Some interesting discussion over at Anna Raccoon today:

http://annaraccoon.com/2014/06/01/a-matter-of-f-a-c-t-witch-hunts-old-and-new-part-2/

‘Gildas The Monk’ wrote an interesting post entitled: “*A Matter of F.A.C.T – Witch Hunts Old and New Part 2″*

The lady ‘annaraccoon’ has had a great interest in the Savile affair because she just happened to have been living in one of the institutions which Savile was said to have conducted his nefarious behaviour in respect of young girls. Essentially, she says that Savile was not in attendance at that place when he was said to be, according to the accusations. Well …. Something like that.

We must remember that the accusations against Savile do not rely upon actual evidence of wrong-doing. They rely upon the multiplicity of accusations. That is the case also of the accusations against Michael LaVelle (?), Dave Lee Travis and others. None of the individual accusations could be substantiated, and they were thrown out in court. For some reason that I do not understand, Stuart Hall caved in. Perhaps he did a deal, or perhaps he is senile. It does not matter which.

Gildas extrapolated from all that stuff to various similar ‘witch-hunts’ which have occurred in the past, specifically, the ‘satanic ritual abuse’ trials and the persecution/prosecution of Sally Clark, the solicitor, who was accused of killing her children. She was initially found guilty and imprisoned. Only several years later, on appeal, were the statistical claims about the unlikelihood of the cot deaths of two children in a family (millions to one against) found to be false. Sally Clark was released, but the damage had been done. She sank into the depths of despond and died some eight years later an alcoholic. Was she a disturbed character? She may have been. Did she kill her children? She may have done. Was she treated as a witch? DEFINITELY.

The witch-hunter in that case was a guy called Dr Meadow. It was he who claimed that the likelihood of two children dying from cot death in the same family was practically zero. It took the evidence of ‘proper’ statisticians to show that his calculations were false.

Here is a simple example.

If you throw a dice, what are the chances of a 6 appearing? Clearly, it is 5 to 1. If you throw the dice again, what are the chances of a 6 appearing? Again, the chances are 5 to 1. What you CANNOT do is, somehow, multiple the chances on the second throw by the chances on the first throw. Thus, it is quite possible for Sally Clark’s children to have both suffered cot death, and the chances of the second death were just the same as the first death. Dr Meadows multiplied the chances.

======

This ‘witch-hunt’ attitude is important to us smokers. it is reasonable to ask the question, “If a court of law, seriously considering the guilt of a woman who has been accused of killing her children, and bound by the need for ‘beyond reasonable doubt’ evidence can be fooled by an ‘expert’, how much more easily can politicians be fooled by the same sort of evidence?” Remember that politicians do not have ‘beyond reasonable doubt’ as a guiding principle, or indeed, the civil case ‘guiding principle’ of “On the balance of probabilities”. They need no evidence AT ALL! They can act and decide purely emotionally. Or they can decide to act merely upon what they think that the electorate might like, and thus increase their popularity.

Is there anything more likely to lead to disaster? The only thing that militates against disaster is that “life goes on”; farmers still produce food and shops still sell it; houses do not collapse; thankfully, the electricity supply and the water supply survive, no thanks to politicians. Thankfully, experts and politicians have real difficulties in buggering things up, try as they might.

It is all-right generalising in this way, but it is not sufficient, since it presumes that the deficiencies are inevitable. That is not the case. In the UK, we have a CABINET system of government. Cameron is not KING. When the smoking ban was proposed by ‘the experts’, it was THE CABINET’S duty to consider the wider implications. They should have rejected it and sacked the experts forthwith. Thus, the consequences of ‘experts’ giving bad advice would have been clear for all to see. That might have given the ‘global warming’ gang of academics food for thought.

The BBC programme about PP is precisely a case in point. The evidence of efficacy is non-existent. ‘Experts’ who suggest that it is a good idea should be shamed and sacked.

It is the only way.

02/06/2014 at 06:32

I rarely comment – I’ve not much to say, nor to disagree with – but I thought the “odds” thing better explained in Gildas the Monk’s piece:-

If you roll, a dice the probability that you will get a “6” is 1 in 6.

If you roll the dice a second time, the probability that you will get a “6” is…the same. 1 in 6.

