Loading ...
Sorry, an error occurred while loading the content.

Use and abuse of statistics

Expand Messages
  • eyeofhoor <eyeofhoor@yahoo.com>
    In response to the statistical calculations made in the essay ... Odds in the millions can hardly be called astronomical, but are significant. The original
    Message 1 of 8 , Jan 2, 2003
    View Source
    • 0 Attachment
      In response to the statistical calculations made in the essay
      "Signature of the Gods" (Post # 12599) Randall Hobart wrote:

      > With regard to your statistics, the prior discussion of which I
      > have excised from my response, all I'm going to say is that I
      > maintain my position that they're complete rubbish. You find
      > something that you consider significant, calculate the odds, and
      > then call it statistically remarkable because the odds are
      > astronomical.

      Odds in the millions can hardly be called astronomical, but are
      significant. The original work done with the Bible Codes by Jewish
      scholars is claimed to defy odds in the millions; statisticians
      disagreed with the odds, claiming the number is lower, while
      understanding that odds in the millions *are* significant.

      > Maybe you're calculating the odds correctly, and maybe you aren't
      > -- I don't really care.

      That statement speaks volumes about your qualifications for
      formulating an informed response to my work.




      > The point is that extraordinary odds are meaningless in the context
      > you're using them.

      Your previous statement indicates a lack of qualification on your
      part for determining whether the odds figures in question are
      meaningless or not. You can remove your foot from your mouth whenever
      it suits you.

      There is a popular trend among skeptics in which statistical
      calculations of all sorts are viewed as meaningless. For instance, in
      the Skeptic's Dictionary, one argument used against the suitability
      of odds for making valid determinations is the lottery. The claim is
      that even though odds of winning the lottery are in the millions,
      someone eventually wins, thereby demonstrating that odds in the
      millions are meaningless. The fact is that if there 20 million sets
      of numbers possible and 10 million tickets are sold, the odds of
      someone winning the lottery are *even*.


      > If one flips a coin 100 times, the odds of obtaining *any*
      > particular sequence are equally astronomically unlikely, but the
      > sequence one obtains is only remarkable if one predicted it
      > beforehand. One is sure to get *one* of the bazillions of
      > possible sequences. All of your odds are calculated after the fact.

      Your analogy is simplistic and irrelevant to the work presented.
      Here is a proper coin-toss analogy. Imagine 448 buckets grouped
      together. Blindfolded and walking in a circle around the buckets, you
      would have toss eight coins individually at random intervals at the
      buckets and land three in one bucket in order to match the
      statistical feat the alpha twins accomplish as documented in my
      essay.

      > Even if the odds are calculated accurately, it's a misapplication
      > of the technique to say that because something had odds of 1 in 2
      > billion of occurring, you've obtained an interesting or remarkable
      > result.

      You position is ludicrous. Ask a statistician.

      > The outstanding feature of your work, in fact, is that you've
      > devised many systems, all of which have produced these same kinds
      > of odds.

      Your assertion is demonstrably false. As a gematria system, the
      Theban system has had several names, but is the same system I
      introduced in alt.magick in September 1999, and the same system used
      in my contributions to the Holycram archives, and the only system
      that I've used in statistical calculations.

      > The lesson to take from this is that you can find the same kinds of
      > odds in *any* system if you keep looking, *especially* if you have
      > a computer and can easily investigate all kinds of obscure textual
      > features that you couldn't otherwise.

      I had no gematria software available for use when I devised the
      gematria system called the Theban system. Between 1982 and 1992 I
      wrote a number of gematria programs on the original 8-bit machines.
      When my last disk drive died so did all of my software. I scored my
      first gematria system for the IBM computer I use now from Robin's
      Mystical Internet in December Of 1999. Upon adding up the value of
      the words to the left of the line drawn on the grid page and
      discovering its enumeration to be 9797. I ran search for the
      enumeration using Robin's 'Enuminator' program and discovered the
      original alpha twin. In order to find the reest of them I had to write
      M.A.G.I. Your inference that I've scanned the text for unusual
      patterns for the purpose of claiming a remarkable discovery is also
      demonstrably false.

