Will algos&AI take over? Is discretionary trading coming to an end?

Discussion in 'Trading' started by orbit23, Sep 20, 2019.

  1. vanv0029

    vanv0029

    It is not true that chess programs can beat the very best human players. Programs run into John Lighthill's combinatorial explosion. You are believing the GOOG hype. Read Leonard Barden's FT chess columns. Also read Gary Gasparov's "Deep Thinking" book. Kasparov claims that the well financed IBM chess group used psychological war fare against him. Now billionaire sinquefield support of chess canceled IBMs huge budget advantage including development of programs that experts use to analyze positions.
     
    #21     Sep 22, 2019
  2. vanv0029

    vanv0029

    Also, AI (aka HFT) is cheating mostly front running that is only allowed because of unwillingless of the SEC to enforce the law. If Warren is elected, no one will be able to remember how AI trading worked at all.
     
    #22     Sep 22, 2019
    ElectricSavant likes this.
  3. wrbtrader

    wrbtrader

    https://www.elitetrader.com/et/threads/a-message-to-some-day-traders.336101/page-32#post-4930358

    They've already taken over the markets. In fact, anything that's not algorithm...its viewed by the financial institutions as discretionary trading.

    Prior to AI...discretionary trading was viewed as intuition trading...trading from the seat of your pants as in gambling.

    You can adapt via developing your own algorithms or try hard not to play in the typical playgrounds that algorithms are playing within.

    wrbtrader
     
    #23     Sep 22, 2019
    Orbiter likes this.
  4. wrbtrader,

    How does a person determine the "typical playgrounds that algorithms are playing within." ?
     
    #24     Sep 22, 2019
  5. Real Money

    Real Money

    I love the fact that most of the trading is robots and algorithms. It's like playing a video game where the opponent is stupid bots. It's is easy to outsmart AI and take advantage of them. But, nobody is talking about it cuz it's a big secret.

    Here is a clue.

    It's estimated that up to 50-80% of trading on NYSE is algorithms on any given day. Globex is in Illinois and stocks are in NYC. The exchanges are linked via microwave towers that transmit and receive orders between the exchanges at 99% of the speed of light.

    If the ES contract trades something like 1.6 m volume in a day, that's $240 billion dollars.

    Colocated HFT firms use this infrastructure (100m dollars a month) to capture arbitrage opportunity between the exchanges. This is called cash and carry and reverse cash and carry arbitrage. This arb is given to the HFT program trade bots!

    The point is that Morgan, Goldman, Barclays, UBS and rest of them (dozu's pro boyz) are taking advantage of this. And I mean really taking advantage of this.

    If they want to, they can move the market any way they want just by adjusting the cash vs futures spread. This is just the cost of doing business to these guys.

    People need to wake up.
     
    #25     Sep 22, 2019
  6. Real Money,

    I have one question for you please:

    1. How does this help the average trader sitting at home trading by his/her self make money?

    Thanks,
     
    Last edited: Sep 22, 2019
    #26     Sep 22, 2019
    Overnight likes this.
  7. Real Money

    Real Money

    Simple. You get a software that can chart the spread. Then you do all of your analysis on that because it is the market now. Put ES and the spread on the same chart.

    It tells you what you need to know to trade ES all day long.

    (there is some other stuff I'm keeping secret, for now).

    PREM_CHART.png
     
    #27     Sep 22, 2019
  8. tommcginnis

    tommcginnis

    Truly, if there is a sign of abject idiocy keeping hyper-intelligent space aliens from contacting our still-knuckle-dragging human culture, this statement has *got* to be right up there. :wtf: "Beam me up, Scotty -- there's no intelligent life down here." :cool:

    Funny but, I have never read this statement from anyone who was qualified to teach *anything*. :confused: You might want to consider that. :rolleyes:
     
    #28     Sep 22, 2019
  9. IAS_LLC

    IAS_LLC

    I was joking... But I did teach Aircraft Design and Wind Tunnel Aerodynamics while in grad school
     
    #29     Sep 22, 2019
    tommcginnis likes this.
  10. tommcginnis

    tommcginnis

    It's plain in reading things over that 3/4s of respondents to this thread have no idea what "AI" or Machine Learning are -- that's typical across the 'tech front' these days. Regardless of having been around for decades, there are basic tenets that remain unknown. (But, having software ask a question does *not* make it "AI"; having software make it's own range of predictions does *not* make it Machine Learning.)

    In any event, there is routine alpha to be granted to smaller algos that is of zero interest to parties wishing to efficiently, effectively, deploy hundreds and hundreds and hundreds of millions of dollars. There is stuff to be washed in-and-out-of, in less time than it takes to write it or read it: HFT and all that. And there is stuff that takes a week-to-10-days to enter or exit. But between those to high speed or high wealth playgrounds, is lots and lots of swing time.

    Will that alpha-adding swing-time be there tomorrow? I don't see why it should be. Algo design and deployment will have gotten *more* effective than the dribs and drabs by which it is used/profited now. With that drop in fixed cost will come a ramp-up in incidence, and a concomitant decline in required AUM to make it shine.

    But it will be YEARS away. 10 at least; 20 years easy. In the meantime, hunt for the alpha while it's yet available.
     
    #30     Sep 22, 2019
    IAS_LLC likes this.