Announcement

Collapse
No announcement yet.

A discussion on weighing up two sides of an argument

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • A discussion on weighing up two sides of an argument

    Hi all,

    On occasion I have noticed debates or discussions stalling at a point in the discussion because the two viewpoints get stuck trying to come to some agreement over whether or not the evidence we have provides sufficient grounds to move on. This, of course, is a worthwhile focus at times, but there comes a point when its worth is questionable.

    I'm going to present things here in a slightly abstracted form, because it is the notion of how we compare two competing views that I want to focus on, not any specific example. Partly, I wish to avoid the thread getting distracted by JtR specific topics, and focus on something that is applicable to much of what we do here.


    If side A is presenting some explanation for events, and they bolster that by referencing something from the pool of information we have to work with. I’m using the word information here rather than evidence simply to avoid getting distracted by definitions. The information pool includes all of the things we have to examine, from court documents, police letters, newspapers, memoirs, and so forth. Now let’s say the side B argues “you cannot move on because there is some possibility that information is not correct”. How should we evaluate that argument? Is it sufficient to halt the discussion at that point in order to resolve that issue, or is side B’s argument insufficient?

    Well, that really depends on how large that possibility of error is. I’m going to switch to use the term probability here, but only in the sense of percentages, and not in the sense that probability necessarily means “with high likelihood”. The probability, therefore, implied in Side B’s argument could be a value anywhere between and including 0% and 100%. I know side B’s argument doesn’t make any sense if the probability is 0%, but 0% is a value for probability so it technically gets included here.

    Obviously, if we can demonstrate that Side B’s probability for error is 0% (or the converse, that Side A’s probability for being correct is 100%) we move on. Conversely, if side B can show the probability for error is 100%, then Side A has to drop their claim.

    While the above constitutes the ideal, it is never achieved in the real world. This is why to secure a conviction the prosecution has to establish their case beyond reasonable doubt, and not beyond all doubt. Which is to say, the defense has to establish reasonable doubt, not just any doubt.

    Too often arguments get bogged down at a point where one side or the other continues to point to the fact that this ideal has not been reached, that the probability for error has not been reduced to 0%, therefore their presentation must still be considered with as much regard as the other side’s. That, however, is not true.

    To keep it simple, let’s say Side A is arguing for the idea A to be true (that could be something like the time at which a murder occurred, for example, or that Stride is a victim of JtR), and Side B is arguing that is not true (donated as ~A, which just reads as “not A”). Now, because things must be either A or ~A (that covers every possible condition), then the probability of A + ~A is 100%. Making the probability of ~A = 100 – the probability of A. Simply saying that the probability of ~A is > 0% is not a sufficient reason to drop the conclusion A.

    What we do is compare the ratio of those probabilities, and create what is called an “odds ratio”. For example, if we knew the probability of A was 90% (so we had 90% confidence that A was true), side B is arguing “But there’s a 10% chance that ~A is the true state, so you cannot move on and my counter explanation is just as good as yours”. That might seem appealing, but notice, 90% is 9 times 10%, we have an odds ratio of 9:1 that side A is correct. That would be considered, in the context of JtR discussions at least, a safe bet. If, however, the evidence for A was weaker, and Side A only had a 50% chance of being correct, then one is no better off betting on Side A than Side B, we’re at a point where indeed, Side B has every reason to argue that moving on is a bad idea. Obviously, once Side A’s chance of being correct falls below 50%, then Side B’s case is the safer one to bet on, and increasingly so as the chance for Side falls lower and lower.

    Now, in most situations, we can’t put actually numbers to each side. What we do is present ties to information. We debate the relative reliability of various sources of information, and so forth. And from that, we examine which side has more information in support than the other, and we would include in that examination the quality of the sources, and how direct the information is connected to the debated topic. If one side comes out with an overwhelming amount of information in its favour, and the other side only has the argument of “but it might be wrong”, with either little information to show it is wrong, or at best only tangential bits of information far removed from the debated point, that is a signal that the odds ratio fall strongly in favour of the side with the supporting information.

    Does it mean that side is 100% absolutely incapable of being wrong? Of course not, we never can reach that absolute state of certainty (even DNA analyses in courts are presented as having some extremely small chance of being wrong, after all – do we toss that out? Or do we go with the odds of it being right vs wrong? Clearly, it is the odds that matter).