      You stated in a previous post that any pet system could equal the
      results I've obtained with my system. I've designed my software to
      prove whether this is true or not regardless of the outcome. I accept
      your challenge or anyone else's. While the chances of another
      system producing similar results to my system are very slim, I can
      calculate the odds of any system being able to do so by having the
      program tally the number of alpha twins present in the text using any
      system.

      Prophet 718
    • eyeofhoor <eyeofhoor@yahoo.com>
      ... you ... I was rushed when I started this thread and had no time to review the post. The correct number of buckets in the coin-toss analogy is 112, not 448.
      Message 2 of 8 , Jan 2, 2003
      View Source
      • 0 Attachment
        >
        > Your analogy is simplistic and irrelevant to the work presented.
        > Here is a proper coin-toss analogy. Imagine 448 buckets grouped
        > together. Blindfolded and walking in a circle around the buckets,
        you
        > would have toss eight coins individually at random intervals at the
        > buckets and land three in one bucket in order to match the
        > statistical feat the alpha twins accomplish as documented in my
        > essay.

        I was rushed when I started this thread and had no time to review
        the post. The correct number of buckets in the coin-toss analogy is
        112, not 448. The chances of successfully landing three coins in one
        bucket is about one in 4000. For those who do not believe odds in the
        thousands are meaningful, you should try the coin toss. Only one
        attempt is allowed.

        The odds calculations presented in "Signature of the Gods" are
        valid, but not the ultimate evidence of my claims. The ultimate
        evidence to be presented is the sweep of *all* of the cryptic verses
        in the book by presenting *objective* explanations for them,
        including a precise explanation of verses II: 15 - 16 and the
        perplexing reference to The Empress & the Hierophant.

        Prophet 718
      • Rikb
        93! ... You only think so because you re too obtuse to grasp the point I m making. The magnitude of the odds is irrelevant. They could be in the thousands or
        Message 3 of 8 , Jan 2, 2003
        View Source
        • 0 Attachment
          93!

          > > Maybe you're calculating the odds correctly, and maybe you aren't
          > > -- I don't really care.
          >
          > That statement speaks volumes about your qualifications for
          > formulating an informed response to my work.

          You only think so because you're too obtuse to grasp the point I'm
          making. The magnitude of the odds is irrelevant. They could be in the
          thousands or in the trillions, but if you're abusing the technique in the
          first place it doesn't matter. My objection isn't to the precise odds you
          find -- they could be perfectly correct and it wouldn't make them any more
          relevant to the issue of statistical significance. I have no interest in
          checking your arithmetic when the calculations you're doing aren't even
          applicable to what you're trying to demonstrate.

          > There is a popular trend among skeptics in which statistical
          > calculations of all sorts are viewed as meaningless. For instance, in
          > the Skeptic's Dictionary, one argument used against the suitability
          > of odds for making valid determinations is the lottery. The claim is
          > that even though odds of winning the lottery are in the millions,
          > someone eventually wins, thereby demonstrating that odds in the
          > millions are meaningless. The fact is that if there 20 million sets
          > of numbers possible and 10 million tickets are sold, the odds of
          > someone winning the lottery are *even*.

          Odds of one particular person winning are irrelevant to the odds of
          *someone* winning. If you correctly predict that one particular person --
          Joe Smith in Provo, Utah -- will win the lottery, you've done something
          pretty remarkable. If you correctly predict that *someone* -- without
          specifying who -- will win the lottery, it's not at all remarkable because
          we know that someone will win eventually. The odds of the first happening
          say nothing about the odds of the second happening. The problem is, people
          conflate the two. You'd like to think that your calculations are like the
          first example, when they're really more like the second. It's easy to find
          *some* event with odds that would be fantastic had you predicted the same
          precise event.

