Skip to main content

View Diary: The Black Swan Myth (154 comments)

Comment Preferences

  •  "Black Swan" = "I don't have to analyze" (6+ / 0-)

    And contra the comment upthread, I've had it thrown at me twice in the last week, when I have simply posted data (on initial jobless claims).

    It is a way of saying, "I don't have to pay attention to, or analyze the data, becaue imo this is 'black swan' and so beyond analysis."

    In short, hysteria flung like s*** at reason.

    "When the going gets tough, the tough get 'too big to fail'."

    by New Deal democrat on Thu Jul 16, 2009 at 11:44:35 AM PDT

    [ Parent ]

    •  Au contraire (2+ / 0-)
      Recommended by:
      melo, ericy

      If you have a model that tells you some event should occur every hundred years, and it occurs every 20 years, your model is wrong.  This is the central point to Taleb's thesis- the models in use severely underestimate the probability of massive losses, and the use of these bad models only exacerbates the losses.  The banks point to these models in good times and say "We aren't taking any risk", and load up their bets.  Then when the bad event occurs, they go to the government hat in had asking for a bailout, mumbling "Nobody saw this coming".

      Data should inform the models, not ideology or wishful thinking.

      "Competent statisticians will be the front line troops in our war for survival..." George Box, 1976

      by aztecraingod on Thu Jul 16, 2009 at 12:07:15 PM PDT

      [ Parent ]

      •  The problem is ... (0+ / 0-)

        ... people are using the term to mean that analysis of the data is impossible.

        In fact, we do have data from the Great Depression era and before:  we have monthly commodity and consumer inflation data, we have monthly money data, we have yield curve data, we have daily/weekly stock and bond market data, we have some annual statistics for wages, housing and unemployment.

        While the data set isn't great, it is there and some useful analysis can be done -- even if the paradigm is the Great Depression.

        "When the going gets tough, the tough get 'too big to fail'."

        by New Deal democrat on Thu Jul 16, 2009 at 12:10:58 PM PDT

        [ Parent ]

        •  I'll say it bluntly then (1+ / 0-)
          Recommended by:
          New Deal democrat

          if people mean that a black swan means analysis of the data is impossible, those people are idiots.

          What I am saying, and this was my area of research, is that if you are using a model that makes some assumptions about the data you are making inferences on, and those assumptions are demonstrably false, it shouldn't come as a surprise when the conclusions offered by the modeling are worthless (or worse).

          The solution is to use more robust methods, methods that do not assume gaussian returns, that incorporate correlation structures that are robust to heavy tailed distributions, and generally speaking, impose fewer assumptions.

          The problem is that they don't get into this stuff in any detail until you go into graduate school in statistics (and then only if you have an open-minded professor).

          "Competent statisticians will be the front line troops in our war for survival..." George Box, 1976

          by aztecraingod on Thu Jul 16, 2009 at 12:25:20 PM PDT

          [ Parent ]

Subscribe or Donate to support Daily Kos.

  • Recommended (152)
  • Community (68)
  • Elections (42)
  • Environment (41)
  • Bernie Sanders (41)
  • 2016 (39)
  • Hillary Clinton (34)
  • Republicans (31)
  • Climate Change (31)
  • Culture (31)
  • Media (30)
  • Spam (28)
  • Labor (24)
  • Education (24)
  • Congress (23)
  • Science (23)
  • Civil Rights (23)
  • Barack Obama (22)
  • Law (21)
  • Trans-Pacific Partnership (21)
  • Click here for the mobile view of the site