HBO has been showing "The Jinx" which is the story of Robert Durst. A piece of evidence came out in this Sunday's episode which reminds me of Sklansky's application of Bayesian thinking when the size 18 bloody footprint evidence came out in the O.J. Simpson case.

Here is the Wiki on Robert Durst
http://en.wikipedia.org/wiki/Robert_Durst

Spoiler Alert if you want to watch the show in the dark.

Cliffs on the story:
1. Durst's wife disappeared in the early 80's. He is a suspect in that case.

2. Durst fled to Texas in the early 90's when his wife's case was reopened. He disguised himself as a mute woman. He dropped the disguise becoming friends with a neighbor. The neighbor's dismembered body is discovered in garbage bags floating in a lake - the head never recovered. Durst is tried for the murder, claims self defense, claims victim was shot with his own gun in the head during a scuffle. Durst admits on the witness stand that he dismembered the body because he didn't think the police would believe it was self defense. He is acquitted.

3. Durst's female friend who people suspect helped him cover up his wife's murder by making phone calls impersonating his wife is in Los Angeles in the early 2000's when the Wife's case is reopened once again. New York detectives will interview the friend. She calls Durst and tells him she is going to talk to them.

Police receive a letter addressed to the "Beverley Hills Police" telling them a corpse can be found at the friend's address. The friend is found murdered, mob hit style. She is the daughter of a famous Las Vegas mobster and is working on a mob expose for TV. Durst cannot be put in L.A. at the time but can be put in northern California with some questionable time where he has his phone turned off.

4. Here's the size 18 bloody shoe print evidence just revealed Sunday. Going through the female friend's papers a letter from Robert Durst to her is found. The city line for the address is hand written similarly to the handwritten letter to the police, as "Beverley Hills, CA". Beverley is misspelled the same way as the misspelling on the letter to the L.A. police informing them of the corpse.

The question is, if you were on the fence about Bob Durst's guilt in the murder of his friend how much does this coincidentally identical misspelling of Beverley sway your opinion toward guilty?


PairTheBoard

1. I'm never on the fence about anything. Bob Durst's head should be cut off with a hacksaw. Whether innocent or guilty doesn't matter to me.

2. Since the backwards educational system in the US began its ridiculous crusade about proper spelling (vastly overcorrecting a minor problem) it has become an overblown requirement in society to always spell everything correctly and if you do not; you are labeled an idiot and have committed an unpardonable sin of such vastness that your head should cut off with a hacksaw.

3. How common is the misspelling of the word in question and is it usually done in the same way?

From the wiki references this guy appears to be an ahole even if it were true he is not guilty in some particular case say. At some point in time even if i am not supposed as say juror to think that way i have started to not give a damn about being wrong that he is guilty...I am not saying i will close my eyes and decide without evidence this guy did it in some case but i wont have the same small threshold for doubt (to not be ok with a conviction i mean) and care that i do for some other random person. That right there is a dangerous property to have as a human being. When people look at your life you need to be giving them the impression that you are not a complete loser, self serving individual that has always done what was necessary to get away from spots that deserved to be handled a lot better. This guy gave us many examples that he didnt operate ok.

Screw such people , i dont care if i am wrong. Putting them away by error is not really an error! Put them away in prison and keep in the back of you mind the cases open for further clues later if something happens. Such people do not deserve to be free out there...

What was the spelling error by the way? Do some Bayes proper analysis there based on how often people make such errors and how often a person making such error will remain consistent to do it again (lol). You have some level of doubt that you want to meet to not convict someone and if the particular property this guy has (error in spelling) is rare enough you get there. Also other evidence has say your doubt at 5% so this may take it to 1% if eg the math says so. It affects that way your decision...


Even if jurors may be advised to not think this way technically we all apply some informal Bayes approach in deciding about others by using their prior lives. Clearly you do not generally expect a killer to be a fair well mannered person in general that has never had problems with the law. Sure there are such killers but people with trouble past/lives are more likely to get in future trouble also vs a random individual. That doesnt affect the particular case if you operate properly as juror but must affect your doubt levels and in fact to not affect it is the real injustice in general. It must matter to society if you werent an ahole in your life.


PS; you mean Beverly Hills vs Beverley Hills? Maybe this is not a very telling thing in general as i expect many people to make such error if not from that state or location or if not paying attention to such details. Maybe >10% of people do it? I could have been in them in some point in my life though.

Sklansky's point on the size 18 footprint (or whatever rare size it was) was that most jurors would not adjust their lean toward guilty nearly as far as the Bayes calculation dictates. Not sure if this is enough to qualify them for his "moron" label.

