EdgeDriven Gold Portfolio
- Experts
-
Dan Mishima
Edge, Backed by Evidence.
EdgeDriven Algo is an Expert Advisor (EA) brand dedicated to developing algorithmic trading systems that prioritize robustness and reproducibility, all validated through rigorous statistical analysis. - Version: 1.6
- Updated: 11 September 2026
- Activations: 5
EdgeDriven Gold Portfolio is a fixed, multi-strategy XAUUSD Expert Advisor for MetaTrader 5 hedging accounts.
Instead of relying on a single entry model, it combines 20 fixed strategy modules using H1/H4 market data. Each module keeps its own trade ownership and exits, while one EA manages portfolio-level sizing, weekend rules, news controls and execution checks.
The portfolio is fixed. It is not a strategy builder, optimizer or signal selector.
EdgeDriven Algo — Edge, Backed by Evidence.
The results shown below are historical simulations, not live or forward returns. Leveraged Gold trading can cause substantial losses.
At a glance
-
XAUUSD portfolio for MetaTrader 5
-
20 fixed strategy modules
-
Hedging accounts only
-
No Martingale
-
No grid or averaging-down recovery
-
Module-specific protective stops and exits
-
Break-even and bar-count exit logic where defined by the module
-
7,013 historical trades from January 2014 through June 2026
-
Combined historical MT5 Profit Factor: 1.38
-
Restricted USD News Guard
-
Friday cleanup and weekend entry pause
-
No customer-side DLL or external indicator installation
-
No external license server or subscription required
Try it before purchasing
Use the free Market demo in MetaTrader 5 Strategy Tester before deciding whether to purchase.
Not sure which EdgeDriven portfolio to test? Compare Gold, Dollar Yen and Multi Asset.
Preparing your first test? Read the Gold demo and result-check guide.
Run it with your own broker history and trading conditions, then compare your results with the reference evidence below. Differences in spread, commission, swaps, contract specifications, minimum lot, history and execution modeling can all change the result.
A demo test can help confirm setup and broker compatibility. It does not predict future live performance.
Why use a portfolio?
Every strategy has market conditions where it performs better and conditions where it performs worse.
A portfolio approach allows different modules to participate at different times instead of depending on one entry model to perform well in every environment.
That does not mean 20 modules equal 20 independent sources of profit.
All modules still trade XAUUSD and can share the same market exposure, margin usage and losing periods. Several modules may hold positions at the same time, and losses can overlap.
The purpose of the portfolio is to combine different strategy behavior under one fixed construction — not to claim that portfolio risk disappears.
What is fixed
The 20-module membership and intended relative allocation are fixed.
Each module has separate trade state and order ownership. The individual strategy recipes are not user tuning inputs, and the built-in risk profiles do not switch strategies on or off. They scale the same underlying portfolio construction.
Position size is calculated from the selected risk configuration and the module's stop distance, then rounded downward to the broker's allowed volume step.
If the broker's minimum lot would require more risk than the sizing budget allows, the trade is skipped.
The EA does not round the lot upward to force participation, and skipped risk is not reassigned to another module.
This means smaller accounts may participate in fewer signals than the historical reference.
Increasing risk simply to force trades is not a solution.
Historical assessment process
The portfolio was developed and assessed across separate historical periods:
-
2014–2022: Development
-
2023–2024: Validation
-
2025–2026H1: Final Historical Confirmation
The selected 20-module portfolio was fixed before the Final period, and the Final results were not used to rerank the portfolio.
However, these labels should not be interpreted as completely untouched individual-strategy OOS data. Some later data had already been seen during earlier component qualification or diagnostics.
For that reason, the Final period is described here as Final Historical Confirmation, not as live, forward, or completely unseen proof.
