EKF Portfolio EA
- エキスパート
- Olalekan Damilola Aturaka
- バージョン: 1.3
EKF Portfolio Trader — MT5 Marketplace
EKF PORTFOLIO TRADER
EKF Portfolio Trader
Extended Kalman Filter · Multi-Symbol · MetaTrader 5
⚡ VERSION 1.03Overview
EKF Portfolio Trader is an institutional-grade automated trading system for MetaTrader 5 that applies the Extended Kalman Filter (EKF) a recursive Bayesian estimation algorithm, to dynamically track and predict price-state trajectories across an entire portfolio of instruments simultaneously.
Unlike conventional indicator-based EAs, this system treats price as a hidden state variable corrupted by measurement noise. The EKF continuously refines its estimate of the true underlying trend using prediction and update steps, producing statistically optimal, noise-filtered signals with quantified uncertainty (covariance). Trades are only entered when the filter's confidence meets a defined threshold, eliminating impulsive entries on random market noise.
It incorporates multi-symbol portfolio management, dynamic position sizing, real-time risk controls, and full MT5 compatibility, ready to deploy on forex, metals, indices, or any supported CFD instrument.
Key Features
📡
EKF Signal Engine
Recursive state estimation with a 3-component model (price, velocity, acceleration). Tracks the true trend with covariance-bounded confidence, adapting in real time to changing volatility.
🗂️
Portfolio Mode
Manages multiple symbols from a single EA instance. Independent filter states per instrument with unified risk allocation and correlation filtering across the book.
📐
Z-Score Entry Gate
Trade signals are gated by the filter's normalized innovation z-score. Only statistically significant deviations from the predicted state trigger an entry.
🛡️
Layered Risk Controls
Dynamic lot sizing, daily loss circuit-breaker, max concurrent positions, pre-order margin verification via OrderCalcMargin, and free margin checks before every order.
🔄
Adaptive Noise Tuning
Process noise (Q) and measurement noise (R) matrices are fully exposed as inputs, letting you tune the filter between fast trend-following and smooth mean-reversion behavior.
📊
Full MT5 Compatibility
Strategy Tester ready with custom OnTester() criterion. Auto-detects symbol fill modes and supports both hedging and netting account types.
How It Works
1
Predict On each new bar, the EKF projects the prior state estimate (price, velocity, acceleration) forward through the process model, inflating the covariance matrix according to Q.
2
Update The new price observation is compared against the prediction. The Kalman Gain balances trust between the model and the observation based on their relative uncertainties, then corrects the state estimate.
3
Signal Generation The normalized innovation z-score is calculated. If it exceeds the configured threshold and passes the trend, volume, and correlation filters, a directional signal is confirmed.
4
Safe Execution Before any order is sent, the EA verifies account balance, free margin, and uses OrderCalcMargin to confirm sufficient funds. Only then is the order dispatched with ATR-based SL/TP levels.
