Saturday, October 31, 2009

Raymond Jessop's probability of Paternity.

I smell yet another basis for appeal, again, assuming there is a conviction. And Brooke? Bravo again.
From the Plural Life:
"Prior probability assigns a 50 percent likelihood to the calculation. Hudson argued that biased the outcome, because there was an assumption from the get-go that there was a one in two chance that Jessop was the father and that violates the presumption of innocence in a criminal trial.

'Before you ever did the testing the alleged father was predicted to be the father,' he said to Smuts.

'Hypothetically,' she said.

He asked Smuts to explain the scientific theory that supports use of that 50 percent probability and she said: 'It is factored into the equation.'

And then: 'It is assuming the alleged father, the tested man, is either the biological father or he is not the biological father.'

Hudson argued that if a different probability — a lower probability — was used, the outcome might look quite different. He even handed Smuts Jerry Goldstein’s IPhone so she could use its built-in calculator to try it.

But Tanner objected and Walther agreed, shutting down the experiment."
The trial is proving to be intellectually entertaining. I don't mean in a non serious way, perhaps I should say instead of entertaining, it is thought provoking or stimulating.

If I understand it correctly the assumption of the state in claiming 99.99998 paternity certainty for Raymond Jessop is based on a 50/50 chance to start with, that he was the father. Frankly I would have guessed something higher, but the defense is entirely correct to say "we don't know" and in not knowing a different starting number should be used to determine Raymond Jessop's paternity. I'm sure Walther knows this, but she shut down the attempt to graphically prove the fallacy of Amy Smuts' number.

How this works. As a criminal proceeding goes, we must agree on a number of possible consorts "Mrs" Jessop had. If you assume all the fertile males she came in contact with around about the time of conception, just limited to FLDS men on the ranch or frequenting the ranch on or around the date of her child's conception, the number might be 1% or 1/10th of 1% to start with. This would assume between 100-1000 available males, with which to consort. Not knowing the entire equation I can't reliably say what that does to the outcome, but perhaps it makes Raymond Jessop a 10% probable father, as opposed to his neighbor, who would be a 1% probable father. The state is already claiming Raymond's wife was "reassigned" so it may very well be a given that she has had more than one sex partner. Has she had more? Personally I doubt it and I am not trying to besmirch her, but the state has to assume that someone who has not been sexually exclusive might continue to be "non exclusive" so on the face of it, the child's father is probably "non hispanic," "non African," "non Asian." You get the picture.

Raymond's paternity is not certainty. Not even if the manipulated formula came up with a 49.9999% probability would it be certainty or even 74.9999% and this is what the defense and the prosecution are arguing, and Walther did not just "side" with the prosecution in terms of allowing the testimony, Walther sided with the prosecution in a partisan way. She didn't Amy to use the calculator, even when the jury was not present.

There is a reason for this. You don't allow that testimony because you know what the outcome will be. Amy's number would drop precipitously and her "certainty" would become a "best guess." The Jury went HOME over the weekend. Don't kid yourself, this biased blow by blow description of Amy's credibility will get into the jury pool. They are not sequestered. Walter did not allow Amy to use the calculator precisely because it's a "if it doesn't fit, you must acquit" moment.

Instead Walther has chosen to make it a contest in court. That allows what may very well be a stacked jury to decide they "believe" one "expert" more than another and it becomes a contest of "personalities" as opposed to facts.

Face it, a Raymond Jessop that the PROSECUTION acknowledges to have a 50% probability of paternity, shouldn't have even been brought to trial. I'd like to see what those adjusted numbers turn out to be.

Sphere: Related Content


Ron in Houston said...

Um, see prior probability is the starting point of the calculation. The state expert is actually cutting Jessop a lot of slack by putting it at 50%. Considering the other records they have and that FLDS women aren't running around having sex with other men, the state expert should have started with a prior probability of something like 80%.

Either way, if you start out with 99.999% with a 50% prior probability and lower it to a .02% prior probability you've only lowered the likelihood he is the father to 99.5%.

The Pharisee said...

I'm not familiar with the equation so I can't state the outcome.