Current historical reference
The current published reference uses:
-
Product version: 1.006
-
Initial capital: USD 50,000
-
Account type: USD hedging account
-
Leverage: 1:100
-
Risk profile: Standard
-
Common scalar: 3
-
PortfolioRiskScale: 100%
-
Compounding: Off
-
RestrictedNewsGuard: true
-
Test period: January 2014 through June 2026
-
MT5 model: 1 Minute OHLC
Historical Strategy Tester execution uses the frozen restricted-news windows used during development. Live operation uses the MetaTrader 5 Economic Calendar for future restricted USD events.
| MT5 statistic | Historical result |
|---|---|
| Net profit | USD 49,600.78 |
| Profit Factor | 1.38 |
| Trades | 7,013 |
| Maximum modeled equity drawdown | USD 5,176.78 |
| Equity drawdown / equity peak | 6.17% |
| Same drawdown / initial capital | 10.35% |
| Recovery Factor | 9.58 |
The two drawdown percentages use different denominators and are not interchangeable.
Neither should be treated as a future loss limit.
The floating drawdown shown by Strategy Tester is also model-dependent and is not the same as a live-account intratick drawdown measurement.
Historical period breakdown
The figures below are sections of the same continuous historical run. The account was not reset between periods.
| Period | Net USD | PF |
|---|---|---|
| 2014–2022 — Development | 36,878.35 | 1.37 |
| 2023–2024 — Validation | 7,487.26 | 1.35 |
| 2025–2026H1 — Final Historical Confirmation | 5,218.39 | 1.63 |
| 2014–2026H1 — Combined history | 49,600.78 | 1.38 |
The Combined row is descriptive history, not an additional independent validation sample.
The 2026 period contains January through June only.
Losing periods and stagnation
A profitable long-term backtest still contains meaningful losing periods.
Across the same 150 calendar months:
-
92 months were positive
-
58 months were negative
-
Worst calendar month: January 2023, −USD 1,215.32
-
Worst month-end rolling 12-month period: July 2021 through June 2022, −USD 2,140.97
-
Balance stagnation: 362.92 days
-
Hourly-observed equity stagnation: 441.04 days
The equity-stagnation figure is based on hourly observation, not every intratick movement.
These figures describe what occurred in the historical simulation. Future losing periods, drawdowns and stagnation can be longer or larger.
Risk profiles
The EA provides four risk modes:
-
Conservative
-
Standard
-
High
-
Custom
They scale the same fixed 20-module construction; they do not select different strategies.
The current public historical evidence above is the Standard reference for Version 1.006.
Profile names are not safety guarantees, and scaling should not be interpreted as a linear return projection. Lot rounding, broker minimum volume and skipped trades can change which signals are actually taken.
For Custom mode, CustomPortfolioScalar controls the common sizing scalar. It is not a per-trade risk percentage.
Choose risk based on the amount of loss and drawdown you can actually tolerate, not on the return you hope to achieve.
Restricted USD News Guard
Version 1.006 includes the Restricted USD News Guard used in the current historical reference.
In Strategy Tester, the EA uses the frozen historical restricted-news windows used during development.
In live operation, it uses the MetaTrader 5 Economic Calendar for the corresponding future USD event families.
During an active restricted window — five minutes before through five minutes after the event — the EA:
-
blocks new entries,
-
requests cancellation of its owned pending orders,
-
and requests closure of its owned positions.
These actions still depend on terminal connectivity, quotes, trading permission, broker sessions and order execution.
If the live economic calendar cannot be refreshed or becomes stale, the News Guard fails closed for new entries until calendar health is restored. Existing position management continues while the terminal is connected and receiving ticks.
The News Guard is designed to reduce exposure around selected scheduled USD events.
It does not eliminate gap risk, slippage or execution risk.
Friday cleanup and weekend pause
The EA is designed to avoid routine weekend position carry.
From Friday 20:00 UTC, it blocks new entries and requests closure of its owned module positions and cancellation of its owned pending orders.
Unacknowledged requests remain eligible for retry during the Friday/Saturday cleanup window.
The weekend entry block remains active until Sunday 23:00 UTC. The fixed weekday module sessions may delay the next actual entry further.
These are UTC rules, not a universal broker-server time.