But if you roll the dice several times and keep getting a 6, the reasonable indications are that the there is some anomalous force which is excluding the other numbers. But you can’t say how it has been fixed; it could be a magnet, a weight or sleight of hand, the decree of the Sky Pixie or an alien ray, or anything else.

02/06/2014 at 17:08

Mathematical terms can be confusing. It is indeed true that, since there are six sides to a dice, the chances of one side being revealed are ‘one out of the six’. But the ratio of six to other numbers is one to five, That is, one 6 to five other numbers. The importance of this ratio arises when you consider relationships. For example, If you have six people, how many relationships can the first person have with the others? The answer is five. Another example is to consider spaces between objects. Given six objects in a line, how many spaces are there between the objects? The answer is five. But it the objects are in a circle, then an additional space appears between no. 6 and no. 1! In which case there are six spaces.

These considerations matter when the effect of mutual gravitational attractions occur between objects in space.

But for the purposes of this post, these considerations are not important. What is important is that, unlike mutual gravitational attractions, you cannot multiply simple odds, which is what Meadows did.

Perhaps I should have included the rest of the statement which you quoted, but it was not important regarding the fault of Meadows.

02/06/2014 at 17:00

Then there is ‘Conditional Probability’ or some such thing.

If you have two dice and throw a 6 with the first one,what is the probablity that you will also throw a 6 with the second one?

(Snake-eyes)

In this instance you must multiply.

1/6 X 1/6 = 1/36

There is a 1 in 6 chance that the 2nd dice could be a 6; but, only a 1 in 36 probability that it will. 🙂

02/06/2014 at 17:40

Quite so. An easy way to understand that is to count the number of possibilities. You could write it thus:

1+1.

1+2.

1+3.

1+4.

1+5.

1+6.

2+2.

2+3.

2+4.

2+5.

2+6.

3+3.

3+4.

3+5.

4+6.

5+5.

5+6.

6+6.

—–

Total + 18.

========

You then need to count the possibilities of the other dice, because you have only counted the possibilities of dice one. HOWEVER (and this is where ‘spaces between objects’ come in), you cannot count ‘1+1’ twice. If you throw the two dice together, there is only one possibility of both dice showing ‘1+1’, which means that you have to take out also 2+2, 3+3, etc, which means that the real number of possibilities is 30 and not 36.HOWEVER (again) the statistical possibilities of throwing any random combination are still 36!

02/06/2014 at 20:47

I think the key to the Sally Clark case is an abuse of probability by Meadows (a medic natch).

An example of this fallacious argument:

There is about a 1 in 14 million chance of buying a winning lottery ticket. But people do regularly win. So how did they overcome the impossible odds? Surely they must have cheated.

Once a result has occurred, the event is certain. Not unlikely.

See also http://en.wikipedia.org/wiki/Prosecutor%27s_fallacy

03/06/2014 at 00:11

I liked the example given in the section on ‘conditional probability’. I quote:

“

Berkson’s paradox – mistaking conditional probability for unconditional – led to several wrongful convictions of British mothers, accused of murdering two of their children in infancy, where the primary evidence against them was the statistical improbability of two children dying accidentally in the same household (under “Meadow’s law”). Though multiple accidental (SIDS) deaths are rare, so are multiple murders; with only the facts of the deaths as evidence, it is the ratio of these (prior) improbabilities that gives the correct “posterior probability” of murder.”In plain language, the improbability of two ‘accidental’ deaths in a household is matched by the improbability of two ‘murder’ deaths in a household.

The ‘anterior’ probabilities dictate only the rarity. When an event happens, such as two accidental (SIDS?) deaths, then it is the ‘posterior’ probabilities which come into play. Have there been other incidents where two (or more) children have been found dead with a separation of time between the deaths. How many of them were conclusively proven to be murder? How many were definitely accidental? How many were uncertain? It is those ratios which are important, after the deaths have occurred.

03/06/2014 at 00:17

Just as a point of interest concerning lottery techniques. When I choose my lottery numbers I always select the numbers that I feel are most likely to fail. Up to now I have never been wrong – does this mean that I have a unique system?

03/06/2014 at 00:26

LoL. But your system is not foolproof. The only foolproof system is to fail to buy a ticket. I have never won a penny on the lottery ……

03/06/2014 at 02:11

Not being mathematical, I shall have to consult the spirit world for better odds.