an ai hot take with no evidence to support it

Поділитися
Вставка
  • Опубліковано 7 лис 2024

КОМЕНТАРІ • 21

  • @alexanderverhaeghe4484
    @alexanderverhaeghe4484 2 місяці тому +1

    Nice to see your perspective on this topic!

  • @HitAndMissLab
    @HitAndMissLab Місяць тому +1

    Great channel by the way!
    I guess it is "rubbish in rubissh out" theorem. What makes you sure that you are giving ML actually predictive inputs.
    Larry Williams published book back in late 90's with one pattern (forgot it's name) and YT is full of videos that are testing that pattern and it still works x24 years after it was published. So, patterns are out there. Maybe its the way ML "thingks" that prevents it from finding them.

  • @anangelsdiaries
    @anangelsdiaries Місяць тому +1

    Actually pretty fair, I am learning algotrading currently and the idea of using ML raw as an automated trading system just always sounded wrong. I know people overfit backtests with finetuning for regular strategies so using ML would just compound the problem because ML models are typically pretty hard to interpret.
    But it feels like if used for sentiment analysis, or as some other cog in the overall system. ot would make sense? Maybe ML based position sizing? Idk, I am still learning all that stuff.

    • @ethernetwink7230
      @ethernetwink7230  Місяць тому

      @@anangelsdiaries i think ML will be a standard for classification in the future. Things like K means and PCA are proven and awesome. But like you said, overfitting a trading strategy make it a pain to understand what is going on

  • @MLirola
    @MLirola 2 місяці тому +1

    Hey!
    I've seen a few videos of you and the content is amazing.
    I'm from spain and i'm learning quant strategies on my own. I would like to chat about some strategies and code if that's fine to you.
    Cheers.

    • @ethernetwink7230
      @ethernetwink7230  2 місяці тому

      @@MLirola what do you want to talk on?

    • @MLirola
      @MLirola 2 місяці тому

      @@ethernetwink7230 Just how you're developing yourself, in terms of math, strategies, websites to seek information and knowledge and also to increase my network. I would like to meet people that are in the same path as me.

    • @MLirola
      @MLirola 2 місяці тому

      @@ethernetwink7230 Just to see how you're developing yourself in terms of math and strategies. I'm interested on profiles (twitter) and sites to seek information (i have an entire folder just with things about quants) and of course to expand my network, i would like to meet people that are in the same path as i am.

  • @hackerborabora7212
    @hackerborabora7212 2 місяці тому +2

    Try nbeat or Nhit model it's good with time series data try put ratio and volatility and see if work with you don't complex problem just do it can't FORCAST no one can we just find

    • @ethernetwink7230
      @ethernetwink7230  2 місяці тому

      @@hackerborabora7212 if we’re not forecasting what are we trying to accomplish?

    • @hackerborabora7212
      @hackerborabora7212 2 місяці тому

      @@ethernetwink7230 hhhh sorry I'm beginner 😆😆

    • @hackerborabora7212
      @hackerborabora7212 2 місяці тому +1

      It's just mathematics and a kind of striving for a living. This is life. You search for something and believe in it, and it comes true. But not always, because God's laws are above everything.

    • @ethernetwink7230
      @ethernetwink7230  2 місяці тому

      @@hackerborabora7212 Amen

    • @HitAndMissLab
      @HitAndMissLab Місяць тому +1

      What timeframe are we talking about, one minute, five minute or one day?

  • @MLirola
    @MLirola 2 місяці тому +1

    I don't know why it doesn't let me reply you back.
    I'm just interested on how you're developing yourself in terms of math and strategies. I'm also interested in profiles that share valuable information (twitter) and sites to seek information. And of course, i would like to expand my network and meet people that are in the same path as i am.

    • @ethernetwink7230
      @ethernetwink7230  2 місяці тому

      Find someone who compliments your thinking. I have a mentor who's an amazing trader while I am more versed on the math and programming side. So we work really well together.
      Honestly I'm not sure who is good to follow on twitter besides myself. All jokes I mostly post Bible verses and myself playing guitar on twitter
      I would read The Misbehavior of Markets by Benoit Mandelbrot, and other things from people like Nassim Nicholas Taleb.
      Also find a niche you like, maybe order flow, options, futures price action whatever and dive super deep into it. I love fractals, volatility, and options.
      Build what you want to! If you find a way to find a good network let me know HAHA

    • @shoto4352
      @shoto4352 2 місяці тому

      @@ethernetwink7230 do you have any ideas for other niches, because I already know what you've mentioned and I'm aiming to automate strategies with them, but if you knew of other niches that would be great. + Feel like creating a discord?