FXEA365 Atlas Diversified Trend Portfolio
- Experts
-
Kenichiro Sakamoto
I'm Kenichiro Sakamoto, the developer behind FXEA365 — a studio building transparent, rule-based Expert Advisors for MetaTrader 5. - Version: 1.52
- Updated: 10 October 2026
- Activations: 5
FXEA365 Atlas Portfolio — one EA trading a five-market trend portfolio (USDJPY, GBPJPY, EURJPY, BTCUSD, ETHUSD) from one chart (MetaTrader 5)
=== LAUNCH PRICE: LAST ONE AT $35 ===
The price rises after every 3 purchases and keeps rising with no upper limit ($35 → $49 → $79 → $99 → $149 → …). One more sale and the price goes up to $49.
=== QUICK LINKS ===
- User guide and ready-made set files (one per run mode, plus a prop-firm set): https://www.mql5.com/en/blogs/post/779744
- Forward test: Exness demo account (not real money), figures below and updated in the guide
- Questions and setup help: the Comments tab of this page
- All our products: https://www.mql5.com/en/users/app.develop.sk/seller
=== KEY FACTS ===
- Symbols: USDJPY, GBPJPY, EURJPY, BTCUSD, ETHUSD, all traded from one USDJPY H1 chart. All five symbols must be in Market Watch; broker prefixes and suffixes are detected automatically, and a missing symbol is reported in the log
- Account: designed for hedging accounts. On a netting account the EA prints a warning and behaviour will differ from the tested configuration. Built for MetaTrader 5 — a separate MetaTrader 4 edition is available: https://www.mql5.com/en/market/product/208328
- Leverage: 1:100 or higher; confirm your broker's crypto leverage
- Run modes: Defensive, Standard (default since v1.47), Aggressive, Ultra
- No grid, no martingale, no averaging down. The trend engines hold at most one position per symbol
- Stop-loss: every position has a defined stop from the moment it opens. The JPY legs send their stop-loss to the broker's server with every order. The crypto legs' stops are monitored and executed by the EA, so those positions are protected only while the EA is running. A VPS is strongly recommended
=== RESULTS (WITH MAXIMUM DRAWDOWN) ===
Backtests are historical simulations, not a live record, and do not indicate future performance.
- Five-market portfolio, 2019.01.01 to 2026.06.28, deposit USD 10,000, Standard: profit factor 1.48, 1,193 trades, win rate 33.19%, maximum equity drawdown 19.74% (maximum balance drawdown 16.29%), recovery factor 6.17.
- Recent real-tick check, 2026.04 to 2026.07, deposit USD 10,000: Defensive profit factor 1.15, drawdown 11.4%. Standard profit factor 1.11, drawdown 22.7%. Aggressive profit factor 1.09, drawdown 39.0%. Ultra profit factor 0.99, drawdown 70.9% — the only mode that did not make money, which is why v1.47 made Standard the default.
- Monday-open sleeve (v1.47 → v1.49, real ticks, JPY legs only, Standard, 2023.01-2026.09, $10,000): net +$17,997 → +$20,600, maximum equity drawdown 24.04% → 22.29%.
- Previous-day breakout module (v1.49 → v1.50, real ticks, all legs, Standard, each year from $10,000): 2024 net +$8,686 → +$10,813, maximum equity drawdown 19.61% → 17.54%. 2025 +$1,671 → +$1,787, 25.20% → 25.58%. 2026.01-2026.09 -$789 → -$434, 22.27% → 22.28% — a losing period in both versions.
- Forward test (Exness demo account, not real money, default settings, since 2026-08-21): 65 closed trades, -19.28% net, profit factor 0.17, maximum drawdown 19.76% as of 2026-10-31. This has reached the backtest maximum (19.74%); size positions assuming a drawdown of this depth can happen again.
The five-market and run-mode figures were measured before the sleeve and the module existed.
=== HOW IT WORKS ===
1) JPY pairs (MEGAMAX engine) — long only, Donchian channel breakouts, ATR initial and trailing stops, and a maximum holding period.
