I'd like to hear some ideas from traders who are using fully-automated systems regarding the criteria you use to expire entry signals. Currently, I'm manually trading an intraday system that's under development. It generates signals for approximately twenty different futures products, with trades lasting from a couple of minutes to a few hours. About half of the time the market doesn't reach my specified entry price for a given trade before it trades through the first profit target. However, I've been surprised at the fact that many of the entry signals remain valid one or two days after they're initially generated (by "valid" I mean that price returns to the entry level at a later time and then reaches the initial profit target again without hitting the stop loss level). Intuitively, this makes sense as areas of supply/demand imbalances can hang around until a shift in the fundamental landscape invalidates them. In practice the profit and stop levels should be re-adjusted based on the volatility in the market at the time the position is actually opened rather than the sticking with the original levels suggested when the signal was generated. As I trade the system manually, I can make decisions in real-time regarding the validity of a particular trade with respect to the market fundamentals, but obviously an automated system can't make such subjective assessments. Expiring a signal after a pre-determined period of time seems somewhat arbitrary, but having a bunch of standing entry points hanging around seems silly. One idea I've had is retaining signals as long as the amount of volatility at the time of prospective entry is the same or less than when the signal was generated, and provided there isn't a trade in the opposite direction open when the entry is reached (i.e., it's not a runaway train in the wrong direction). Any helpful suggestions are welcome. Thank you. Regards, William
Perhaps you could set a time limit based on the amount of volitility, e.g. a longer time to expire on securities that have higher volitilities since it has a greater chance of reterning to the entry level. Another suggestion would be to set a time limit based on the predicted length of the trade, if it is a 1 hour trade, only let the order stay valid for maybe 5-10 minutes. Also, you could cancel the order after it trades through your first profit target, and if it returns later, resubmit the order if the system determine's its still valid. The hard part of course is determining whether its still valid or not. Hope that helps. Paccc