My apologies Cliffy for including my points in your quote. Not to good with this stuff.
Cut and paste.
I actually do have a point to make. It would be in the best interest of the defense team to establish a high “percentage of probability of guilt”. Whether they do this subliminally or more directly they certainly try to do it.
Sorry, but I’ll have to weigh in again. Cliffy has explained perfectly, but the Kid still doesn’t get it.
Yes, you can assign a number to an intuitive yes-or-no answer, and even get paid a lot for doing. That doesn’t mean it ain’t B.S.
What you asked is “What … probability of guilt meets the criteria of beyond a reasonable doubt?” You can’t seem to get it through your head that
(1) The probability of guilt is NEVER CONSIDERED by a judge. A judge may estimate the chances of a conviction, if he is exceedingly irresponsible, but the probablity of guilt is something that a judge will not think about because it in no way relates to his job. Criminal court judges make decisions about evidence, courtroom conduct, etc., but only in a few cases will he decide guilt. So why would he even go there?
- “A reasonable doubt” has nothing to do with “probability of guilt”; the defense raises doubts, the prosecution attacks the doubts, the jury decides: reasonable or not? Yes or no. There may be dozens of doubts raised in a some cases, and the jury will decide for each doubt whether it’s reasonable. If none of the doubts is reasonable, the defendant is guilty. If even ONE of the doubts is reasonable, the defendant is acquitted. The jury does NOT consider “Do I doubt he’s guilty? Is that reasonable?” That’s not what the words mean.
Now, in any given jury, there’s probably several (maybe even a majority) who don’t understand any of the above, and will just say “I’m about 95% sure he’s guilty, and that’s good enough for me.” But you asked about judges, and a good, responsible judge, if you ask him that question, will say “It doesn’t work that way…”
The quoted survey, of course, glosses over the reasonable doubt issue and asks what certainty of guilt the respondent would require. That’s rather definitely NOT what you asked, and if that’s what you wanted to know, I’ll have to suggest that next time, you avoid blithely tossing around technical terms that you don’t understand, and just SAY WHAT YOU MEAN.
nametag I suggest you reread my response to cliffy. YOU are missing the whole point of the exercise. Just because you don’t understand what I mean doesn’t mean I didn’t say what I mean. As you can see, I asked a question and don’t ask provided an answer. Apparently the guy who wrote the article and all the judges who responded to the survery understood my question. Why can’t you? Don’t get bent out of shape because someone provided a cite that you said wouldn’t exist.
Except for one-egg twins (triplets, etc.) it is virtually impossible for two individuals to have the same DNA (many, many orders of magnitude larger than the number of people on earth). But they don’t measure all your DNA, just a small snippet. It is still much more individual than, say, fingerprinting in which there is too much subjectivity. And far beyond reasonable doubt. Considering that personal identification is usually considered the gold standard of identification and is notoriously flawed, it can be taken as the nearest thing to proof available.
What is reasonable doubt depends primarily on your wealth and skin color.
You should use my sig.
oops, here it is
I`ve been using it for about a month now.
Golly, not to wade into this too deeply, and I do understand the question asked in the OP…different jurisdictions will have jury instructions to help jurors determine what constitutes “reasonable doubt.” I don’t know of any jurisdiction that from a legal standpoint includes a numerical value in its instruction.
As far as scientific evidence goes, I don’t have the case cite at hand (looks like Daubert if I’m reading my upcoming cite correctly), but for scientific evidence to be admitted the methodology has to be “generally accepted as reliable” in the relevant scientific community and appoints judges as the “gatekeepers” of what is and is not considered to be generally accepted as reliable. There were some cases over the last few years challenging fingerprint evidence but AFAIK they’ve pretty much petered out.
Ah, but here’s your problem. Work on creating expert systems (including legal expert systems) fell down rather badly in the 1980’s because knowledge engineers tried to capture expert’s understanding of domains in terms of probabilities. It turns out that this is not really the way that experts think (or if it is, they don’t have access to it). You may be able to assign a degree of belief to something, but the experts became uncomfortable manipulating this value as if it were a formal probability (c.f. Rule-based Expert Systems by Buchanan and Shortliffe). If you do elicit numbers from experts (e.g., your judges), they often turn out to be conflicting or meaningless.
Now you may ask what this has to do with your OP -- well, even if humans are "souped up fuzzy logic systems", this doesn't help you evaluate the probability of a situation.
a. You have no access to the "internal" values of your fuzzy logic system. All you have access to is intuitive qualitative characterizations of how likely something is. As in "No way", "Probably BS", "Likely", or "Beyond a reasonable doubt".
b. Actual probabilities as stored in a fuzzy logic system would depend on the level of training of the system (how many comparable situations it has been presented with). Each observer is therefore going to assign different probabilities. Without consensus, hard numbers aren't going to be meaningful.
c. Producing a single numeric probability of a complex situation is amazingly hard to do. In a non-closed world, there are just an amazing number of possibilities. To achieve a probability, you’d have to enumerate all the possibilities and assign likelihoods to all of them.
d. Finally, and probably most key – humans really, really suck at intuitively grasping probabilities. Study after study has shown this. (My sophomore probability class proved in it my case.)
