikely is reflected in Gerry McCann's confidence that the current situation is "indefinite".
 
Unfortunately for Gerry McCann, the situation isn't the straightforward dichotomy he supposes. There is an equally stark alternative reality, summed up as: 'Either an intruder entered 5A on the night of May 3rd, 2007, or Madeleine is dead.'
 
In embellishing his own 'straw man', Gerry McCann has himself unwittingly introduced a patchwork of possibilities. Two contingent factors are involved: 'whereabouts' and 'state of being'. Organizing these into a simple matrix presents four distinct conditions, one of which must define Madeleine's situation post May 3rd, 2007.

OUT THERE
          Yes No
ALIVE Yes   X
          No X   


As there's 'nothing to say she's not out there alive' we can place a cross in the bottom left quadrant. (Both the syntax and the speaker's vocal inflection indicate that the phrase being negated in this statement is 'alive', not 'out there', an appropriate paraphrase being 'she's not alive out there'. This is easily confirmed by substituting alternative, less emotive phrases, e.g. 'she's not at home, asleep').
 
We can confidently place a cross in the upper right quadrant also, as this condition represents Madeleine safely with her parents, which is self-evidently not the case. That leaves two conditions, one of which must apply.
 
Gerry McCann has been hiding behind the obvious difficulty in 'proving the negative'. But it's really not necessary, because one can test the remaining positive condition independently by other means. Simply disproving it alone exposes the remaining condition to be true - body, or no body.
 
The original investigation found no evidence of a break-in. Could an abductor have just walked into the apartment? Consider the following statements:
 
"...that's exactly what I felt like, you know, a few minutes before our world was shattered and probably 3 or 4 minutes before Madeleine was taken."
 
"Part of the reason we ended up coming through the back was the noise coming through the front door."
 
Notice that they 'ended up' coming through the back, apparently. They did not decide to do so 'early on.' So when was this decision arrived at? It must have been after 9.05 p.m., because Gerry McCann used his own front door key on that occasion, according to his original statement to police. The intruder then had just 3 - 4 minutes to enter via the patio and leave with Madeleine via the window, opening the shutters in the process, while Gerry McCann and Jeremy Wilkins were standing outside (if the abductor didn't open them to get in, he must have opened them to get out. Not only did Gerry McCann and Jeremy Wilkins not hear this commotion but Matthew Oldfield did not report the shutters/window open following his visit to the apartment at 9.30).
 
Subsequent sightings of two different men (as described) moving in different directions, and at different times, counteract each other as candidates.
 
It did not happen. It could not have happened. And if no unauthorised person accessed the apartment, then Madeleine was not abducted from it.
 
Thus Madeleine McCann is dead, confirmed by the '100 days' statement, in which Gerry McCann refers to 'the belief that Madeleine was alive when she was taken'. What could possibly have happened for the child to have died within 3 minutes of having been witnessed asleep by her father?

[debunker]

I give up. Not only dopeople misunderstand the math, they can't read either. It is not a 'dead' group, merely a non-return. People go missing for many reasons.

[Heriberto Janosch]

I will try to explain better.

We have two levels of analysis here: the aggregate level (a lot of cases) and the single case level.

We are dealing with no random events like crimes also. So the probability maths theory can not be used here.

You can not scientifically use results from the aggregate level to explain a single case. Or take a single case to explain the aggregate level.

An example. Suppose we know that 15% of Madrid women are pregnant now. So now 85% are not pregnant. Now you meet a woman from Madrid, María. You can not use the statistics to say nothing about María state. She is not 15% pregnant for sure. She is 100% pregnant or 100% not pregnant.

But you can use the statistics to estimate how many maternity hospitals to build and how many maternity health personnel you will need in the next years. Aggregate level (past) to explain aggregate level (future). You can predict more or less how many women will be pregnant the following years, but not who of them.

To know more on María single case and her actual state, you have to study María single case. You can not assign to her "probability", "likelihood", or "chance". It is not scientific.

So as an example if "in the 90% of the last 2,000 solved past cases of missing people who were missed for more than 48 hours the person was found dead", we can not say the Madeleine has 90% probability of being dead. And that is all.

So, you can not say scientifically 50/50, 99/1, 90/10, etc. It is only guess. She is alive or she is not alive now. The single case is studied as a single case. So the way to know if she is alive (or not) is searching for her and investigating her case.

I expect to have been more clear now.

Heri. 

[Redblossom]

Statistics for toddlers are not the same as for teenagers and adults.They return of their own account if tthey just ran away or escape if they didnt.

It is more probable that she is dead than alive, but obviously it is unknown which is true. The dog alerts play a part here too, however much some people want to show that they mean nothing at all vis a vis a possible death.
No one knows.
Likewise, there is no evidence whatsoever that she is alive.

How many three year olds snatched by a stranger have been found alive after years? And how many toddlers and young children have been found dead?
The reality is depressing. Abducted teenagers and young adults taken for sexual predatory reasons are a whole different kettle of fish.They tend to live.