          > > If one flips a coin 100 times, the odds of obtaining *any*
          > > particular sequence are equally astronomically unlikely, but the
          > > sequence one obtains is only remarkable if one predicted it
          > > beforehand. One is sure to get *one* of the bazillions of
          > > possible sequences. All of your odds are calculated after the fact.
          >
          > Your analogy is simplistic and irrelevant to the work presented.
          > Here is a proper coin-toss analogy. Imagine 448 buckets grouped
          > together. Blindfolded and walking in a circle around the buckets, you
          > would have toss eight coins individually at random intervals at the
          > buckets and land three in one bucket in order to match the
          > statistical feat the alpha twins accomplish as documented in my
          > essay.

          The task as presented here is *fundamentally* different from the case
          of your "accomplishment." In the case you present above, you specify
          beforehand a particular domain and the precise outcome that one must
          achieve. In fact, all one would have to do is find *some* circumstance with
          the same odds, not the exact one that you specify. That's *much* easier. I'm
          sure there are enough circumstances in Liber AL with 1 in 4000 odds and
          greater, using whatever gematria system you care to choose, that anyone
          could find a few given a little time and facility with numbers -- it just
          depends on what kind of number relationships you decide are important (which
          of course you are free to specify arbitrarily after the fact, invent new
          ones, etc.).

          >
          > > Even if the odds are calculated accurately, it's a misapplication
          > > of the technique to say that because something had odds of 1 in 2
          > > billion of occurring, you've obtained an interesting or remarkable
          > > result.
          >
          > You position is ludicrous. Ask a statistician.

          I've taught statistics, including the kinds of odds calculations you use, at
          the university level and never got less than an "A" grade in my own graduate
          level statistics classes. How many statisticians have endorsed your methods
          and conclusions?

          >
          > > The outstanding feature of your work, in fact, is that you've
          > > devised many systems, all of which have produced these same kinds
          > > of odds.
          >
          > Your assertion is demonstrably false. As a gematria system, the
          > Theban system has had several names, but is the same system I
          > introduced in alt.magick in September 1999, and the same system used
          > in my contributions to the Holycram archives, and the only system
          > that I've used in statistical calculations.

          Granted. How inconvenient of you to keep changing the name ;-) In any
          case, I think that if you tried using another system, you'd find that you'd
          see the same odds. It wouldn't necessarily be with "alpha twins," but you'd
          find something.

          > You stated in a previous post that any pet system could equal the
          > results I've obtained with my system. I've designed my software to
          > prove whether this is true or not regardless of the outcome.
          > I accept your challenge or anyone else's. While the chances of another
          > system producing similar results to my system are very slim, I can
          > calculate the odds of any system being able to do so by having the
          > program tally the number of alpha twins present in the text using any
          > system.

          But again, one would not have to do this to equal your so-called
          accomplishment. All they would have to do is find *some* feature of the text
          that had a 1 in 4000 chance of occurring randomly. You're constraining the
          conditions of "success" in a way that you yourself weren't constrained in
          your initial search. Here's another analogy. Say you dial a random phone
          number, and that the person who answers the phone happens to be a 6'2" tall
          blond woman wearing a green mohair sweater and posessing a mole 2.4
          centimeters from her hairline above her left eye. You can't say that you're
          special or that something earthshattering has happened because no other
          person is likely to get the same result in a lifetime of making random
          calls. The odds against that *particular* result were perhaps 6 billion to
          one, but that's irrelevant to the odds of reaching an unique individual,
          which anyone would have done.

          93 93/93
          RIKB
          http://www.horusset.com/RIKB
          ---
          Outgoing mail is certified Virus Free.
          Checked by AVG anti-virus system (http://www.grisoft.com).
          Version: 6.0.434 / Virus Database: 243 - Release Date: 12/25/2002
        • paulrhume <paulhume@mailsnare.net>
          ... 6 2 tall ... You could tell all that over the phone? Egad Holmes, how do you do it?!
          Message 4 of 8 , Jan 3, 2003
          View Source
          • 0 Attachment
            > Here's another analogy. Say you dial a random phone
            > number, and that the person who answers the phone happens to be a
            6'2" tall
            > blond woman wearing a green mohair sweater and posessing a mole 2.4
            > centimeters from her hairline above her left eye.