Live conversion therefore depends on a correct computer UTC clock and a valid observed broker-server offset.
Weekend cleanup is an execution request, not a guarantee that every position will close at the intended time or price.
Stops, sizing and execution checks
Each module retains its own protective stops and exit rules.
Accepted hard stop-loss and take-profit orders remain broker-side. Software-managed exits, pending-order management and cleanup logic require the terminal to be running and connected.
Before opening or submitting new exposure, the EA checks items including:
-
broker volume step and limits,
-
minimum lot,
-
stop-based nominal risk,
-
available margin,
-
broker stop-distance constraints,
-
and pending-order capacity.
These checks can prevent an order from being submitted.
They do not guarantee a fill, a specific exit price, or a maximum realized loss.
Position ownership and duplicate-instance protection
The portfolio uses a dedicated Magic range based on MagicBase .
With the default setting:
MagicBase 2026090800 reserves MagicBase+1 through MagicBase+20.
Do not change the Magic range while the EA still has open exposure.
Persisted state and recovery logic are used to restore module management after supported terminal restarts. A terminal-local ownership check also helps detect overlapping active Magic ranges.
This protection does not coordinate separate MT5 terminals or separate VPS instances.
Do not run a second copy that can manage the same account and Magic range.
During migration or updates, make sure the previous instance can no longer manage the same exposure before enabling the new one.
Operating requirements
-
MetaTrader 5 hedging account
-
Netting accounts are not supported
-
Recommended and validated market: XAUUSD
-
Attach one instance to an XAUUSD H1 chart
-
Synchronize both H1 and H4 history
-
Reference account currency: USD
-
Reference leverage: 1:100
-
Check broker minimum lot, volume step, margin, quotes and stop-distance rules
-
Keep the operating-system UTC clock correct
-
Keep MT5 or the VPS connected when EA-managed actions are required
-
Reserve an exclusive Magic range
-
Do not run duplicate instances in the same account/Magic context
-
No customer-side DLL or external indicator installation is required
The EA can evaluate the current chart symbol, but the commercial evidence and recommended use are for XAUUSD. Other instruments and arbitrary broker specifications are not performance-validated.
Main inputs
| Input | Default | Purpose |
|---|---|---|
| RiskProfile | Standard | Conservative, Standard, High or Custom. Scales the fixed portfolio construction. |
| PortfolioRiskScale | 100 | Percentage multiplier applied to the selected profile. Higher values can increase losses and margin requirements. |
| CustomPortfolioScalar | 3 | Common scalar used only in Custom mode. Not a per-trade risk percentage. |
| UseCompounding | false | false uses the recorded starting-capital basis; true uses current account equity. |
| TradingEnabled | true | Allows new entries when all checks pass. Turning it off does not cancel existing pending orders or disable owned exit management. |
| MagicBase | 2026090800 | Reserves MagicBase+1 through MagicBase+20. Do not change while exposure remains open. |
| EmergencyStop | false | Blocks new entries and requests cancellation of owned pending orders. It is not a flatten-all command. |
Accepted pending orders and broker-side SL/TP can remain active even when new entries are disabled.
Turning MT5 Algo Trading off can prevent new EA-generated management requests while already accepted broker-side orders remain active.
What the historical evidence does — and does not — show
The reference model is:
1 Minute OHLC, reported history quality 78%, not real ticks.
The reference custom-symbol specification uses:
-
nominal spread: 0.250 USD price units
-
commission: 0
-
native tester swap and fill behavior
Intraminute order chronology, fill prices and floating drawdown are therefore model-dependent.
The historical test does not establish:
-
future profitability,
-
a maximum future drawdown,
-
identical behavior at another broker,
-
independence between the 20 modules,
-
identical live execution,
-
or future returns.
There is currently no verified product-linked live or demo-forward return record presented here.
A technical demo, setup test or restart test is not a forward-performance record.
Historical evidence should be read together with its modeling assumptions, costs and operating limitations.
Before enabling live entries
A cautious setup sequence is:
-
Install the Market version through MetaTrader 5.