Input Parameters
📁 Portfolio Configuration
| Parameter | Default | Description |
|---|---|---|
| InpSymbols | EURUSD,GBPUSD, XAUUSD,USDCHF, AUDUSD | Comma-separated list of symbols to trade simultaneously. Up to 10 symbols supported. |
| InpTimeframe | M30 | Analysis timeframe used for bar processing and all indicator calculations. |
| InpMaxSymbols | 10 | Hard cap on the number of symbols the EA will manage concurrently. |
🔬 EKF Model Parameters
| Parameter | Default | Description |
|---|---|---|
| InpPhi | 0.85 | State transition decay factor applied to velocity and acceleration components. Range 0.7–0.95. Lower values mean faster decay to zero. |
| InpProcessNoise1 | 0.0001 | Process noise variance for the price state (Q₁₁). Higher = filter tracks price changes faster but is noisier. |
| InpProcessNoise2 | 0.00001 | Process noise variance for the velocity state (Q₂₂). |
| InpProcessNoise3 | 0.000001 | Process noise variance for the acceleration state (Q₃₃). |
| InpMeasurementNoise | 0.001 | Observation noise variance (R). Higher = smoother filter output, slower response to price changes. |
📶 Entry Signal Configuration
| Parameter | Default | Description |
|---|---|---|
| InpEntryZThreshold | 1.5 | Minimum normalised innovation z-score required to trigger a trade. Range 1.0–3.0. Higher = fewer but higher-conviction entries. |
| InpMinBarsBetweenTrades | 10 | Minimum number of bars that must pass after a trade before the EA will enter again on the same symbol. |
| InpUseVolumeFilter | true | Enable volume confirmation. Signals are rejected unless current bar volume exceeds the rolling average by the multiplier below. |
| InpMinVolumeMultiplier | 1.2 | Current bar volume must be at least this multiple of the 19-bar average to pass the volume filter. |
🛡️ Risk Management
| Parameter | Default | Description |
|---|---|---|
| InpRiskPercent | 1.5% | Account equity risked per trade as a percentage of balance. Lot size is calculated dynamically from this value. |
| InpMaxDailyRisk | 5.0% | Maximum cumulative daily loss as a percentage of balance. Trading halts for the rest of the day when this is reached. |
| InpMaxDailyTrades | 20 | Maximum number of new trades allowed per calendar day across all symbols. |
| InpMaxOpenPositions | 5 | Maximum number of concurrent open positions managed by this EA at any one time. |
| InpMaxCorrelation | 0.7 | Prevents opening positions on highly correlated symbol pairs in the same direction, limiting portfolio concentration risk. |
📏 Stop Loss & Take Profit
| Parameter | Default | Description |
|---|---|---|
| InpATRPeriod | 14 | ATR period used to calculate volatility-adaptive stop and target levels. |
| InpSLMultiplier | 2.5× | Stop-loss distance = ATR × this multiplier. |
| InpTP1Multiplier | 2.0× | First take-profit level = ATR × this multiplier. |
| InpTP2Multiplier | 3.5× | Second take-profit level = ATR × this multiplier (used as primary TP in execution). |
| InpTP3Multiplier | 5.0× | Third (extended) take-profit level = ATR × this multiplier. |
| InpUseTrailingStop | true | Enable ATR-based trailing stop to lock in profits as the trade moves in favour. |
| InpTrailingMultiplier | 1.5× | Trailing stop distance = ATR × this multiplier. |
🔍 Additional Filters
| Parameter | Default | Description |
|---|---|---|
| InpUseTrendFilter | true | Enable EMA trend alignment filter. Signals are only taken in the direction of the prevailing EMA trend. |
| InpEMAFast | 20 | Fast EMA period for trend direction detection. |
| InpEMASlow | 50 | Slow EMA period for trend direction detection. |
| InpMinADX | 20.0 | Minimum ADX value required to consider a trend valid. Signals are skipped in ranging/choppy markets below this level. |
| InpADXPeriod | 14 | Period for the ADX trend strength calculation. |
🕐 Trading Hours
| Parameter | Default | Description |
|---|---|---|
| InpUseTimeFilter | true | Restrict trading to a defined session window using broker server time. |
| InpStartHour | 8 | Hour at which the EA begins accepting trade signals (broker time, 0–23). |
| InpEndHour | 20 | Hour at which the EA stops accepting new trade signals (broker time, 0–23). |
⚙️ System Settings
| Parameter | Default | Description |
|---|---|---|
| InpMagicNumber | 987654 | Unique identifier stamped on all orders. Change this if running multiple instances on the same account. |
| InpTradeComment | EKF_Portfolio | Comment string attached to all trade orders for identification in the terminal history. |
| InpDebugMode | false | Enable verbose logging to the MT5 Experts log. Useful for diagnostics; disable in live trading. |
⚠ Version Notice: Past performance in backtesting does not guarantee future results. Trading leveraged instruments carries significant risk of loss. Only trade with capital you can afford to lose.
EKF · PORTFOLIO TRADER · v1.03
MetaTrader 5 · MQL5 Marketplace · New Release