            You could tell all that over the phone? Egad Holmes, how do you do
            it?!
          • rikb
            On Fri, 03 Jan 2003 15:53:17 -0000, paulrhume wrote ... LOL. This would be assuming, of course, that one asked questions of the person on the other end of the
            Message 5 of 8 , Jan 3, 2003
            View Source
            • 0 Attachment
              On Fri, 03 Jan 2003 15:53:17 -0000, paulrhume wrote
              > > Here's another analogy. Say you dial a random phone
              > > number, and that the person who answers the phone happens to be a
              > 6'2" tall
              > > blond woman wearing a green mohair sweater and posessing a mole 2.4
              > > centimeters from her hairline above her left eye.
              >
              > You could tell all that over the phone? Egad Holmes, how do you do
              > it?!

              LOL. This would be assuming, of course, that one asked questions of the
              person on the other end of the line, as one would "ask questions" of a text
              by investigating numerical patterns. But this does point up an important
              aspect of the issue. One is perfectly free to ask any question whatsoever, so
              between common sense and trial and error, one is likely to find a small set
              of attributes rather quickly that designate the person as very uncommon if
              not unique, in *some* sense. One of the reasons the Bible Codes have been
              firmly consigned to the trash heap of millenial silliness is that skeptics
              were able to demonstrate the relative ease with which one could get very "low-
              probability results" from any text of sufficient length by massaging the data
              and asking their "questions" in a variety of ways (altering spelling for
              example, which is quite variable in Hebrew anyway, adding or taking away
              articles and prepositions, etc), and then showing that Bible Code proponents
              had done the same thing in order to get their results.

              I should mention, since I am making myself out to be the bastard skeptic of
              the hour, that I do not deny the value or utility of doing literal-qabalistic
              work -- I quite often engage in it myself. I am only making the case that its
              value and utility are not due to the degree of statistical improbability one
              can calculate for one's results.

              --
              93 93/93
              RIKB
              http://www.horusset.com/RIKB
            • eyeofhoor <eyeofhoor@yahoo.com>
              ... aren t ... Ad Hominem. ... The magnitude of odds is irrelevant to probability calculations? You re making no sense. Certainly one can calculate fantastic
              Message 6 of 8 , Jan 5, 2003
              View Source
              • 0 Attachment
                --- In thelema93-l@yahoogroups.com, "Rikb" <rikb@h...> wrote:
                > 93!
                >
                > > > Maybe you're calculating the odds correctly, and maybe you
                aren't
                > > > -- I don't really care.
                > >
                > > That statement speaks volumes about your qualifications for
                > > formulating an informed response to my work.
                >
                > You only think so because you're too obtuse to grasp the point
                > I'm making.

                Ad Hominem.

                > The magnitude of the odds is irrelevant.

                The magnitude of odds is irrelevant to probability calculations?
                You're making no sense. Certainly one can calculate fantastic and
                meaningless odds using events and factors that are unrelated, which
                is the nature of the criticisms you're directing at my work, and in
                error at that. The odds are calculated much like lottery odds are and
                are applied to English sentences that are a single and continous
                construct that express the idea that secrets are hidden within them.

                There is a chance the numeric sequences noted are random, but the
                probability calculations indicate they are probabably not random.

                snipped.


                > In fact, all one would have to do is find *some* circumstance with
                > the same odds, not the exact one that you specify. That's *much*
                > easier. I'm sure there are enough circumstances in Liber AL with 1
                > in 4000 odds and greater, using whatever gematria system you care
                > to choose, that anyone could find a few given a little time and
                > facility with numbers -- it just depends on what kind of number
                > relationships you decide are important (which of course you are
                free
                > to specify arbitrarily after the fact, invent new ones, etc.).