2) Crypto (COMET engine) — Donchian breakout with a long-term simple moving average trend filter and an extension filter.
3) Monday-open sleeve (v1.49) — Monday 01:00 broker time, one extra buy on each JPY pair, closed at 02:00, with a hard 50-pip stop-loss on the server; skipped if the spread stays above 3.5 pips for 30 minutes. UseMondayFlow=false turns it off.
4) Previous-day breakout module (v1.50, USDJPY) — one buy-stop at the previous day's high and one sell-stop at its low, each with a server-side stop-loss and take-profit; unfilled orders are removed the same day. Set UsePrado to false to turn it off. Its time window assumes a server clock of UTC+3 (input PR_TimeBroker); v1.51 fixes buy orders when this input is not 3.
=== RUN MODES AND SET FILES (real-tick check, 2026.04-2026.07) ===
- Defensive: the most conservative sizing, drawdown 11.4%
- Standard (default): drawdown 22.7%
- Aggressive: drawdown 39.0%
- Ultra: drawdown 70.9%, no longer the default
- Prop-firm set: Defensive mode — the lowest-risk mode measured above (11.4% drawdown in the real-tick check), with everything else at the shipped preset.
The published results correspond to the Run-Mode presets as shipped; the set files in the guide only change the Run-Mode.
=== FAQ ===
Q: Does it use martingale or grid?
A: No grid, no martingale and no averaging down. Every position has a stop from the moment it opens; the crypto legs' stops are managed by the EA, so keep it running on a VPS.
Q: Why is the win rate only about one third?
A: It is a trend-following system: it relies on the average win being larger than the average loss, not on winning often.
Q: Why was the default changed from Ultra?
A: In the 2026.04 to 2026.07 real-tick window Ultra was the only mode below profit factor 1.0, at a 70.9% drawdown. v1.47 changed only the default; the trading logic was not changed.
Q: How do I check it before buying?
A: Run the free demo on USDJPY H1 with "Every tick based on real ticks" and all five symbols available, 2019.01.01 to 2026.06.28, $10,000, Standard, and compare with the portfolio figures above.
=== RISK WARNING ===
The forward test has already reached the backtest's maximum drawdown, and 2026 to date is a losing period in the backtest too. Start on a demo account or a small account on Standard or Defensive, and never trade money you cannot afford to lose.
=== LAUNCH PRICE: LAST ONE AT $35 ===
The price rises after every 3 purchases and keeps rising with no upper limit ($35 → $49 → $79 → $99 → $149 → …). One more sale and the price goes up to $49.
=== QUICK LINKS ===
- User guide and ready-made set files (one per run mode, plus a prop-firm set): https://www.mql5.com/en/blogs/post/779744
- Forward test: Exness demo account (not real money), figures below and updated in the guide
- Questions and setup help: the Comments tab of this page
- All our products: https://www.mql5.com/en/users/app.develop.sk/seller
=== KEY FACTS ===
- Symbols: USDJPY, GBPJPY, EURJPY, BTCUSD, ETHUSD, all traded from one USDJPY H1 chart. All five symbols must be in Market Watch; broker prefixes and suffixes are detected automatically, and a missing symbol is reported in the log
- Account: designed for hedging accounts. On a netting account the EA prints a warning and behaviour will differ from the tested configuration. Built for MetaTrader 5 — a separate MetaTrader 4 edition is available: https://www.mql5.com/en/market/product/208328
- Leverage: 1:100 or higher; confirm your broker's crypto leverage
- Run modes: Defensive, Standard (default since v1.47), Aggressive, Ultra
- No grid, no martingale, no averaging down. The trend engines hold at most one position per symbol
- Stop-loss: every position has a defined stop from the moment it opens. The JPY legs send their stop-loss to the broker's server with every order. The crypto legs' stops are monitored and executed by the EA, so those positions are protected only while the EA is running. A VPS is strongly recommended
=== RESULTS (WITH MAXIMUM DRAWDOWN) ===
Backtests are historical simulations, not a live record, and do not indicate future performance.
