But, Mr. Francis, you are a rapist.

[Dinsdale]

I’m not at all familiar with either Mr. Francis or his product. And I’m certainly not going to condone rape or filming underage porn.

But has he always been a rapist/underage pornographer? The earliest criminal charges seem to date from 2003, at which time he is already described as a multi-millionaire. And the story linked in the OP describes him starting his business in 1998 with a $50G credit card charge.

From a borrowed $50G to multi-millions in 5 years, from taking pictures of drunk young women showing their tits?

Now THAT’S exactly the kind of can-do capitalism that would have made our founding fathers proud. Seriously - who could not envy such a success story?

[Ensign_Edison]

@Bricker

I incidentally find it hilarious (although not suprising) that you failed to understand that Dead Badger was criticizing YOUR position above; your demand that I respond to him is just freakin’ funny.

And very consistent with the rest of your comprehension skills.

…He’s criticizing your position too, and the only “position” of mine he criticized was one he made up for me.

[Dead_Badger]

Erm, Bricker, I’ve posted several objections to your claims as well. I think prevalence of a crime has no outright implications for claim reliability. If you assume equal claim rates from blacks and whites, then I don’t particularly have a problem with what you’re saying, but I don’t think that such an assumption is supportable. If you’re simply using no data for claim rates, then you can’t simply omit that node from the probability tree; it’s a block after which you can’t compute contingent probabilities.

I have great sympathy for your effort to get your “prior probability” point across, but believe that your actual application of probability theory here is quite fatally flawed.

Ensign Edison, putting words in my mouth referred to your blithering fuckwit remark. I’m not keen to have someone with such an apparently deficient understanding of probability take my posts and use them to attack people. If I want to call Malacandra a blithering fuckwit, I will do so myself, thankyouverymuch. As for my not understanding that race can be replaced by other factors, I would like to draw your attention to the numerous places where I, er, replaced it with other factors. But never mind, eh?

[Dinsdale]

Correction:

has he always been an alleged rapist/underage pornographer?

[Ensign_Edison]

@Dead Badger

Ensign Edison, putting words in my mouth referred to your blithering fuckwit remark. I’m not keen to have someone with such an apparently deficient understanding of probability take my posts and use them to attack people. If I want to call Malacandra a blithering fuckwit, I will do so myself, thankyouverymuch. As for my not understanding that race can be replaced by other factors, I would like to draw your attention to the numerous places where I, er, replaced it with other factors. But never mind, eh?

I did not mean to say that you think he’s a blithering fuckwit. I was using your post because it contained some examples of reasons why I think he’s a blithering fuckwit. I apologize that I was not more clear.

And that’s kind of why I asked. It looked like you did understand, but then you were making like you did not. You specifically said that a certain statement belied the idea that you are now saying you understand and agree with. Hence my confusion, eh?

[Bricker]

@Dead Badger

Not without data regarding claim rates. Indeed, if whites raped whites 9 times more often than they rape blacks, but only 1 percent of rape allegations against whites were from black people, then we might then infer that in fact black people’s rape claims against whites have a higher prior probability of being true. Even then that would contain the dubious assumption that blacks and whites are equally likely to report such a rape if it genuinely occurs, along with the other simplifications I mentioned earlier. The best approximation we can get to a prior probability of a claim’s validity is to assess the percentage of claims that result in a conviction. Obviously this is a poor approximation, but equally obviously it intrinsically factors out the prevalence of the crime.

A rare crime merely means that one’s sample size for estimating the reliability of a class of claims is smaller, so in this instance the only effect that the prevalence of white-on-black crime would have is making our reliability estimate less accurate.

Right you are, which makes my initial assumptions important. If all others things are equal – that is, if we assume that reporting rates are similar, etc etc. Obviously it’s a poor approximation, and obviously it’s predictive value is slight. Tiny.

But it’s not zero.

[you_with_the_face]

@Dead Badger

To continue arguing from the middle: you with the face, do you not understand that analysing the probabilities is a matter of discovering which factors are relevant?

