Interested in thoughts/feedback on this. Because we can trade either long or short into a currency pair, I think many of us implement our trading strategies using the same levels, targets, and expectations both long and short. I admit, I’ve always thought of it that way. To be clear, I’m not talking about whether or not to trade against a trend or MA. The question is, when you execute your trading strategy (whatever it is) do you look for the exact same trigger levels, have the same pip gain expectations, use the same risk:reward ratio or stop, etc. regardless of long or short?

I’m watching my MT4 Strategy Tester run through bajillions of optimization passes in my typical “find needle in haystack” manner, and something stood out. In this example, I have an EA which uses a primary trading strategy that can start with \$200 in 2005 at 20:1, and end up with 120k today at a 1.80ish profit factor. I’m working on introducing a secondary strategy function to it, a very simple thing, that with no tuning, can bump that to \$148k, but knocks the PF down to 1.19 (more profit, but also more slop). The secondary function simply triggers if the current period high and low are beyond x standard upper/lower BB deviations, relative to the last period. Set to all 4s for BB deviations, that gets me the \$148k. To tune the slop out of this function and get better win rates, I created 4 variables. One for top and one for bottom deviations on long trades, and one for top and one for bottom deviations on short trades. I then told Strategy Tester to loop each BB variable through values of 1-6, against every other combination of 1-6 on all the other variables for the last 15 years. Needless to say, this test is taking days, and is still not done.

One thing that has jumped out at me is that I’m seeing the profit and profit factor ranges I’d expect as I tune something up, but they’re NOT at all corresponding to any sort of symmetry between long and short. For example, the best mix of profit and PF so far involves long trades with a lower BB deviation of 6 and an upper deviation of 5, with short trades being 1 and 4 inversely. At first I was surprised that best success combinations don’t look the same both directions, but maybe it makes sense. Why would we expect a currency pair to move both directions with the same momentum? Maybe they tend to be stronger or longer one way vs the other (regardless of being on the right side of a trend).

This is only one simple example, but it’s making me rethink the way I view bidirectional expectations in my trading strategies that have previously been symmetrical. I think I’m going to start tuning EAs with entirely different values for long vs. short all over the place. Just wanted to see if anyone else has thoughts on this subject. Cheers!

2 Likes

This is something I have never had the chance to explore. Up until now I have used the same trend features and characteristics to determine whether I should be long or short, and to gauge the strength of trend such that i either hold and pyramid or hold as far as the next pull-back or hold only until the next favourable close, but always regardless of whether I am long or short.