The indication is RON, that Walther didn't want to know what the outcome was or wasn't sure Amy would perform well under that sort of pressure and she shielded the witness which wasn't in front of the jury from looking stupid, and from having to answer the question.

You should know I got "A's" in calculus but that was many years ago and I'd have to tune way up and see all the work to do it myself today, but I want to know what happens when you change that input.

Unless you can do the math, then I'm not accepting your answer of a ".02%" change in probability.

Ron in Houston said...

Well don't accept the math if you don't want to. It doesn't change the fact that as only a part of the calculation changing the prior probability doesn't make huge changes in the probability of paternity.

As to Walther "shielding" the witness, once again you don't know one damn thing about how this works. This was the defense's suppression hearing. They'll get a chance to ask these questions in front of the jury.

However, by asking these questions in the suppression hearing, they've tipped their hand and the witness will be more prepared at trial.

The Pharisee said...

You haven't shown me any math, just like President Clinton frequently claimed he wasn't going to "Apologize" because "he had already" you are now claiming to have done Math.

It's a spread sheet Ron. It depends on how big the input number affects the outcome, it's not additive.

By the way, I know how POLITICS works. Walther is political. What harm is there in giving Smuts a Calculator? Hmmmm?

The jury wasn't there Ron.

Why not watch her flop or succeed?

You tell me what your qualification is as a mathematician or statistician or lawyer, and I might listen to you. You don't and your next "I'm right you're wrong and a stoopy poopy head" comment won't get published.

Now go back to your blogs no one reads.

The Pharisee said...

Here is the evidence for Ron's contention:

At first blush, it seems credible.

I didn't publish his remark, because it contained claims of lies I told and false hopes I raise that, well, I didn't tell and didn't raise.

I love lawyers. The biggest evidence that Ron is a lawyer is when he tells you what you said, back to you in his words, it isn't what you said.

Nevertheless if this site and this math formula is correct and the conclusions also, then the additive effect of assuming a much lower probability of paternity at the outset, wouldn't be that great.

However, there is no accuser as the site mentions. "Mrs" Jessop isn't claiming the child is fathered by Raymond Jessop, and if it were up to her, she probably wouldn't make that assertion. She is also, an adult, not a child.

There is no plaintiff claiming paternity with which to make the "50%" assumption and even then a malicious claimant (as the state has to be viewed) isn't giving an unbiased view.

I still maintain Walther shielded the witness and as much as she could, the press, from a failing and flopping "Expert" witness. Personally, I'd have been curious to see what Smuts would have done with the calculator.

In that the State took large numbers of DNA samples, and in that the FLDS are closely related from a genetic standpoint, I'd like to see test runs on all the "dads" from which DNA was taken at YFZ, and see their results.

This may in fact be coming.

I'd also like to know who was excluded from DNA sampling, because my assertion would be if you get down to something like 99.5 % probability or perhaps 98.5 and also find someone who is 90% probable in the available "dad" pool, it's a lot less conclusive. Even if someone came back at 33% frankly that skews the numbers a bit.

I should mention that inserting a .02 probability means there is a 1/50 sample size of "possible dads." I suspect what the defense wants to do is increase that sample size to something much larger, preferably "all Caucasians."

One in a billion looks like this: 0.000000001 and is considerably different than .02.

I'll try to do the math, and get back to you with the result. If the defense can make the case that a larger net should have been cast and a probable suspect is out there, they're going to do that and they'll work that angle on appeal, provided there is one.

kbp said...

"However, by asking these questions in the suppression hearing, they've tipped their hand and the witness will be more prepared at trial."

I read of nothing I feel the other side would not have anticipated. The outcome I see is the state's witness looks to be weaker than the defense anticipated.

To qualify as an "expert" has always seemed rather easy, the courts most often allowing them in regardless of how well they do. The expert is now on record showing what she does not know.


That link you posted comes back to this comments page.

The Pharisee said...

Try it again.

I_hate_bigots said...

Let's say your last name was Smut.

Since only one family of five people has the name Smut, the probability of someone having that name is 99.999%.

If the State can prove a person by the name Smut committed a crime, they could arrest Sally Smut and say there is a 99.99% probability she is guilty.

But what they don't tell you is compared to her family members, there is only a 20% chance she is guilty.