How could race/ethnicity be relevant, in and of itself, when assessing the probability that one specific claim is true? The mechanism for its supposed revelance should be explained well before the numbers come out and are used in assessing probability, don’t you think?

Extrapolating population-derived statistics to an individual case and attempting to estimate a probability that that particular case is true, based only on a data point that at most only has a correlative relationship with prevalence (which has no bearing on credibility whatsoever) makes. no. sense.

The whole point is that one can examine various contributing factors, and discover either that they have no statistically significant effect (as one would in the case of shoe size, for example), or one might find that they do, and investigate further to see if this is a causative factor, or is merely co-dependent on a hidden, unexamined factor.

This doesn’t conflict what I’m saying at all. I’ve been arguing against a troop of idiots who have decided to use race as if it is a causative factor for veracity, rather than something that is possibly correlated with a causative factor for prevalence.

What Bricker is saying is that IF one studies it and IF one discovers that shoe size does have a statistically significant bearing on a particular outcome, THEN you can indeed use that as a predictive factor.

What Bricker has been saying is if prevalence data for a particular type of crime (such as white-on-black rape) is low, then that means future reports of those types of crimes are less than fully credible. I wish it was more complicated and nuanced than that, but it is not.

[holmes]

@Bricker

Please try to follow the argument. I know it’s complicated and has lots of big words. But try to focus your meager reasoning power on this sequence of posts…

But it has everything to do with the discussion we were having and the point you posed, which I was answering and you evidently forgot immediately after posting it. Being imprisoned at Riker’s Island is NOT a neutral statement; it DOES imply something about character or behavior, and it IS correct to say that a random pick of a Riker’s Island inmate will be black.

See, how it works is, you say something, and I answer you.

Bricker you made a neutral statement, NBA members being black. That was the point I responsed to. You, as usual changed the question into a 'gotta". Perhaps if you laid out your point about Rikers, instead of trying to trap me; I would have responsed fuller.

No one including me, is disputing the law of averages or odds. What’s being disputed is whether or not that law can be applied to individuals, which is ultimately what this is about.

Yes, it’s correct to say thaty you may have odds of picking a black person in Rikers, what not correct saying that since your random picking turns up a black person, their blackness corresponses to their behaviour. All you’re doing is creating a confirmation bias.

They’re criminals because they are, their blackness is incidential, the behaviour and their blackness are not interchangible, it isn’t what makes them a criminal, any more than me saying that if I randomly pick a prisoner in Rikers, he’ll have hair; therefor hair is predictor of behaviour and all things being equal, i should treat people with hair, differently than I do people without it.

Further we’re not talking about self-selected and tightly controlled groups like your NBA or people in Rikers examples, so I don’t see how you can translate them into the general public, which is what I’ve been posting about.

[Random]

So, to recap some recent posts, Bricker asserts (I’m paraphrasing somewhat) that YWTF has demonstrated that she is incapable of applying logic or rational thought to a discussion.

Specifically, he says:

@Bricker

Yes, I admit stupidity. I continue to read your posts as though you have some scintilla of relevance and some modicum of a possibility of reasonable contribution to a rational discussion. That is, I admit, a pretty damn stupid assumption.

YWTF responds:

@ywtf

@Bricker

Yes, I admit stupidity.

It’s about damn time.

I submit that Bricker’s assertion is proven.

[Ensign_Edison]

Random, your woosh is showing. She snipped that part specifically to make a joke about it. See how she only included that part of the quote, so she could humorously take it out of context?

[Dead_Badger]

@Ensign Edison

…the only “position” of mine he criticized was one he made up for me.

My second post in this thread quite directly took issue with your contention that:

@Ensign Edison

You absolutely can’t use statistics to determine whether an individual person is probably lying unless those stats are about that particular person’s history of lying.

You absolutely can. Most potential discriminating factors will of course prove to have no influence on a person’s trustworthiness, and I would be very surprised if ethnicity proved to, but your bald statement that statistics can not be applied to this problem is simply false. Statistics will mostly prove that these factors are irrelevant, but it is most certainly statistics proving this, not some arbitrary notion of irrelevance as dictated by an observer.

