How feasible is algorithmic trading?

Discussion in 'Automated Trading' started by toli, Feb 28, 2023.

  1. toli

    toli

    So in other words trading on intuition rather than rigorous rules?
     
    #11     Feb 28, 2023
  2. VEGASDESERT

    VEGASDESERT

    no, they are rigorous but not as exact.


    for example, i could see a climax bar form that didn't get to my
    profit target but i'll still exit the trade where an algo may have managed
    the trade to break even as the market pulled back towards entry because the
    profit target was not hit,

    i know that aglos can get massively complex to account for all that but
    like the i said, algos cannot be too simple hence you're back to no free lunch.
     
    Last edited: Feb 28, 2023
    #12     Feb 28, 2023
    MACD likes this.
  3. toli

    toli

    So using indicators instead of algorithms?
     
    #13     Feb 28, 2023
  4. Not exactly. "Using math to predict the market"... no so much. Price TA is a matter of "price pattern recognition".... converting that into math for the computer is waaaaay complex. The problem for "pattern recognition" is that the "start" of the pattern can occur anywhere in the chart. What starting point do you use for a calculation when you don't know where it is until after it's past.. and how do you program that?
     
    Last edited: Feb 28, 2023
    #14     Feb 28, 2023
    SimpleMeLike and MACD like this.
  5. NO! Not trading on intuition... but rather trading on logic and probability.

    Intuition is like "gut feeling", isn't it? I recall a long time successful trader's view on "gut feeling". He said, "when you get a gut feeling about a trade... get up out of your trading chair and proceed to a nearby door... and bang you head against the jamb until the feeling goes away".
     
    #15     Feb 28, 2023
  6. toli

    toli

    Well how do you decide on that as a human?
     
    #16     Feb 28, 2023
    rb7 likes this.
  7. VEGASDESERT

    VEGASDESERT

    remember every smart quant knows what other smart quants are doing, they
    all compete for the same fills and profits hence the algos always need to be changed
    because the need to outsmart each other to get in front of the other algos,
    the game never ends.
     
    #17     Feb 28, 2023
    MACD, Tradess0610 and toli like this.
  8. As a "calculation", I don't think you can. But you can see the trades set up on the chart.
     
    #18     Feb 28, 2023
  9. Bad_Badness

    Bad_Badness

    Algo trading is the new buzzword for retail traders. Like "AI-xyz" or back in the day "TA".

    It is so broad a term that is easily becomes meaningless. If one were to catalog the types of Algo trading it would be VERY large, despite the secrecy behind it. Just like TA.

    Algos are just an extension of TA. Although it is faster, and can eliminate human intervention for order execution. The big difference is people like to conflate order execution and signal generation to make it even more a "silver bullet". This is oversimplifying the situation.

    So it is no more or no less "feasible" in the broad sense than, say TA. It is just another tool and how you use it, how precisely, how accurately, how you maintain it for changing markets, makes the difference between success and churning or even failure.

    Bottom line, it still takes SKILL.
     
    #19     Feb 28, 2023
    MACD, themickey and toli like this.
  10. lariati

    lariati

    I am embarrassed I gave that ridiculous video any clicks. To believe chatGPT is going to write a winning trading strategy is to have not the first clue how chatGPT works. The point of that video is to make money from people clicking on the video. Youtube has really become like a giant flee market of people trying to sell each other cheap plastic tchotchkes.

    There are not all these winning strategies just laying around for you to tweak or for chatGPT to train on.

    Most of the time in this context winning algorithmic trading strategy means really simple losing discretionary strategies that are simple enough to easily write as a script and happen to have lined up with the back testing period as to look like winning strategies.

    An algorithmic strategy like a hedge fund would be something like putting on 5000 short trades and 5000 long trades this week across the entire universe of tradeable instruments. That can't really be done manually without an army of traders. There is no point in automating a moving average crossover on a single instrument when you can just look at the chart and enter the order by hand.
     
    Last edited: Feb 28, 2023
    #20     Feb 28, 2023