• Aceticon@lemmy.world
    link
    fedilink
    arrow-up
    70
    ·
    8 months ago

    Neatly showing why when all you have is two data points you can’t just assume the best fit function for extrapolation is a linear one.

    Mind you, a surprisingly large number of political comments is anchored in exactly that logic.

      • Aceticon@lemmy.world
        link
        fedilink
        arrow-up
        19
        ·
        edit-2
        8 months ago

        Good point and well spotted!

        PS: Though it’s not actually called exponential (as it isn’t enr-3-month-periods but rather 2nr-3-month-periods ) but has a different name which I can’t recall anymore.

        PPS: Found it - it’s a “geometric progression”.

        • wischi@programming.dev
          link
          fedilink
          arrow-up
          15
          ·
          edit-2
          8 months ago

          By tweaking a few parameters you can turn every base into any other base for exponentials. Just use e^(ln(b)*x)

          PS: The formula here would be e^(ln(2)/3*X) and x is the number of months. So the behavior it’s exponential in nature.

          • Aceticon@lemmy.world
            link
            fedilink
            arrow-up
            1
            ·
            edit-2
            8 months ago

            By that definition you can turn any linear function a * x + b, “exponential” by making it e^ln(a*x +b) even though it’s actually linear (you can do it to anything, including sin() or even ln() itself, which would make per that definition the inverse of exponential “exponential”).

            Essentially you’re just doing f(f-1(g(x))) and then saying “f(m) is em so if I make m = ln(g(x)) then g(x) is exponential”

            Also the correct formula in your example would be e^(ln(2)*X/3) since the original formula if X denotes months is 2X/3

        • some_guy@lemmy.sdf.org
          link
          fedilink
          arrow-up
          3
          ·
          8 months ago

          PPS: Found it - it’s a “geometric progression”.

          A terminology that I learned from the Terminator 2 movie. Only that was, I think, a “geometric rate”.