We are, therefore, at opposite ends on the “other factors” question. I believe that any and all factors are fair game for study, while you appear to believe none of them are. Hence your confusion about my apparent agreement, I s’pose. You and you with the face appear to me to be making categorical statements that can only be proven by examining the numbers. You’ll notice that I’ve stated that most factors will prove to be useless to examine; where we part ways is that you insist that this means they shouldn’t be.

@Bricker

If all others things are equal – that is, if we assume that reporting rates are similar, etc etc.

But I think you’re effectively assuming the answer, in that case. In the post linked in the other thread, you stated as your extra information only the knowledge that the absolute rates of the respective crimes are very different. This gives you no information about the claim rates, and if you assume these to be equal, then you are effectively saying “if black people’s claims are less reliable, then their claims are less reliable”.

Basically, there is no predictive value in assuming the value for a node in a directly dependent chain of events. You’re positing a possible posterior outcome, assuming it is the case, then applying it back to make a predictive statement. You just can’t do this. We agree, there are lots of approximations involved, but the lack of data for the number of claims relative to the number of occurrences is an absolute dealbreaker when making predictions. You simply can’t assume your way around that. At heart, we are making the following calculations (I won’t duplicate them for both ethnicities):

P(black claim valid) = ( number of times victim actually raped ) / ( number of total claims )

We approximate this as:

P(black claim valid) ~= ( number of convictions ) / ( number of total claims )

Now, this is a nasty approximation, but there is clearly some dependency between the number of actual rapes and the number of convictions, so we proceed. But we still need to calculate that ratio, and to do that we need some indication of the number of claims. Not only do we not have that in this case, we have no approximating factor to substitute it with. We’ve therefore got two fractions without denominators. Sure, you can say that if both denominators are the same then the black claims are less valid, but that’s begging the question. So we might do the following substitution, and say

P(black claim valid) ~= ( number of convictions ) / ( total number of actual w-o-b rapes * black reporting rate + number of false accusations )

(where reporting rate is the likelihood of a genuine rape victim reporting the crime)

This is really where it breaks down. I think it’s a supportable assumption that, given a rape, whites and blacks are equally likely to report it. So let’s assume that’s equal. And again, we can approximate conviction rates to the total number of w-o-b rapes, but that still leaves us with the false accusation rate. At this point I think it becomes clear that further assumptions are truly begging the question, since you’re simply plugging in an arbitrary false accusation rate, which is the very thing you’re trying to assess in the first place.

Some gaps in knowledge can be filled with approximation. Some, you have to admit that this is a hole that cannot be filled with anything but data. Data on claim rates is most certainly the latter, IMO.

[Dead_Badger]

@you with the face

How could race/ethnicity be relevant, in and of itself, when assessing the probability that one specific claim is true?

How can you not, at this point, have acknowledged the very specific meaning of prior probability? If I were black, I would have had a much lower prior probability of going to the university I did; this would have no bearing on the posterior determination of whether I did or not and no-one is claiming it would. This distinction is so incredibly fundamental that it beggars belief that you’ve apparently participated in something like 12 pages of discussion on the issue without seemingly even beginning to grasp it.

“In and of itself” is just a red herring; Bricker is not saying that the person’s blackness would be the cause of their trustworthiness or otherwise - merely that it is a factor which could be examined, and which, due to dependencies on other less observable factors might give a result which is different to the same figure for whites. I disagree with him that the posited data allows one to infer such a conclusion, but you seem to be taking the position that even to entertain the idea that personal characteristics might give an indication of one’s behaviour is intrinsically bizarre. I do not agree, but see no chance of agreement on this point unless we can first establish what is meant by prior probability. And I see very little chance of that at all, on the display thus far.

Incidentally, I in no way think any such calculations should be allowed within a country mile of a courtroom, but that’s immaterial.

[saoirse]

It looks like catsix got her wish.