I don't know how rare the "Beverley Hills" misspelling is. I suspect not as rare as the size 18 or 24 or whatever it was footprint that perfectly matched O.J.'s size.

PairTheBoard

Our Lord and Savior David Sklansky does not have a particular good grasp of reality on the inner workings of the general, the plebian, the hoi pollio, the throbbing masses, or even the hoity-toity. This is not meant to be dismissive. His point is well taken. It just overshoots.

It should be interesting to see how Durst responds to this evidence when they spring it on him in the next interview this coming Sunday on "The Jinx". This is new evidence, never discovered by the police, just uncovered and being kept secret by "The Jinx" producers. These Durst interviews with "The Jinx" are the only ones he has ever done. He has had plenty of time to anticipate all the questions and practice his answers. In fact, he was once caught alone in the interview room, unaware the camera was still rolling, practicing his delivery of a difficult line, "I didn't tell the Whole Truth. No one tells the Whole Truth".

He will not have had a chance to practice his response to this new evidence of his letter with the "Beverley Hills" misspelling. It's not as damning as the size 18 bloody footprint for O.J. but it certainly looks bad for Durst.


PairTheBoard

It would sway me but I'm not sure it would sway me sufficiently for me to go from the fence to beyond the threshold of reasonable doubt. On a lesser burden of the balance of probability it would.

It sounds likely to me that Durst suffers from oppositional defiant disorder which can be comorbid with Asperger's. The Beverley spelling is illuminating but the track record of all these people close to him dying is much more significant.

Since he has Asperger's he lacks the social component most people have and so the big question is whether or not he can appeal for consideration on the grounds of mental incompetence.

There are probably all kinds of things that could have triggered his aggression towards people. It really makes you wonder why the family let him get away without establishing some type of legal guardianship over him so he wouldn't commit these crimes.

Durst arrested yesterday in New Orleans hotel, registered under a false name.

Today's story in the New York Times regarding the arrest and finale of the HBO documentary "The Jinx".
http://www.nytimes.com/2015/03/16/ny...sted.html?_r=0

When confronted with the damning "Beverley" letter, Durst shows no reaction other than a burp. He calmly denies having written the Cadaver letter to the Beverly Hills police, saying it's a common misspelling and all block letters look about the same - like typed letters.

But after the interview, Durst goes to the bathroom not realizing his mic is still live. After two years the Jinx producers discovered Durst's bathroom whispers to himself on tape. He whispers things to himself that sound like a confession, e.g. "I killed them all". On the other hand, I think he can claim he was speaking sarcastically or ironically or whatever. i.e. Realizing how bad it looked and being resigned to being Bad Luck Bob once again.


PairTheBoard

What? Does he talk to his penis? This all seems rather odd. Bathroom whispers; that sounds like a new internet meme. I'm confused. Perhaps he is talking about nothing instead of something.

I sense a conspiracy. That usually works well for an explanation when things get convoluted to this degree.

This is a youtube of when he whispers, "I killed them all. Of course". It will probably be copyrighted off youtube soon. And there's nothing for his entire conversation with himself where he also says, "This is a disaster".




PairTheBoard

"(I) Killed them all, of course."

Cool. No crime in that by itself. Sort of an open-ended comment and not directed at anyone or any group, etc. Actually, by itself it is a rather empty phrase. I didn't hear the "I" . Sounded more like: Kill them all, of course. to me.

That people mutter to themselves in toilet facilities seems odd to me.

Yeah i sure hope they have more than that lol. I bet even the whoever you want to put there well respected person (president, author, famous scientist) could have thought in their mind how to go about doing the perfect crime as an exercise in how criminals make errors and how to top them or see why they think that way etc. Now if you actually lived alone and talked it out it would look funny to an observer but many people talk out like that when alone. Now you may not do it in the bathroom lol but i have caught myself many times arguing out loud things that a random listener would find strange without context. And then i also thought about this happening too lol (ie being recorded doing that and having to defend it too) !

Maybe the guy can tell them he was thinking like this; All you guys think i did it, imagine if i came out and said to myself (as thought) " right now there is evidence that incriminates me although i am innocent. Say a coverup took place that i participated at but i am really innocent anyway of the main crime (only of hiding evidence). That evidence has a chance to be revealed in a few years but maybe not yet (for some good reason say, eg a house may be demolished soon and uncovered but not now). Imagine if i incriminate myself now on purpose with something else stupid that can be defended easily, i win the trial and then i have double jeopardy in place when/if the real evidence comes. (i assume this is true not? http://en.wikipedia.org/wiki/Double_jeopardy) "

So i am thinking out loud sure i did it etc. Even offer some other evidence in a fictitious discussion with myself playing it out how to do it. If you were to record me while i was playing out these thoughts it would look ugly without being the real deal.

