A Statistical Method For Stop Placement

About the Author Jeff Swanson

Jeff is the founder of System Trader Success - a website and mission to empowering the retail trader with the proper knowledge and tools to become a profitable trader the world of quantitative/automated trading.

follow me on:
  • BlueHorseshoe says:

    Hi Jeff,

    Each time you run the code a different series of random entries will be generated. Of all the possible outcomes for this system, in terms of adverse excursions, the worst MAE can obviously be found like this:

    vars:x(0),y(0);
    x=o-maxlist(o-l,h-o);
    if x>y then
    y=x;

    Why not just identify a model that describes how excursions from the open are distributed between zero and y, and then situate yourself in terms of this with your ISL?

    I’m not sure how helpful using a random entry actually is here.

    I might be mis-thinking this though!

  • Ilya says:

    What’s the endgame here? Trying to optimize where to put a stop, and then testing it out of sample?

    A couple of things: according to Larry Connors, stops *hurt*. That is, assuming you have an edge, no matter how bad your loss is, given that you’re at a certain unrealized loss L, to get out at that point would be to realize that loss L, when you can let the trade ride and get out at some value x < L (assuming you have an edge). Otherwise, you're going to pay a premium for abandoning your edge. At least as far as I understand it. Basically, he advocates waiting for the trade to come back, no matter what. I'm not sure I agree with him though.

    What I say though is that if your trade goes off the rails the way it does, it may mean that a model is incomplete and that there's a certain property of your trading system not being picked up by the existing indicators, which should give you a more systematic stop loss. EG if you "buy the dip", you probably need some way of identifying when a dip isn't a dip, but instead a trend reversal going against you, and get out ASAP.

    • BlueHorseshoe says:

      Hi Ilya,

      No liquid indexed market (as opposed to a single stock) has ever experienced a continuous parabolic rise or decline. This means that if you buy the dip and get caught in a reversal, the new trend that follows the reversal will also have a pullback, and that the pullback is a better place to exit.

      The “Connors Model” of markets seems to be based on the simple truth that they never move in straight lines for very long.

      Remember, there are other ways of controlling risk besides a hard stop. I trade a variation of the Connors approach, and I control risk by using zero leverage and through position sizing. It just means that I have to look for more opportunities to exercise the strategy in more markets.

      People don’t like being told “de-leverage!”, but it can be a great way to manage risk instead of stop-losses.

      BlueHorseshoe

  • Mark says:

    This article was already posted back in August 2012. Jeff, are you out there or is this website on autopilot these days?

    • I’m still here. I do rotate past articles that I think are interesting. You would be surprised how many people don’t see them until they rotate back to the feature article.

  • Mikey says:

    why does your stop sign only have 6 sides?

  • Lorenzo says:

    Hi Jeff, great tool and idea! You write:We are not done yet on determining our ISL. However, it will have to wait for a future article. Is it available already?

    • I was hoping to see more from the original author so I could translate it to EasyLanguage. But I never saw a follow-up article.

  • >