setting stop loss point in automated trading?

Discussion in 'Automated Trading' started by elit, Sep 26, 2006.

  1. elit


    How do you decide and set the stop loss point in your automated trading systems?

    Is using a percentage efficient, or do you do any advanced calculations?

    I suppose a good stop loss would be on the point where your trade is proven wrong, i.e. that the technical indicators does not show the signal any more. Although this could often be very far from entry.
  2. cashonly

    cashonly Bright Trading, LLC

    Seeing as how you are automated, you may want to check into volatility based stops. This lets you adjust for changing market conditions and from the studies I've read, helps to keep you in profitable positions longer. While it may have more loss on some trades, the ability to stay in other trades for more profits more than makes up for it. It's well-suited for automated tradiing as the computer is quickly doing the calculations.

    Also, when you're setting stop loss points, particularly for NASDAQ, you definitely want to consider using the NBBO (and taking into account size) as opposed to actual prints because someone getting screwed on an illiquid ECN should not be a reason for you to bail on a position.

  3. elit


    Is a system with exits based on trailing stop loss a good or bad idea?

    And I've heard something about chandelier exits, any input?
  4. Using stop loss orders dramatically icreases your transaction costs in terms of slippage.

    Getting out of a losing position...
    Is precisely the kind of thing...
    That cannot be automated to even approximate what a good trader can do.

    This would be the primary reason why 100% mechanical systems are ** quite sub-optimal **.

    Any intelligently designed ATS...
    Would do the 80% no brainer trades automatically...
    But have an experienced trader handle the 20% trickiest manouvres...
    Especially getting out of a problem in a fast market.
  5. It really depends on your trading style.

    Here's my suggestion.
    - If you have no clue where is the best to place your stop-loss order, simply use percentage stop (if you wish to limit your risk) or volatility stop (if you wish to lower the chance of stopping out prematurely, but it may increase your risk)
    - If you know where to place your stop-loss order, that's where you place your order. This is called a technical stop. For example, based on your experience, you feel that when it hits your stop price, this should indicate the trend ends and you should stop out. That's the whole point of a technical stop.
  6. fader


    in most, if not all, testing i have done, a stop always costs you something, if not in terms of direct profits, at the very least the cost is slippage - so you have on one side the cost of the stop and on the other side you have a reduction in volatility of your returns (assuming that this is what the stop is designed for)

    so you have the typical p&l vs volatility, i.e. risk vs return, tradeoff - i first set the risk/return target parameters for my overall portfolio, then use these metrics to come up with stops for each individual strategy (adjusting for correlations between strategies etc.)

    all the best.
  7. Chabah


    I am starting my own automated system and looking at exactly this question. You might look at "Trade Your Way to Financial Freedom" by Van Tharp, a new edition is coming out in November. I read "Financial Freedom Through Electronic Day Trading" by Van Tharp and Brian June, and while a bit dated the position sizing (which relies on your stop) and money management (again related to your stop) are just fine. Take a look at Trader Mike too, check out Position Sizing.

    For automated ways of setting stops, check out Professional Stock Trading by Mark Conway. I am still reading that one, but one recommendation is to compute the Average True Range (ATR) of the stock and setting your stop equal to one ATR.

    Personally I am more concerned with the proper strategy to update your stop order(s) as the price changes. If I get my stuff together I may post it here as a journal.

    Good luck,