- Five-market portfolio, 2019.01.01 to 2026.06.28, deposit USD 10,000, Standard: profit factor 1.48, 1,193 trades, win rate 33.19%, maximum equity drawdown 19.74% (maximum balance drawdown 16.29%), recovery factor 6.17.
- Recent real-tick check, 2026.04 to 2026.07, deposit USD 10,000: Defensive profit factor 1.15, drawdown 11.4%. Standard profit factor 1.11, drawdown 22.7%. Aggressive profit factor 1.09, drawdown 39.0%. Ultra profit factor 0.99, drawdown 70.9% — the only mode that did not make money, which is why v1.47 made Standard the default.
- Monday-open sleeve (v1.47 → v1.49, real ticks, JPY legs only, Standard, 2023.01-2026.09, $10,000): net +$17,997 → +$20,600, maximum equity drawdown 24.04% → 22.29%.
- Previous-day breakout module (v1.49 → v1.50, real ticks, all legs, Standard, each year from $10,000): 2024 net +$8,686 → +$10,813, maximum equity drawdown 19.61% → 17.54%. 2025 +$1,671 → +$1,787, 25.20% → 25.58%. 2026.01-2026.09 -$789 → -$434, 22.27% → 22.28% — a losing period in both versions.
- Forward test (Exness demo account, not real money, default settings, since 2026-08-21): 65 closed trades, -19.28% net, profit factor 0.17, maximum drawdown 19.76% as of 2026-10-31. This has reached the backtest maximum (19.74%); size positions assuming a drawdown of this depth can happen again.
The five-market and run-mode figures were measured before the sleeve and the module existed.
=== HOW IT WORKS ===
1) JPY pairs (MEGAMAX engine) — long only, Donchian channel breakouts, ATR initial and trailing stops, and a maximum holding period.
2) Crypto (COMET engine) — Donchian breakout with a long-term simple moving average trend filter and an extension filter.
3) Monday-open sleeve (v1.49) — Monday 01:00 broker time, one extra buy on each JPY pair, closed at 02:00, with a hard 50-pip stop-loss on the server; skipped if the spread stays above 3.5 pips for 30 minutes. UseMondayFlow=false turns it off.
4) Previous-day breakout module (v1.50, USDJPY) — one buy-stop at the previous day's high and one sell-stop at its low, each with a server-side stop-loss and take-profit; unfilled orders are removed the same day. Set UsePrado to false to turn it off. Its time window assumes a server clock of UTC+3 (input PR_TimeBroker); v1.51 fixes buy orders when this input is not 3.
=== RUN MODES AND SET FILES (real-tick check, 2026.04-2026.07) ===
- Defensive: the most conservative sizing, drawdown 11.4%
- Standard (default): drawdown 22.7%
- Aggressive: drawdown 39.0%
- Ultra: drawdown 70.9%, no longer the default
- Prop-firm set: Defensive mode — the lowest-risk mode measured above (11.4% drawdown in the real-tick check), with everything else at the shipped preset.
The published results correspond to the Run-Mode presets as shipped; the set files in the guide only change the Run-Mode.
=== FAQ ===
Q: Does it use martingale or grid?
A: No grid, no martingale and no averaging down. Every position has a stop from the moment it opens; the crypto legs' stops are managed by the EA, so keep it running on a VPS.
Q: Why is the win rate only about one third?
A: It is a trend-following system: it relies on the average win being larger than the average loss, not on winning often.
Q: Why was the default changed from Ultra?
A: In the 2026.04 to 2026.07 real-tick window Ultra was the only mode below profit factor 1.0, at a 70.9% drawdown. v1.47 changed only the default; the trading logic was not changed.
Q: How do I check it before buying?
A: Run the free demo on USDJPY H1 with "Every tick based on real ticks" and all five symbols available, 2019.01.01 to 2026.06.28, $10,000, Standard, and compare with the portfolio figures above.
=== RISK WARNING ===
The forward test has already reached the backtest's maximum drawdown, and 2026 to date is a losing period in the backtest too. Start on a demo account or a small account on Standard or Defensive, and never trade money you cannot afford to lose.
