One of the weakest parameters of some trading systems is the stop loss. It serves as an emergency exit when trade goes in wrong direction and losses get too high. Trailing stops diminish returns as well and I've never seen a system where they are a profitable part of a mechanical strategy. For some time I've been experimenting with a clumsy dynamic stop loss calculated on average true range. Basically, if a trade goes in the wrong direction and it appears it won't come back (and no reverse trade has been triggered), the stop loss is there to turn everything off. I say "dynamic" as it is not based on a % of capital or static number of pips, but is calculated based on average candle size at the time of entry and how the average true range is changing over time. However, I need to refine this approach, or better, look for a new concept. We all think about profit stops and the challenge of predicting where price will go, what about stop losses, and predicting where a trade is no longer viable and needs to be killed? Share your ideas!