-
Open the intended broker XAUUSD H1 chart.
-
Synchronize H1/H4 history.
-
Check account type, leverage and symbol specifications.
-
Confirm the computer UTC clock.
-
Confirm the intended RiskProfile and PortfolioRiskScale.
-
Confirm the Magic range is unused.
-
Attach the EA with TradingEnabled=false .
-
Check the Journal for readiness and obvious setup errors.
-
Enable new entries only when you are ready for the EA to trade.
Testing the setup successfully does not guarantee future live results.
Who this product is designed for
This product is better suited to users who:
-
understand that backtests are historical simulations,
-
can tolerate losing months and extended periods without a new equity high,
-
want a fixed multi-strategy XAUUSD portfolio rather than a manually optimized signal generator,
-
are willing to choose a risk setting based on their own drawdown tolerance,
-
and can maintain a properly configured MT5/VPS environment.
It is not designed for anyone expecting guaranteed monthly profits, guaranteed drawdown limits, or a strategy that can be left completely unmonitored regardless of broker or terminal conditions.
Evidence gallery
The product gallery includes actual English MT5 Strategy Tester evidence for the published historical reference.
The MT5 report and Balance / Equity screenshots use the original tester pixels. The surrounding brand frame and explanatory captions are presentation elements.
Other gallery panels summarize methodology, risk and setup.
All performance material in the gallery is historical simulation unless explicitly identified otherwise.
Frequently asked questions
Does this use Martingale or grid recovery?
No.
Position size is based on the configured capital basis, the fixed module allocation and the protective stop distance. It does not increase lot size in a recovery ladder after losses.
Multiple modules can still hold positions at the same time, so simultaneous positions should not be mistaken for independent account risk.
Why is a hedging account required?
Each module must retain separate ownership of its positions and exits.
A hedging account allows those module positions to remain separate.
This does not mean the portfolio is hedged against losses.
Netting accounts are not supported.
Is USD 50,000 the minimum deposit?
No.
USD 50,000 is the current Standard historical reference configuration, not a universal minimum or a personalized capital recommendation.
Actual participation depends on the broker's minimum lot, volume step, contract specification, leverage and margin.
If the calculated volume is below the broker minimum, the EA skips the trade rather than increasing the lot.
Why might my backtest differ?
The published reference uses custom Gold history, 1 Minute OHLC modeling, the disclosed spread/commission assumptions and the stated time convention.
Your broker history, spread, swap, commission, symbol specification, timezone and modeling can differ.
The published 78% history quality is not a real-tick test.
Does turning TradingEnabled off close positions or cancel orders?
No.
It blocks new entries, but existing pending orders may still fill and owned position management continues.
EmergencyStop additionally requests cancellation of owned pending orders, but it is not a flatten-all command. Open positions retain their defined exits.
Is there a live track record?
No verified product-linked live or demo-forward return record is currently presented on this page.
The published evidence is historical simulation.
If forward monitoring is added later, it should be read separately from the historical backtest.
Is this the USDJPY product with a different symbol?
No.
EdgeDriven Gold Portfolio has its own 20-module XAUUSD construction and its own evidence.
The USDJPY product uses a different portfolio structure and separate production evidence.
Their combined account performance has not been validated here.
Version 1.006
Version 1.006 restores the development-consistent Restricted USD News Guard to the production EA and adds live/future synchronization through the MetaTrader 5 Economic Calendar.
Historical Strategy Tester runs use the frozen restricted-news windows used during development. Live operation uses the dynamic calendar path for the corresponding future USD events.
The 20-module membership, strategy logic, intended relative construction and reference risk design remain unchanged.
Because the News Guard changes the executed trade sequence, the published historical reference evidence was regenerated for Version 1.006.
Risk disclosure
Leveraged trading can cause substantial losses.
Historical simulated performance does not guarantee future results. Historical drawdown is not a guaranteed future loss limit, and stop-loss orders do not guarantee the execution price.
Choose settings only after deciding what level of loss you can afford to accept.
