It looks like this is the best I can do (using the basic tools made available by EABuilder) in coming up with an expert advisor that endeavors to duplicate the results I get trading the major Forex pairs manually... I guess I'm going to have to wait until I finish learning how to code EAs myself before I'll be able to refine this "blueprint" to the point that I can hand over the reins to (more sophisticated, complex, and nuanced) automated trading. I nonetheless feel that the $97 I paid for full access to the website's product was not money wasted. It's easy enough for me to discern when the price of a given pair is rising or falling, but conveying what I "see" to a "dumb" computer in language that empowers it to see the same thing and then act accordingly turned out to be a real challenge for me. The effort has been well worth it, though, forcing me to write definitive guidelines on exactly what it is that makes the system work. It's made explicitly clear precisely which graphics (not pictured in the image below) reveal the probable direction in which rates are headed and how to interpret those "signposts" in a manner such that one ends up with a valid and reliable forecast. The corresponding alerts should enable me to continue trading profitably—but with greater efficiency and far less monitoring of charts—so that the bump in daily returns I'm likely to experience should quickly reimburse the money I spent, God willing.
I transferred the graphics from the above approach to a different time frame and something caught my eye. I applied what I saw to a new strategy and this is what I got... It was profitable for a while, but MT4 was not drawing the chart AS the test was running, so I was unable to see what went wrong after trade 42 in real time. I think I'll get better entry and exit levels if I take this same EA and translate it down to one-minute charts. I'll also try to figure out why the methodology I was using did so poorly at the end, and correct it, but all of this will have to wait until tomorrow. (By the way... no glitches since incorporating what HobbyTrading had to share about using shift 1 rather than shift 0.)
It seems to me that this should be a winning strategy, not merely break even. Tomorrow I'll try transferring it to a lower time frame and see if that doesn't reveal some way I can make it as profitable as I think it ought to be. And the second time I ran it, it painted green SELL arrows along with red ones (and green BUY arrows as well). I have no idea what was causing this to happen... I got another idea that looked like it might be profitable, but it was only breaking even (more or less) until the end. Also, it resulted in only 14 trades in the same amount of time the other approaches were executing dozens... All that notwithstanding, the real value of all this is that it is enhancing my approach to trading manually, making my "job" that much easier.
I'm going to leave the back testing behind for a while. I was very pleased with the chart configuration that produced the graph below, even though it was not profitable as an automated system, so I will begin using it to trade manually... In the meantime, I will resume learning about coding until hopefully my knowledge is sufficient to create a nuanced enough expert advisor to mimic real-life human trading.
Before quitting this morning, I was curious to see whether I could come up with a "dumb" EA that would be able to turn a profit due to my basing it strictly on the nature of automated trading and NOT on any kind of sophisticated market observations (which I am unable to integrate into any kind of source code that I'm able to generate at this time). As it turned out, I was indeed able to do so... (This is only ten month's worth of trading, from November of last year.)
..EA BENCHMARK... .. .Tuesday, September 21, 2021 My plan was to leave the back testing behind for a while, but as I was setting up my new charts, I got yet another idea I was curious to investigate. I haven't actually reached Shangir-La. But, given my extremely limited know-how with respect to coding expert advisors, I think the version that generated the chart below has provided me with a basic blueprint I'm likely to be able to expand on once my education in this area is complete... Now it's time to get back to the learning process!
I believe this is the EA I want, but that it didn't maintain profitability because of the necessity of having to us "shift 1" for position exits. Therefore, later today or sometime tomorrow, I plan to translate it down to one-minute charts to see it this doesn't eliminate most of the lag associated with closing out positions at optimal levels. (I had intended to set aside back testing for a while and go back to learning more about coding. But, as each trial EA leads to enhancements to my chart configurations—the enhancements, in turn, lead to additional ideas for expert advisors.)
Attacking the market from this angle looked great in the beginning. But, from trade 32 on, the approach only managed to tread water...
FRIDAY / SEPTEMBER 24, 2021 Again, mission accomplished. I have two versions of this same approach. The first resulted in a greater number of trades and a more substantial return in the end, but took a somewhat erratic path to get there... Whereas the second got off to a poor start, but then evidenced a more or less steady climb thereafter... My plan now is to convert this same methodology to lower time frames to generate a higher percentage of winning trades along with a relatively negligible occurrence of drawdown. What I'm hoping is that once I know how to code expert advisors all on my own, it will equip me with the ability to combine these two versions into one automated system. Theoretically, it seems to me that this ought to be possible using some kind of "if/else" language, but I guess I'll find out soon enough.
Finally! I am at last finished with the back testing. What I will do now is continue day trading manually using the methodology that yielded the following results... ...until and unless I am someday able to self-code an expert advisor (EA) that can execute this system in ALL of its nuanced complexity.