I dare anyone here to claim they never thought out how to do a crime and get away with it as an exercise in thinking to spot the errors of a process. It has nothing to do with actually wanting to do a crime or philosophically/ethically objecting to it. Its called free thinking. If you are alone you can easily be heard talking like that when doing things where no particular focus is needed. Well probably not the bathroom but some people sing in the shower for example. That is another thing i wouldnt do but i have noticed myself arguing things during a documentary out loud or in a political/recent events debate etc without anyone there to listen to me. Thats standard. If you record me only when i say out loud the partial statement and not the one that started it i can be effectively be heard saying all kinds of things that make no sense or can be made to imply things never intended etc.

This guy is probably guilty anyway but this is not evidence. A very egocentric person with high idea about themselves can always talk to themselves like that for attitude like say sure i did it i killed them all, go prove it lol idiots!

Now if you reveal while talking to yourself some detail like say i did it but you will never find that suitcase in the bottom of that river with the evidence because i made sure its well secured, oops you revealed it now. Thats another story.


http://www.aol.com/article/2015/03/1...6pLid%3D628711

I can see the guy being innocent and thinking to himself -

Thinking:

"I Killed them all"

Thinking:

"This is a disaster. He was right. I was wrong"

etc.

PairTheBoard

This is everything Durst said on the bathroom recording played by the Jinx documentary - in order. I would describe his tone as a combination of pissed and resigned with notes of sarcasm.

There are significant pauses between each line:


There it is.
You're caught.
You're right of course.
But you can't imagine.
Arrest him.
I don't know what's in the house.
Oh, I want this.
What a disaster.
He was right.
I was wrong.
And the burping.
I'm having difficulty with the question.
What the hell did I do?
Killed them all...of course.


PairTheBoard

The photos of the 2 addresses with "Beverley" are actually more incriminating than even the spelling error or the easily defended but still bad looking at surface (for public consumption say) mic recording. If you look at the letters they are significantly close to each other as if made by the same person.






That too of course is only a small piece of evidence that can be defended with proper statistics lol. However if you add a lot of such evidence that little by little kills the doubt probability you can get there even without the real deal. Pretty soon they will introduce math in courtrooms because it is getting ugly out there and i bet a lot of defenses can use math to destroy evidence if they can convince the court of the proper math (gl with that but imagine if forced to teach jurors and the judge Bayes updating/inference or something).

http://en.wikipedia.org/wiki/Bayesian_inference
(eg see the section "in the courtroom")

A proper trial must try to if possible offer some math to the case. Like say 100000 people in some area could have done it (of course not all with the same probability) . Then the killer has properties A1,A2,..Aj and the accused has all these properties and a random person in the city has each one of these with p1,p2,...pj chance etc. Eg the killer was tall over 1.9m. The killer was male. The killer was blond. The killer had rented a car that was red that day. The killer had bought a particular chemical within past 60 days. Now put this all together and find the chance the accused is innocent and got unlucky by simply doing all these also by accident.

That can be a pretty tricky mathematical problem also if there are correlations (or some valid ambiguities/uncertainties/errors in evidence collection, systematic errors etc). So the 2 lawyers can find themselves battling with math arguments lol trying to go below the critical doubt level or remain above it in some math intense various probabilities updating dueling game that involves tutoring the jurors and assigning homework too lol! The judges would love that (noway)! Force the jurors to convict only if they can prove to themselves they understand the math enough (or feel guilty if not! ie introduce doubt on their own competency on top of the other doubt too!)

The Jinx producers took the two letters to a forensic handwriting guy. He said the sample was too small to make a good comparison, especially considering the block lettering. So the Jinx producers dug around for everything Durst had ever written where he had used block lettering and gave it to the forensics guy. That was enough for the guy to conclude the incriminating "Cadaver" letter was definitely written by Durst.

Of course the Durst lawyers will have their own forensic handwriting guys.


PairTheBoard

I anticipate David Sklansky being called as an expert witness on the application of Bayes Theorem. When voir dire-ed Sklansky will respond, "Hey, I was right about O.J.'s size 36 shoes wasn't I?"


PairTheBoard

I sure hope they don't bring math to the courtroom. It's not hard to mislead people with math if you make it a little more complex than they actually understand, and then trying to convince them that they were misled would be near impossible. Plus, then you would have to define the probability for 'beyond a reasonable doubt.' You'd essentially be defining how many innocent people we convict, which is a number we should work on minimizing, not writing into stone.

