#41 【Survival Report W】 Progress update on development of the Box-Through EA
Currently, I am tackling the development of my fourth EA, but this time it is proving to be more difficult than ever.
Therefore, quite a lot of days have passed since the last article.
... Wait, I just realized— the status of the previous article was still "private"! Sweating
Now I’ve set it to "public." (´⌒｀。)
During development, I have been implementing various ideas and repeatedly testing them, but I haven’t been able to achieve results that satisfy me.
This time, I would like to candidly summarize the struggles and detours up to this point as my "record of effort."
Mid to late May: Starting from the "box-out" logic
I named the fourth EA’s logic "box-out" and began development.
Initially, the spec targeted the open price of the baseline bar, and I would enter a position when the 1-minute bar fell below it—a very simple setup.
Later, I expanded the target to include not only the open price but also the close price, and changed the rule to enter a position when the price moved outside the body (open to close). I also started optimizing by widening the execution bars beyond just the 1-minute, to 5, 15, 30 minutes, etc.
End of May to early June: Shift toward trend following and increasing condition complexity
By the end of May, I radically shifted to a logic that would enter when two (or a specified number of) bullish or bearish candles occurred in succession.
I added logic to ignore doji candles, and even incorporated the idea of reversing the position immediately after a stop loss, which has made testing very messy.
In June, I changed the target to the high and low, evolving toward a clearer break strategy.
Mid to late June: Detours and obsession with "fine trailing"
During this period, the entry conditions were the main source of struggle.
"Enter if the same color appears twice in a row"
"No, only look at the color of the very last candle and enter"
"Compare the one-before-last and two-before candles to decide the condition"
and so on—rules kept flipping with each test.
I also added logic to skip Saturday and Sunday candles that cannot be visually seen, painstakingly defining candle handling by day of week.
Furthermore, to maximize profits, I implemented a "trailing" feature that progressively raises stop loss and take profit points.
Late June to early July: Transition to a new logic and introduction of capital management
At the end of June, I thoroughly revisited the existing logic and launched a new logic focused solely on the USDJPY pair.
It is a mechanism that sets the target when a bullish candle of a certain size (BodySize pips) appears.
When July arrived, I added compounding functionality and a strict capital management rule that stops operation if capital falls below 10%.
Also, to minimize窓 openings during year-end holidays, I calculated backward from December 28 to identify the "last business day" and "first business day of the new year" and restricted entries with a very fine calendar logic.
Most recently, on July 7, I added a condition to close immediately if the price dips below the prior candle's low, reinforcing safety.
The battle continues
After spending so many days, I have coded numerous ideas and implemented multiple risk-avoidance measures (weekend and year-end holiday gaps avoidance and fine trailing settings, etc.).
Yet, even now, I still don’t see satisfactory results, which is the harsh reality of EA development.
In this maze without a correct answer— the market— I keep hitting walls, but I have no intention of giving up.
Still, since good results have not yet appeared, the fight continues.
Next time:
I will share good news... at least that’s the plan. w
Paid content list
Article: A winning method even with scalping banned—the beard-trimming method
150 yen! Cheap!... President, cheap! ٩(ˊᗜˋ*)و
Article: 90% of your losing trades are due to the broker’s fraud.
Article: The Cross Method :
In chaotic markets after the cross appears, it ruthlessly takes away the shattered remnants of the losers,
an ultra-precise strike against conventional wisdom!
150 yen! Cheap!... President, cheap! ٩(ˊᗜˋ*)و
E-book:【Broker-Managed】Beard-trimming Method in the Stop Scalp Era
E-book:【Stop-hunting Managed】Raising the leftovers of losers in a chaotic market
“Cross Method”
E-book:【Peak Profit from Losses】Mercilessly taking the flesh and blood of losers