[you_with_the_face]

@Dead Badger

Statistics will mostly prove that these factors are irrelevant, but it is most certainly statistics proving this, not some arbitrary notion of irrelevance as dictated by an observer.

I can’t assume that something is relevant unless evidence demonstrates that it is. That proof has yet to arrive wrt race and rape. In the absence of such evidence, I will treat it as irrelevant in a discussion specifically dealing with plausibility. Just as any agnostically-minded scientist should.

And honestly, I think you’re giving Bricker too much credit. His argument really boils down to “if few there are few reports of a particular something, then that means that its statistically unlikely that such a thing would ever happen”. He’s not using any math to reach this determination. He thinks action on a roulette table is analogous to human behavior, for pete’s sake.

[Random]

@Ensign Edison

**You [Bricker] are exactly that person who knows nothing about the law…

Wow. Do I even have to spell out how obviously stupid this is, even if this is the only SDMB thread Ensign’s ever read?

As EE has been around long enought to read a few threads, this proves that he’s an utter moron.

Okay, as apparently the morons are out in force today: Ensign, Bricker is a lawyer. Even if he were a bad lawyer (which I, another lawyer, am reasonably sure that he is not), saying that he “knows nothing about the law” is nonsensical.

[you_with_the_face]

@Dead Badger

How can you not, at this point, have acknowledged the very specific meaning of prior probability?

Why are you talking about prior probability when the discussion is about an event that has already happened? If you are going to argue against what I’m writing, please argue in the context that I’m speaking about, not something that doesn’t apply to any of my arguments.

If I were black, I would have had a much lower prior probability of going to the university I did; this would have no bearing on the posterior determination of whether I did or not and no-one is claiming it would.

Um, plenty of people are claiming that. In fact, that’s exactly what they are claiming. Bricker included. And that’s what I’ve been arguing against! How can you not understand that, if you read the other thread and the Lying Whore one? This has gotta be the biggest lapse in reading comprehension ever documented on the board.

This distinction is so incredibly fundamental that it beggars belief that you’ve apparently participated in something like 12 pages of discussion on the issue without seemingly even beginning to grasp it.

Did you read the “Lying Whore” thread? It beggars belief to think that you might have walked away from that thread with the conclusion that I have a problem understanding probability.

“In and of itself” is just a red herring; Bricker is not saying that the person’s blackness would be the cause of their trustworthiness or otherwise - merely that it is a factor which could be examined, and which, due to dependencies on other less observable factors might give a result which is different to the same figure for whites.

Why are you so hellbent on adding all this texture and nuance to Bricker’s position that is simply not there? He’s not talking about doing any studies or examinations, and his conclusions are not peppered with “mights” or “possibles”, either. He believes that prevalence is equivalent to statistical probability. He has used racial crime data to assess the probability that an event that has already happened is indeed plausible. Those two things alone are what I’ve been arguing against. Nothing more and nothing less.

[Ensign_Edison]

@Random

Wow. Do I even have to spell out how obviously stupid this is, even if this is the only SDMB thread Ensign’s ever read?

As EE has been around long enought to read a few threads, this proves that he’s an utter moron.

Okay, as apparently the morons are out in force today: Ensign, Bricker is a lawyer. Even if he were a bad lawyer (which I, another lawyer, am reasonably sure that he is not), saying that he “knows nothing about the law” is nonsensical.

Random, are you deliberately wooshing me? I was making an analogy. I was comparing Bricker’s activity here as being LIKE a person who knows nothing about the law etc. Jesus, are you drunk?

[you_with_the_face]

And I apologize to the OP for this hijack. It wasn’t my intent to restart a trainwreck. I really just wanted to give Bricker a poke, and that was all.

I’m not going to add to it any further.

[Dead_Badger]

Eh. I’m not going to continue this here. I may start a GD thread about probability in this sort of instance tomorrow, but I’m not going to start explaining the difference between prediction and observation and the importance of when the outcome is studied in this thread, because it’s just too depressing to have to.

My apologies to the OP for contributing so heavily to this hijack.

[holmes]

Fair enough.