EdgeDriven Dollar Yen Portfolio

Nine USDJPY strategies, one EA to manage. EdgeDriven Dollar Yen Portfolio evaluates H2 signals and combines separate position management for each strategy with shared lot sizing, news and weekend controls. It runs as one Expert Advisor on a MetaTrader 5 hedging account.

You get several USDJPY approaches in a single installation, with a common sizing adjustment through ProductRiskScale. Entries and exits follow defined rules, without martingale sizing or grid averaging to recover losses. Original MT5 reports, losing-month figures and a separately identified live record let you examine the evidence before buying.

The portfolio uses price, momentum and volatility conditions. Its composition and relative weights are fixed; this is not a strategy builder or optimizer. All nine strategies trade USDJPY, so their exposure and losses can overlap.

The historical results below are MT5 simulations. The separate live monitoring record is identified below. Neither establishes future profitability.

What one EA brings together

• One EA instance on the broker's USDJPY H2 chart, with synchronized H2 history.
• Nine fixed strategies with pending-stop entries, protective stop-loss, break-even management and bar-count exits.
• No martingale sizing or grid averaging to recover losses.
• No customer-side DLL, external indicator installation, license server or subscription required.
• MetaTrader 5 hedging account; netting accounts are not supported.

Each module can own one position at a time. The EA evaluates the current chart symbol and H2 data internally. Attaching it to another instrument can produce orders there; USDJPY is the recommended and performance-tested market.

This is a separate nine-module implementation with its own evidence, not a symbol setting of the twenty-module Gold product. Their joint performance and account risk have not been verified.

Try it with your broker data

The free demo runs in MetaTrader 5 Strategy Tester, not on a live chart. Use your intended account size and broker history to check initialization, trade participation, losses and drawdown before purchasing. Broker spreads, commissions, swaps, symbol specifications, time conventions and execution can change the results.

Dollar Yen demo and result-check guide. For product selection, see the comparison of the three EdgeDriven portfolios.

Test results and portfolio selection

Development (2014-2022) was used for portfolio construction and candidate selection. The top candidates were fixed before Validation (2023-2024), which narrowed the candidates and set their priority. The nine-module composition and priority were fixed before Final Historical Holdout Confirmation (January 2025-June 2026).

After the Final results were evaluated, no membership, weighting or ranking changes, re-optimization or rescue were made. This separation applies to portfolio-combination selection. Individual strategies had already been exposed to some later-period data during research or qualification, so Final is not completely unseen data for every strategy.

Explanation of the validation framework.

Reference conditions: Version 1.103; USDJPY H2; MetaQuotes-Demo real-tick mode; USD 10,000 initial capital; hedging account; leverage 1:100; ProductRiskScale=1.0; BaseRiskPercent=0.9; Compounding=false; RestrictedNewsGuard=true; NewsGuardDiagnostics=false.

The first three rows are separate tests that each start with USD 10,000. Combined is a continuous test over the full history, not another independent validation sample or the portfolio-selection score. The end date is 1 July 2026, exclusive.

Profit Factor is the MT5 report value at displayed precision. The drawdown percentage shown with the maximal money equity drawdown uses the associated equity peak, not initial capital, as its denominator. It is not a future loss limit.

The Combined test contains 96 positive and 54 negative calendar months. The worst posted-cash month was January 2015 at USD -212.10. Monthly cash includes posted profit, commission and swap; it is not a mark-to-market monthly return. July-December 2026 are outside the test.

Commission is zero in these reference reports; historical spread and reported swap are included. Real-tick mode remains a historical simulation and does not establish future fills or broker-independent results.

Live trading record you can inspect

The MQL5 Signals monitoring record started on 14 September 2026 on a JPY-denominated GaitameFinest-LIVE account. The published live setup uses leverage 1:25, ProductRiskScale=4.0, BaseRiskPercent=0.9, Compounding=false and RestrictedNewsGuard=true.

The live record and the historical reference use different currencies, leverage and risk scales. Neither their cash profits nor their percentage returns and drawdowns are directly comparable. Evaluate each record under its own settings. The live history is still short and should not be treated as proof of future performance.

The Signal is provided for monitoring. Purchasing this EA and subscribing to a Signal are separate services.

Position sizing

ProductRiskScale changes the nominal loss budget used to calculate lot size. Keep BaseRiskPercent=0.9 as the reference when comparing scales. At scale 1.0, each module receives a nominal 0.1% entry budget; adding one budget from each of the nine modules gives 0.9%. These are sizing budgets, not loss limits.

Example: USD 10,000 starting balance, BaseRiskPercent=0.9, Compounding=false.

Lot size is calculated from the budget and stop-loss distance, then rounded down to the broker's volume step. Below-minimum volumes are skipped; the EA does not increase the lot or redistribute the unused budget to force a trade.