    I think it is important, and would be helpful to many discussions, therefore, to keep in mind that it is not a valid argument against a point to simply say “that information might be wrong”. What one has to do is draw upon information to show that the odds of it being wrong are truly in favour of that over that of the other Side. If neither side has much information, we’re getting around that 50% point, but if one side has a lot of information, of a reasonable level of quality, and the other side has nothing but that fall back position, the side with the supporting information is considered to have the greater odds of being correct.
    Yes, there is a lot of subjectivity in what I’m outlining here, though there may be times when it can be more objectively determined. What’s important, though, is that the argument “there’s a possibility the information is wrong” is not information that gets one great odds; it does not make your argument a safe bet, nor does it make the other side the unsafe bet.
    • Jeff

  • #2
    It's for this very reason I'm extremely curious to see what happens if the Zodiac Killer ever gets solved.

    The only other case comparable to JtR which has ever been solved is EARONS. For 40 years there were in depth debates. Disagreements about many aspects of the case and arguments over reliability of information and aspects. Many people had a suspect in mind or theories they wouldn't budge from. Ultimately none of it mattered. Joseph DeAngelo never came up on anyone's radar once. No one got close to the correct answer. What this shows to me is that both sides can easily be wrong including the source material. In GSK attack #24 it was reported that the victims heard a car beeping it's horn and someone banging on the door while the attack was taking place. For years this led to a debate about a possible accomplice. In mid 2018 the victims in an interview revealed they had recently discovered it had been some close friends of theirs who had come to see if they wanted to hang out. Despite this I still see people erroneously use attack #24 when trying to implicate JJD's wife, brother or another person as an accomplice, despite their being no evidence whatsoever for one. This shows that for all we know any of the witnesses or inquests in the JtR case could be wrong but since it was so long ago we will almost certainly never know for sure.

    Now back to the Zodiac. Ideally I would like it to be a suspect people already knew. If it was someone completely unknown like EARONS I feel JtR will almost certainly be the same. However if it is a known suspect it will give me confidence that we can solve this 130 year old mystery. What I would be interested to see is the fallout of a culprit being undeniably declared by the FBI to be the Zodiac. Mainly because if anyone thinks people on the JtR forums are protective of their pet suspects, which I don't btw, or theories about JtR, oh boy, the Zodiac forums are way way worse. The length some of them will go to in order to push there suspect and discredit others is startling and debates are much more heated which many personal attacks.

    For example, if Arthur Leigh Allan was confirmed to be Zodiac by the FBI it would essentially mean this first main suspect was the actual killer. Many people have spent thousands of hours and pounds trying to prove the killer couldn't be ALA and that Graysmith is a liar about many of his claims. However this would mean Graysmith was right from the very start, the person the police at the time thought was Zodiac was actually Zodiac. Despite only circumstantial evidence supporting claims he was Zodiac and seemingly disproven DNA (very arguable IMO that neither sample were no where near good enough for a proper comparison), fingerprint matches and handwriting analysis which would all point at it not being ALA. Would this mean JtR can also be correctly identified with circumstantial evidence? Would Druitt or Kosminski be seen as more likely suspects due to being the main contemporary police suspects like ALA?

    Alternatively, what if Richard Gaikowski was declared the Zodiac. A suspect some view as a complete joke and others view as the undeniable culprit. What what happen then, a suspect the majority of researchers have discredited with what is seemingly strong information turns out to be the actual killer. Would this put suspects like Deeming back into the picture despite the strong evidence he was in prison during the WM?

    If Rick Marshall or Laurence Kane was the Zodiac Killer would in hindsight the living officers wished maybe they had focused on one of them instead of ALA since both were known to the police. RM was even the original suspect in the first draft of Graysmith's iconic book, but his publisher wanted a more interesting suspect so the second draft switched to ALA. Would this put suspects like Tumblety, Champan, Bury, James Kelly or Cutbush in more focus?

    Before I go off on ever more of a tangent I would like to ask anyone who reads this a hypothetical question. What would you do if the canonical five were all exhumed, and complete DNA was recovered from all five which matched with suspected modern descendants. On three of the five complete foreign DNA was recovered which was all the same. This DNA was submitted for tracing on a database and came back with a match for Charles Lechmere. I know a scenario like this is completely impossible but my question is, in this situation with overwhelming evidence would you accept the solution?

    Alternatively, if the DNA matched was Michael Ostrog what would reaction be? A suspect thought impossible due to overwhelming information and evidence showing he wasn't JtR actually was. Would you accept it even if all the circumstantial evidence said no, yet the forensic evidence said yes?

    Also a side note, in this impossible hypothetical situation the testing and analysis would be overseen by the MET Police so there wouldn't be any of the nonsense Shawl BS like questionable methods and peer reviewing.