                There are no other circumstances that can be demonstrated through
                gematria provided the application is typical and used in conjunction
                with continuous sequences of words. If you choose to change the
                inherent specifications of the exercise then you can disprove
                anything. I would like to know what other exercises there are in
                gematria usage that can produce similar results. It's easy enough to
                claim my data is meaningless, but until you can disprove it (I've
                already tried) your talk is meaningless. I'm willing to conspire
                against myself. Name a scenario that you think can compete with the
                numbers I've generated and I'll pursue it and provide the results.


                Prophet 718
              • Rikb
                93! ... The magnitude of odds is irrelevant if you re calculating the *wrong* odds. I couldn t put it more clearly than in my last example -- the one you have
                Message 7 of 8 , Jan 5, 2003
                View Source
                • 0 Attachment
                  93!

                  >
                  > > The magnitude of the odds is irrelevant.
                  >
                  > The magnitude of odds is irrelevant to probability calculations?
                  > You're making no sense.

                  The magnitude of odds is irrelevant if you're calculating the *wrong*
                  odds. I couldn't put it more clearly than in my last example -- the one you
                  have chosen to ignore commenting on.

                  > Certainly one can calculate fantastic and
                  > meaningless odds using events and factors that are unrelated, which
                  > is the nature of the criticisms you're directing at my work, and in
                  > error at that.

                  I don't know where you're getting that. The nature of my criticism is
                  just that you're confused about how to interpret odds calculations.

                  > > In fact, all one would have to do is find *some* circumstance with
                  > > the same odds, not the exact one that you specify. That's *much*
                  > > easier. I'm sure there are enough circumstances in Liber AL with 1
                  > > in 4000 odds and greater, using whatever gematria system you care
                  > > to choose, that anyone could find a few given a little time and
                  > > facility with numbers -- it just depends on what kind of number
                  > > relationships you decide are important (which of course you are
                  > free
                  > > to specify arbitrarily after the fact, invent new ones, etc.).
                  >
                  > It's easy enough to
                  > claim my data is meaningless, but until you can disprove it (I've
                  > already tried) your talk is meaningless.

                  There's nothing to disprove if you've drawn your conclusion through the
                  misuse of statistics. If you can find a professional statistician who is
                  willing to state that I'm wrong, and that the way you've used statistics is
                  valid and your conclusions justified given those statistics, then I'll
                  reconsider your "proof."

                  93 93/93
                  RIKB
                  http://www.horusset.com/RIKB

                  ---
                  Outgoing mail is certified Virus Free.
                  Checked by AVG anti-virus system (http://www.grisoft.com).
                  Version: 6.0.435 / Virus Database: 244 - Release Date: 12/30/2002
                • John Bell
                  Do what you want. There are three kinds of lies: Lies, Damned Lies, and Statistics. Mark Twain Want what you do. Bliss: Alamantra What do you want? ... From:
                  Message 8 of 8 , Jan 5, 2003
                  View Source
                  • 0 Attachment
                    Do what you want.

                    "There are three kinds of lies: Lies, Damned Lies, and Statistics." Mark Twain

                    Want what you do.

                    Bliss:
                    Alamantra

                    What do you want?

                    ----- Original Message -----
                    From: "Rikb" <rikb@...>
                    Date: Sun, 5 Jan 2003 15:37:00 -0500
                    To: <thelema93-l@yahoogroups.com>
                    Subject: RE: [t93] Use and abuse of statistics

                    > 93!
                    >
                    > >
                    > > > The magnitude of the odds is irrelevant.

                    --
                    __________________________________________________________
                    Sign-up for your own FREE Personalized E-mail at Mail.com
                    http://www.mail.com/?sr=signup

                    Meet Singles
                    http://corp.mail.com/lavalife
                  Your message has been successfully submitted and would be delivered to recipients shortly.