IANAJ, but to me, beyond a reasonable doubt means somewhere in the neighborhood of 90% sure.
This should not be a General Question, and I don’t think he’s looking for a Right Answer. I don’t think there’s even a Consensus to be gained here because all the judges will not reply to this poll.
What he’s searching for are opinions, and with that I humbly wait for the Moderators to ship this thread off to the land of IMHO, where such whimsical queries based on questionable correlations will be greeted with much deserved acceptance.
IIRC paternity tests when they 1st came out were not accurate enough to be used in courts (something like 90%) and now it is accurate enough (99%).
Finagle, you claim it is not possible for a person to access the “internal values” they might use in “fuzzy logic” calculations. I believe this is not necessarily the case.
As referenced in the article don’t ask linked to, it is generally assumed that rational agents act to maximize their expected utility. I believe it is also a typical assumption that for most people the utility of a small amount of money is proportional to its value.
Lets say you have a general feeling that X is possibly true. Imagine that there exists a game, where a moderator, who knows whether X is true, will give you $100 (or some other convenient amount) if X is true and nothing if X is false. The question becomes, then, what is the largest amount you would be willing to pay to play this game?
In principle, this sort of procedure can be used to determine the subjective probability that a person attaches to any event. Though no such test could actually be carried out in situations, like jury trials, where the real truth value is unknown, this does demonstrate that a numerical measure of certainty exists.
Unfortunately, this assumes that human beings are rational, which is at best a good approximation to the truth. (e.g. I have an irrational distaste for risk, and would hesitate to bet on a coin flip unless the odds were significantly in my favor.)
JasonFin,
Absolutely right. At my former job we used to bet on EVERYTHING. The standard unit of betting was a 100 futures contract which expired at 100 if the proposition was true and 0 if false. There is a huge network on Wall St. of people who bet on everything from College Basketball to murder trials (I cleaned up on OJ) to what the next beaujolais will score in Wine Spectator. The midpoint of the current bid/offer of any given contract was a reasonable assessment of the subjective probability of that proposition being correct.
We know what reasonable doubt implies in grammatical terms. But the conversion to calculus is something that the judiciary in every province flatly refuse to answer perhaps because it would prove divisive. To me 96% seems about right however, this now presents a secondary problem?
What mathematical formula can I now utilise to formulate what constitutes a 96% probability of a reasonable doubt? Am I saying the whole jury should be 96% sure, or just me? If the whole jury, then I am 8% or one twelfth of the whole? Therefore if I decide guilty and 2 members say not guilty but that is enough to carry a majority, should I change my vote to not guilty rather than let a mere 80% certainty convict the defendant?
Does the conversion to calculus cause more trouble than it is worth?
Tidying up my previous post
Quantifying Reasonable Doubt
We know what reasonable doubt implies in grammatical terms. But the conversion to calculus is something that the judiciary in every province flatly refuse to answer, perhaps because it would prove divisive. To me reasonable doubt represents a certainty of at least 96%. However, these figures now present a secondary problem that I hadn’t bargained for?
Thinking this figure might help to solve the problem, may have been somewhat naïve.
The paradox
What mathematical formula can I now utilise to formulate what constitutes a 96% probability of a reasonable doubt in grammatical terms? How can I quantify that I have achieved 96% probability? The inevitable answer is I can’t. Moreover, am I saying the whole jury should be 96% sure, or just me? If I am saying the whole jury should be 96% sure, then I am 8% or one twelfth of the whole? Therefore if I decide on a guilty verdict along with 9 other fellow jurors ,but 2 other members say not guilty, but that is none the less enough to carry a majority verdict on a 10 / 2 basis, should I consider changing my vote to not guilty rather than let a mere 80% overall certainty convict the defendant?
The conversion to calculus appears nothing more than an imponderable circular argument.
Unusually, calculus doesn’t help us on this occasion; reluctantly I have to concede that the law has it right on this point, with the exception of allowing a 10 / 2 majority convict?(particularly on capital offences)
There really needs to be an answer to this question, because a case really can come down to hard probabilities, even if they don’t always.
Example: Suppose we have a rape case. The jury is completely satisfied by the evidence that the sex was non-consensual. The prosecution calls as an expert witness a geneticist who tested the semen and the suspect, and says that based on the mathematical analysis he performed, there is an x% chance that the semen came from the suspect. The jury accepts that the geneticist is honest and has performed his tests and analysis correctly. Should the members of the jury vote to convict or not?
Keep in mind that this thread is 10 years old.
Actually temperature is a matter of quantity. That’s exactly what the Kelvin (and Rankine) is: degrees Celsius (or Fahrenheit) above absolute zero. Of course that’s not what the car company meant. Since absolute zero is -273.15 C, on a warmish day of 26.85 C (or 80.33 F), the tinted window would reduce the temperature from 300 to 225 kelvins (or -48.15 C or -54.67 F).
Oh, and what the car company probably meant was that the windows would decrease the temperature difference between outside and inside by 25%. So if an untinted car would end up 20 degrees hotter than ambient, one with the tinted windows would only be 15 degrees hotter than ambient.