[Lab Development Logic Introduction] Heading for the Week-Open 60 minutes only: Zero-parameter EA "GBPJPY Week-Open Raw" — Answers to five design questions
This is the fifth installment of our product introduction. The subject is "GBPJPY Week-Open Raw." It is an MT4 automated trading EA that buys only once every Monday and closes 60 minutes later. Rather than listing the specifications in order this time, we will introduce it by answering questions that might come up about this EA.
Logic Overview
During the two weekend days, the market is closed. The supply and demand accumulated during that period surface immediately after the new week begins. This EA takes only the upward skew in GBPJPY that appears at that moment once every week, without processing, and does so. There is no element that predicts the market direction.
Q2. Isn’t it okay without stop loss?
By default, no stop loss is set. Each entry is a single lot, with no averaging down or position scaling, so losses do not grow without bound. However, a single downside move can be larger than an EA with a stop loss.
The worst single drawdown during the validation period was -119.6 pips on 2016-11-21. In March 2020, three of the worst five drawdowns ( -115.7 pips, -83.3 pips, -74.7 pips) concentrated. If you cannot tolerate this behavior, you can set an arbitrary loss amount in the input SL_PIP. Risk management is left to the purchaser's discretion.
Q3. Wouldn’t adding filters improve it?
Adding conditions tends to make the backtest numbers look better. On the other hand, the real-world performance diverges more from the backtest. This EA leaves only “when to buy and when to close” and does not reference moving averages, ranges, or volume.
Q4. Is this not just a lucky timing?
Shifting the entry time by a few minutes before or after, or varying the holding time between 45 and 180 minutes, retained the edge. Since there are no adjustment parameters, there is no room for overfitting to past data. However, the entry time itself is chosen based on past data validation.
Q5. What are the actual results?
Validation environment: MT4 / GBPJPY M1 / All ticks (modeling quality 25.00%) / No chart misalignment errors / 2015.07.01–2026.06.29 (about 11 years) / Initial margin 10,000 / Spread 1.0 pip / 0.06 lots fixed / Trading fees included
Per trade, the average win is +19.2 pips, the average loss is -14.6 pips (both values before fees). The design has no stop loss, but the average loss is still below the average win. Losses exceeding 50 pips occurred 9 times out of 565. The maximum consecutive losses were 4, and a 5-win streak did not occur at any time during the validation period. Use this as a reference when considering adjustments during live trading.
Yearly Performance
Negative performance occurred only in 2020 (-6.5 pips, almost balanced). However, since 2021, the growth has been larger than before. Do not assume this level will continue in the future.
Notes for Use
・Entry time is based on the broker's server time. If you use a broker with a different time zone from the validation, adjust with the input EntryHour
・Spreads may temporarily widen immediately after the start of the new week. The validation uses a fixed 1.0 pip spread
・Orders are not placed during weeks when the trading start is delayed at the beginning of the week, so there are about 51 orders per year
・Drawdown scales with the size of the trading lot
・Holdings do not span days (maximum 60 minutes), so swaps are not incurred in principle
For MT1-applied EAs, the maximum modeling quality on all ticks is 25%. This EA waits for a candlestick close before ordering, so even at this quality, execution is faithfully reproduced.
This is an automated trading tool based on past validation and does not guarantee future profits.
For inquiries about the logic and operational aspects, please visit the Getting Help Desk.