Effect of automated trading on Markets

Discussion in 'Trading' started by duard, May 10, 2005.

  1. anyone notice how super aggressive the black boxes are this week on the nYSE?
     
    #31     May 12, 2005
  2. aggressive in spurts.......... but with no size.
     
    #32     May 12, 2005
  3. Does chess have the elements of fear and greed or intuition? This is where things aren't as clear in the markets as they are in chess.
     
    #33     May 12, 2005
  4. No, it is not realistic. :)

    I'm actually not speaking of having them reveal their software. I am speaking of requiring certain accounts with machines on them to be declared. Eurex does this. A lot of times, what I find chokes up the market as far as execution and just size bouncing around is autospreaders. When those things kick on, it seems that things in futures slow to a grind.

    When the ES went down the other day and then came back up, it took forever for trades to show up in the dow and I'm at a direct clearing firm AT the CBOT. I guess I don't have a point.
     
    #34     May 12, 2005
  5. EricP

    EricP

    Your point can be taken both ways. I would suggest that computers are NOT subject to being influenced by fear, greed or intuition, and that this is a good thing. Those qualities are not typically associated with profitable trading, and avoiding them is usually best for the trader.

    I think your intention was to point out the human brain might be better able to analyze and capitalize on market action that is caused by the fear and greed of others. This might be the case, but from my experience a well designed computer system can still outperform the human brain acting alone.

    One area where computers will never compete, IMO, is the longer term area of investing. Investing relies on long range market forecasts, earnings assumptions, and many factors to predict the future success of a company relative to it's current share price. I don't think that a computer 'system' is as likely be to as effective for investing as an excellent stock analyst (although I may be wrong). However, as the trading timeframe is reduced, the pendulum swings sharply in favor of automated systems, IMO.

    -Eric
     
    #35     May 12, 2005
  6. Please explain autospreaders. Is that what I am seeing in my post above?

    Geo.
     
    #36     May 12, 2005
  7. I agree. I think that the computers are programmed to attempt the very old practice of front-running. Since you can not see guys in the pit taking orders by phone in the e-markets, you have to design sophisticated algorithms to spot big orders being worked into order flow without showing their true size. If the trader working the order is under time pressure, then the computers make him pay up for liquidity.

    One question that I have is whether the computers are adding to liquidity or just leeching off the market when a fat order arrives. Thus, when the market really needs liquidity to handle a major event, the computers shut down because they do not see big orders that can be front-run with low risk.
     
    #37     May 12, 2005
  8. EricP

    EricP

    I think automated systems are being used for every kind of trading that manual traders are implementing. Momentum, scalping, trend-following, counter trend, etc. If a system can be successfully traded manually, then someone is likely trying to save their time and increase their speed by automating it if possible.

    -Eric
     
    #38     May 12, 2005
  9. I would not be surprised that on a good day a discretionary trader can trade a similar strategy 5-10x better than a system. But I think the Bots will win the battle against human over the long haul due to systems' ability to crank out consistent trades (winners and losers). As quoted in the recent SFO article, Ed Seykota said: "If you can't quantify it, you probably can't manage it."
     
    #39     May 12, 2005
  10. FredBloggs

    FredBloggs Guest

    im not convinced that a machine can beat a trained human brain with several years experience.

    the only advantage a bot has is speed and no emotion.

    no way can a bot be trained to get the feel for a market the same way an intuitive trader can. no way on Gods earth. not even neural networks. in fact, n nets become LETHAL if they are OVER trained or UNDER trained. they are not aware of basics like the larger sentiment picture - the road map of market psychology gained from a chart or t&s.

    bots are binary. they trade the world according to 0 or 1. they can not distinguish all the finer points in the world. (eric - chess is still a binary problem - markets are not)

    besides, none of you seem to have yet asked the questions of what triggers the bots, or what capacity they play in the market. everyone seems filled with some neurosis that they will be beat. the markets grind to a halt i hear you cry! whats new - they always do around lunch time!! - unless you trade fixed income in which case its like the night shift anyway!!!

    you must first distinguish the difference bots play between buy & sell sides of the industry.

    only when you understand them (instead of sitting round worrying) will you be able to beat them - should you decide they are a threat, or are one of these mega competitive types that believe everything should be beat, rather than taking care of no. 1 first regardless of no. 2 is ahead or behind....

    believe it or not, markets still work on the same principals they have worked on for the last 200 or so years. in the next 200 years i dare say they will still be moved on the same principals.
     
    #40     May 12, 2005