I was just now reading a bit about this idea and found an article that suggests we use probability differently in criminal cases. Instead of having juries work out the probability of guilt given the evidence, P(G|E), and convict if it's above some large number, have them work out the probability that this evidence could have come about if the defendant is innocent, P(E|G'), and convict if this is below a certain small threshold (ie the defendant's explanation isn't reasonable). I feel that using either probability alone is inadequate and would lead to inappropriate convictions.

Yeah but it sure beats any day to have a 1% doubt threshold target rather than the 2-3% i think that is the error rate for innocent people that were convicted. My own would drop to 0.1% if the person had a good prior life and decent history and raise 1-2% if the guy was an ahole. Hard to estimate but as a basic idea on what could be estimated roughly.

The problem is that indeed sometimes its very hard to put numbers to these things but tell me how you view a DNA evidence where the chance of an error in the collection of evidence is like 0.1% and the chance of corruption from prior historical evidence around 1% (rigged evidence etc). Suddenly that 1 in 100 mil starts looking unrelated to the rest!

And its certainly very good to be able to conclude that the evidence is not as clean as they wanted to make it and the doubt is now at some 5-10%. I bet a defense that proves that the doubt ought to be larger than 5% is a good result if you then tell people that this means 1 in 20 people are innocent!

I am afraid though that juries have very high doubt thresholds maybe even 5-10% anyway and the smaller error rate is because sometimes the job is good enough to be unlikely the police is targeting the wrong person moreover their inability to present a perfectly good case. I do think unfortunately that without forcing the jurors to have some target number in mind they can define doubt as some small probability that is below 50% but not at 1% or lower level, more like at a typical minority of a population level say 5-10-15% type thing.

Going by the feel to me is very unreasonable. I would feel ugly about it. I mean yes they can present to me evidence that he has all the properties the killer has that they have identified as important to show but if this still leaves him a 5% chance to not be that person because the evidence is not super strong in every property then i wont be very happy with it. But someone that is not paying attention may think wow he did match 5 different properties , remarkable!

There are also issues of poor defense or poor collection of information. Say a drug is found in the system of the victim and there is evidence the accused had bought such drug/pills recently but then they fail to show the victim also had access to such pills and the accused also had bought such pills in the past for personal reasons etc. It can get pretty ugly estimating things with poor information. Now this is an example of a dirty kind of evidence but it may still be useful up to a point if a little more clarity if offered about eg the frequencies these people were using such drugs or combination of other things that would have made it less likely to be drugs the victim used that day given what else they were doing at the time that typically would be conflicting with such usage of their own choosing etc.

I mean certainly with all the chaos and imperfect information available its unreasonable to ever claim you can estimate your doubt. But you could sometimes estimate a bound and maybe error for your doubt and then use that. Without being told what to aim i would love to know if the jurors were ever asked to declare after every decision what chance they had in mind for the accused being guilty. I bet many would be happy with 90%,95% and not think about the rate they would convict innocent people that way, only that in their minds the situation looked pretty ugly for the defense with ugly meaning 90-95% confidence of guilt because overall in their lives they consider 90-95% close to sure thing, never really participating in things/activities that require more care than a crude guess game of relatively good confidence.

Compare that with eg a doctor that needs to be very careful in diagnosis and prescriptions in certain cases (more than others maybe) or a particle physicist that is thinking of 5sd results lol (not necessarily applied here but having been exposed in their lives to situations that such confidence is necessary to exist for them to offer a perspective of what is rare and what isnt so rare and therefore risky in cases of importance such as one man's life). People have different lives and appreciate risk differently. If your life is ugly i doubt you have a very small doubt threshold. The more you can imagine risks not presented by the defense the more this can elevate your doubt higher than someone else that doesnt care to be a bit imaginative and unhappy with certain evidence.

That (the bolded) is a terrible basis for justice. Just terrible.

Huh?

The murderer lives in a town of a million and has a size 18 shoe. Three people wear this size and all three have no alibi. If you use the first criteria, when trying one of them you get a probability of 1/3 and it is an easy acquittal. If you use the second, you get the number .000003, the probability that an innocent person wears this size, and you would convict a probable innocent man.

If only you could have quoted one more sentence of my post, you wouldn't have needed to make this post at all. Too bad the last sentence in my post was unreadable and unquotable.

No. Far more than three percent of innocent defendants are convicted. Three percent of convicted defendants are innocent. That means that jurors are willing to be far less than 97% sure to convict. It means either that or it means that near certain jurors are incompetent. Probably both.

It is only because of the fact that most defendants are both guilty and obviously so that we avoid discovering how inept jurors really are. I'll let PTB and masque collaborate to elaborate.