At scale 1.0, three USD 10 entry budgets represent USD 30 of combined initial budgets. Costs, gaps, slippage, simultaneous positions and pending-order fills can increase actual losses. Changing scale from 1.0 to 2.0 doubles the calculation budget, but rounding and skipped trades prevent a guarantee of proportional returns or drawdown.

Compounding=false uses the stored starting balance. Compounding=true uses current account balance for new sizing. The example table explains the calculation; it is not a backtest of all four settings.

USD 10,000 is a reference test balance, not a universal minimum deposit. A USD 100 execution check skipped trades rather than exceed its budget at minimum lot. High-capital, low-leverage and doubled-spread checks are engineering checks, not additional recommended risk profiles. Test your intended capital and broker conditions, allowing for losses worse than the historical drawdown.

News, weekend and execution controls

Restricted USD News Guard uses the fixed historical event windows from development in Strategy Tester and the MetaTrader 5 Economic Calendar in live operation. From five minutes before until five minutes after a selected event, it blocks new entries and requests cancellation of owned pending orders and closure of owned positions.

If live calendar data cannot be refreshed or becomes stale, new entries remain blocked until calendar health is restored. Existing position management continues while the terminal is connected and receiving ticks. The guard reduces scheduled-event exposure; it does not eliminate execution risk.

From Friday 20:00 until Sunday 23:00 in broker server time, new entries are blocked. On incoming ticks, the EA requests cancellation of its pending orders and closure of its positions, retrying when trading is possible. Broker timezones change the corresponding UTC hours. At the end of the pause, an entry still requires a normal H2 signal and all trading checks.

MT5 weekday/hour entry statistics count entry deals; profit/loss statistics group exit deals by their closing time. Scheduled cleanup can concentrate realized results in the Friday bucket. That does not establish a Friday entry advantage. The USDJPY rule uses server time; Gold uses a separate UTC rule.

Before submission, the EA checks volume steps and limits, stop/freeze distances, session availability, order capacity and margin, including pending-order margin reservation. It manages protective stops, break-even changes, pending-order expiry and bar-count exits under each fixed module's rules.

Accepted protective stops remain at the broker and pending orders can fill while MT5 is offline. EA-managed changes, expiry and timed exits require a connected terminal, fresh quotes, permissions and broker acceptance. Gaps and slippage can cause losses beyond the sizing budget.

Inputs

Reserve the MagicBase through MagicBase+8 range exclusively. Do not change it while owned positions or pending orders remain. Reference defaults are a comparison baseline; risk inputs do not enforce a guaranteed loss limit.

Setup, restart and support

1. Install through the MetaTrader 5 Market and test the free demo with your broker's data.
2. Check the hedging account, USDJPY H2 chart, history, risk settings, symbol specifications and available margin.
3. Reserve an unused Magic range and set TradingEnabled=false in Inputs before attachment. The compiled default is true.
4. Check the Journal for initialization and errors. Enable new entries only when you intend to allow orders.

TradingEnabled=false blocks new submissions but retains owned management and does not cancel existing pending orders. EmergencyStop blocks entries and repeatedly requests owned pending cancellation and position closure when execution is possible. It does not close unrelated magic identifiers or guarantee instant liquidation.

Management is limited to the chart symbol and reserved Magic range. Local state stores the starting-capital basis and last processed bar; current orders and account history inform management after reinitialization. Preserve terminal state during updates. The stored capital basis is local and is not automatically transferred by this package.

Before an update or VPS migration, prevent duplicate management from the old instance and reconcile positions and pending orders. There is no lock coordinating separate terminals. Running other EAs requires distinct ownership ranges and sufficient shared account capacity. Tester reinitialization checks are not certification of live reconnects, crashes or VPS power loss.

Keep MT5 or the VPS connected for managed actions. Turning Algo Trading off can prevent EA management requests while broker-side orders remain active. Reconnection cannot guarantee the intended exit price or time.

Support is available through product comments or MQL5 messaging. Include the version, sanitized Journal messages and relevant symbol specifications. Never share credentials. Support covers product operation, not personalized investment advice.

Images and release notes

Images 3 and 4 show the saved native MT5 Balance / Equity graph and tester results for Version 1.103. The original screenshot files are used without redrawing values or curves. The graph includes modeled Balance, Equity and Deposit Load. Portfolio diagrams, period tables and monthly summaries are supplementary explanations, not native MT5 screens.

The gallery shows historical simulations, not the separate live monitoring record. Setup and restart tests are not forward performance records.

Version 1.103 restores the development-consistent restricted news logic and adds live Economic Calendar synchronization. Portfolio membership, signal logic, relative construction and reference risk design are unchanged. Because news controls affect executed trades, the historical reference was regenerated for this release.

Leveraged currency trading can cause substantial losses. Historical performance, diversification, stops and execution controls do not guarantee future returns, a maximum realized loss, identical broker behavior or independence between modules.