• 520@kbin.social
    link
    fedilink
    arrow-up
    3
    arrow-down
    2
    ·
    10 months ago

    Oh there are practical applications if you know where to look. For instance I’m training a trading AI to spot patterns in numbers and indicators that aren’t necessarily common knowledge.

    • helenslunch@feddit.nl
      link
      fedilink
      arrow-up
      3
      ·
      10 months ago

      if you know where to look.

      I obviously don’t.

      I’m training a trading AI to spot patterns in numbers and indicators that aren’t necessarily common knowledge.

      For what purpose?

      • 520@kbin.social
        link
        fedilink
        arrow-up
        3
        arrow-down
        2
        ·
        edit-2
        10 months ago

        For what purpose?

        The short version: making trading decisions based on technical data.

        The long version: there are traders out there that do their operations using statistical formulas called indicators.

        There are lots of such indicators out there, and hey, it’s pretty easy when they all point one direction, but what happens if they don’t? What happens if you only get a few conclusive indicators and the rest are inconclusive? What if you were to combine specific indicators to get a more conclusive answer?

        If there is one thing AI is good at, it is exploiting numbers and patterns in service of a very specific and easily measurable goal. Mine is trading profits.

        • helenslunch@feddit.nl
          link
          fedilink
          arrow-up
          7
          ·
          10 months ago

          Yeah I don’t buy it. People have been trying to create such algorithms for decades. Problem is no algorithm can account for the complexity and completely illogical decisions of humans.

          • jj4211@lemmy.world
            link
            fedilink
            arrow-up
            3
            ·
            9 months ago

            Also, stock prediction can change the thing being predicted. Once a critucal mass of people believe in a certain prediction strategy, then their own actions screw up the prediction. It’s not like predicting some natural or external thing.

          • 520@kbin.social
            link
            fedilink
            arrow-up
            2
            ·
            10 months ago

            Oh yeah, this thing will never be able to predict events like GME or that kind of thing.

            The thing is though, I don’t have to beat the first couple or even the first 50 trades. I just have to beat the babdwagoning masses by detecting upticks and downward trends, something my software can do and act upon faster than any human.