The currency I trade the most frequently is the USDMXN. My dedication to this currency pair has been an evolving process. It all started when I watched a video by Rob Booker describing the consistency of USDMXN price action in the morning. In short, he described how there is usually a reversal around 9 am Eastern Time every morning. I started working on an expert advisor to implement his strategy with a few of my own twists. It then occurred to me that while I'm watching the price drop, waiting for indications of a reversal, I should hold a short position. So I enabled the robot to short the pair one time before attempting the one long position in the morning. What I discovered was that the short positions had much higher win rates than the long positions. This new robot back-tested and ran pretty well. Once it got to the point that I thought it was ready for prime time, I decided that I would put it online and walk away. The first week or two went great. Then the currency pair got in a strong downtrend. While the short positions were still making profit, the robot kept trying to take long positions and ended up stopping out. After a week and a half of consistently hitting my stop loss on long positions, I took the robot offline for a short hiatus. On a side note, my philosophy on stop losses is that they should be large. Like a net under a trapeze artist, they should almost never come into play. In lieu of a stop loss, each robot should have enough intelligence built into it to know when the trade is going south and close it before the stop loss even becomes a factor. Needless to say, with my large stop losses, it doesn't take too many days of hitting them before things get ugly. Reflecting on my losses after the short hiatus, my conclusion was "duh." Even though the Tequila Sunrise (TSR) is a counter trend trading robot (runs on the 5 minute charts or short term), I should not be betting against the intermediate trend on the 1 or 4 hour charts, especially when it is a strong trend. I considered using ADX as a criteria for trade entries but didn't feel like it was reliable enough. So my fix was as follows: I tightened my stop loss just a bit and started performing quick assessments of the intermediate trend every couple of weeknights. The TSR has settings that allow the user to disable short or long positions (you can actually do this in MT4 but I built the settings into the EA as well). If the pair is in a strong trend on intermediate time frames, I will disable trading in the opposite direction. Even during these strong trends, there are still money making opportunities with reversals in the morning. However, by disabling a trade in the opposite direction of the intermediate trend, I prevent the robot from getting stuck in a bad trade. Of course this is a function that could be automated and I plan to work on that. Additionally, I need to work on smarter exit strategies for TSR so the stop loss rarely is hit. I've got the TSR back online with better results but the jury is still out and the current down trend is still in progress.
How has this been performing?
Any progress on automating the strategy for accommodating the intermediate trend?
reed33:
I took the EA off for a little while as I was trying some other strategies. It had done well for a while but was in a slump. I actually put it back on at the end of November and it has been doing pretty well since. I haven’t put a lot of effort into automating the assessment of the intermediate trends since assessing it manually doesn’t take too much time. I’ve developed a new version that trend follows during the remainder of the day and then counter trend trades when appropriate. It didn’t do great in backtesting and has been on the shelf for awhile.How has this been performing?
Any progress on automating the strategy for accommodating the intermediate trend?