    Whilst this situation is impossible for JtR is certainly is possible and hopefully will happen in the not too distant future for the Zodiac Killer. The conclusion and fallout will definitely be interesting and something to compare with our case of JtR. In the meantime we will have to keep researching in hope of finding the missing piece of the puzzle. But like the location of Donna Lass with the Zodiac. Even if we do finally identify JtR that's only half of the answer, working out the how and why is another great task to investigate, and whilst the Zodiac might still be alive to ask for the answer, JtR is certainly dead so it's up to all of us to solve everything.

    ​​​​​​​Astatine

    Comment


    • #3
      Now we have a thread to argue about arguments.
      My name is Dave. You cannot reach me through Debs email account

      Comment


      • #4
        Once you start defining the parameters of how an argument should be argued we are in the realms of parody, or worse, control of free speech. Argue as you see fit. If the other's side case is strong enough then it will be argued as such. Truth as an odds calculator, whatever next?
        "When the legend becomes fact... print the legend"
        - The Man Who Shot Liberty Valance (1962)

        Comment


        • #5
          Hands up who actually read the OP in full? Be honest...
          Thems the Vagaries.....

          Comment


          • #6
            Sorry, perhaps i should have provided some more info.

            What I'm talking about is an area of philosophy of science called confirmation theory. It's the formal study of how reasonable conclusions are made based upon evidence and so forth.

            What I was outlining is the core ideas, and for those familiar with statistical analysis, corresponds most closely to Bayesian analysis.

            Anyway, for those who are not interested in the logic and reasoning the underlies how evidence is used to confirm or reject theories that's fine. It can be dry unless you're into such things. I know a bit about this area, but it's certainly not my area of speciality.

            If nothing else, some might be interested to know that there are people who make a career out of arguing about how arguments work. It's not about just holding your own and refusing to back down, for example.

            There may be no body keen on looking under the hood at how evidence is used to evaluate an idea, or how arguments work with respect to rational thought. That's cool, it will die on the vine. If anyone is interested, I'm happy to be involved too.

            - Jeff

            Comment


            • #7
              Argument Clinic - Monty Python - The Secret Policeman's Balls - YouTube
              My name is Dave. You cannot reach me through Debs email account

              Comment


              • #8
                Bayes's ideas were popularized by some dude named Pierre.
                My name is Dave. You cannot reach me through Debs email account

                Comment


                • #9
                  Originally posted by Al Bundy's Eyes View Post
                  Hands up who actually read the OP in full? Be honest...
                  Nope.

                  Didn't have any maps.
                  My name is Dave. You cannot reach me through Debs email account

                  Comment


                  • #10
                    Originally posted by erobitha View Post
                    Once you start defining the parameters of how an argument should be argued we are in the realms of parody, or worse, control of free speech. Argue as you see fit. If the other's side case is strong enough then it will be argued as such. Truth as an odds calculator, whatever next?
                    Truth is a binary state, something either is or is not true. The problem is we often cannot know that, but we can work out probabilities. For example, peaked caps get mentioned a lot, but how common were they? If every man in the are wore one, is the fact that a man with a peaked caps seen with stride and with Eddowes meaningful? No, because every man wears one in my example. What if the only time peaked caps appear us those two sightings? Every other male wears a brim hat let's say. Now, our two sightings still could be two different people, but that rare type of headress starts favouring arguments for same killer (and by the way, I'm using that as an example topic, peaked caps were common, but not universal).

                    So it's not truth that is the odds function, but the strength of our belief that one side is true. Where would a rational person distribute their belief, if you will.

                    - Jeff

                    It is about how to evaluate such types of data, and how to weigh up the strength of arguments. Th

                    Comment


                    • #11
                      Ha ha, love that sketch. A buddy and I used to do mp sketches in school variety shows and we did that one. Lots of fun

                      - Jeff

                      Comment


                      • #12
                        Originally posted by DJA View Post
                        Bayes's ideas were popularized by some dude named Pierre.
                        I missed the Pierre years, so I can't be sure, but from what I've seen I'm going to suspect he didn't really know what he was talking about.

                        - Jeff

                        Comment


                        • #13
                          Originally posted by JeffHamm View Post

                          I missed the Pierre years, so I can't be sure, but from what I've seen I'm going to suspect he didn't really know what he was talking about.

                          - Jeff
                          In fact he is highly intelligent.
                          Canadian currently over at Howard's site.
                          My name is Dave. You cannot reach me through Debs email account

                          Comment


                          • #14
                            Originally posted by JeffHamm View Post

                            Truth is a binary state, something either is or is not true.
                            No it'snot
                            My name is Dave. You cannot reach me through Debs email account

                            Comment


                            • #15
                              Originally posted by DJA View Post

                              No it'snot
                              Yes it is.

                              - Jeff

                              Comment

                              Working...
                              X