Introduction: Why the ium Collection Exists
After 13+ years managing non-public funds exceeding $1M in FX and futures markets, one fact grew to become inescapable: no single algorithm can dominate each market. Foreign exchange pairs behave in another way from gold. Gold strikes nothing like Bitcoin. And inventory indices have their very own rhythm completely — pushed by classes, gaps, and institutional order circulation that no generic EA can seize.
That realization gave beginning to the ium Buying and selling Collection — a household of 5 specialised Professional Advisors, every purpose-built for a particular asset class, but sharing a unified technological spine:
- Stratum EA — Multi-currency FX buying and selling with OCM Forex Energy filtering
- Goldium EA — Gold-native XAUUSD specialist with DXY Greenback Energy evaluation
- Indexium EA — World indices (NAS100, US30, SPX500, DAX40, UK100) with VWAP Confluence Engine
- Satoshium EA — Bitcoin BTCUSD specialist with 24/7 session administration and weekend safety
- Newsium EA — Financial Intelligence Engine offering real-time elementary evaluation to the complete collection
This text is a complete technical walkthrough of each system, algorithm, and structure resolution that powers the ium collection. In case you are a dealer who desires to grasp precisely what’s working behind the dashboard — that is your doc.
Reside sign efficiency may be tracked right here:
Technical blogs for every EA:
1. The Shared DNA: Modular Structure
Each ium EA is constructed on a modular structure consisting of 10+ internally built-in sections. Every part is a self-contained engine that may be understood independently however works in live performance with the others. The structure follows this pipeline:
H4 Macro Pattern → H1 Construction Detection → M15 POI Mapping → M5 Momentum Affirmation → Entry Sample → Confluence Scoring → RL Q-Worth Adjustment → Threat Calculation → Commerce Execution → Adaptive Studying Suggestions Loop
Right here is the part map throughout the complete collection:
| Part | Perform | Stratum | Goldium | Indexium | Satoshium |
|---|---|---|---|---|---|
| 1 | Core Definitions & Information Constructions | ✓ | ✓ | ✓ | ✓ |
| 2 | Dealer-Agnostic Image Mapper | Multi-pair FX | Gold-native | Index-native | BTC-native |
| 3 | HTF Analyzer (H4 Pattern + H1 Construction) | ✓ | Gold-tuned | ✓ | ✓ |
| 4 | POI Supervisor (OB + FVG + Liquidity) | ✓ | ✓ | ✓ | ✓ |
| 5 | M5 Momentum & Divergence Engine | ✓ | ✓ | ✓ | ✓ |
| 6 | Entry Sample Detection | ✓ | ✓ | ✓ | ✓ |
| 7 | Threat Calculator + Confluence Scoring + RL Agent | ✓ | ✓ | ✓ | ✓ |
| 8A | Newsium Integration Module | ✓ | ✓ | ✓ | ✓ |
| 8B | Asset-Particular Energy Filter | OCM Forex | DXY Greenback | VWAP + Correlation | OCM Forex |
| 9 | Canvas Dashboard (Drag & Drop, 3 sizes) | ✓ | ✓ | ✓ | ✓ |
| 10 | Alert & Notification System | ✓ | ✓ | ✓ | ✓ |
| 11 | Predominant Buying and selling Logic + Commerce Executor | ✓ | ✓ | ✓ | ✓ |
Every EA incorporates roughly 10,000+ traces of pure MQL5 code — no exterior DLLs, no Python bridges, no third-party dependencies. Every thing runs natively inside MetaTrader 5, making the collection totally appropriate with the MQL5 Market distribution necessities.
2. Good Cash Ideas (SMC) Value Motion Engine
On the coronary heart of each ium EA lies a full-spectrum Good Cash Ideas implementation — the identical institutional buying and selling methodology utilized by prop companies and hedge funds worldwide. Not like easy indicator-based techniques, SMC focuses on the place institutional cash enters and exits the market.
2.1. Multi-Timeframe Evaluation (H4 → H1 → M15 → M5)
The SMC engine operates throughout 4 timeframes concurrently, every serving a definite analytical function:
- H4 (Macro Pattern): Determines the general market bias utilizing twin EMA crossover (default 50/200), ADX pattern energy measurement, and Premium/Low cost zone classification. The H4 module solutions the query: “Are we in a bullish or bearish macro atmosphere?”
- H1 (Structural Evaluation): Detects Break of Construction (BOS) and Change of Character (CHoCH) occasions. Tracks swing highs and swing lows with configurable lookback. Identifies Order Blocks at structural break factors and maps Liquidity Swimming pools the place equal highs/lows cluster.
- M15 (Level of Curiosity Mapping): Scans for Truthful Worth Gaps (FVG) and combines them with Order Blocks to create composite POI zones. Tracks fill share on every FVG and manages POI lifecycle from creation to mitigation.
- M5 (Momentum Affirmation): RSI, Stochastic, MACD histogram, and quantity ratio evaluation. Detects each common and hidden divergences (BULL_REG, BEAR_REG, BULL_HID, BEAR_HID). Offers the ultimate momentum filter earlier than entry.
2.2. Order Blocks (OB)
An Order Block is the final opposing candle earlier than a powerful institutional transfer. The ium engine tracks every OB with full metadata:
- Course — Bullish OB (purchase zone) or Bearish OB (promote zone)
- Excessive/Low/Mid costs — The entire zone boundaries
- Timeframe origin — Which timeframe generated the OB
- Contact rely — What number of instances worth has examined the zone
- Mitigation standing — Whether or not the OB has been consumed by worth
- Age validation — OBs expire after configurable bar rely (default: 50 bars)
As much as 50 Order Blocks are tracked concurrently per image, prioritized by recency and timeframe weight.
2.3. Truthful Worth Gaps (FVG)
FVGs symbolize imbalances in worth the place institutional orders left unfilled gaps. The engine tracks:
- High/Backside/Mid — Zone boundaries of the imbalance
- Fill share — How a lot of the hole has been retraced ({a partially} crammed FVG at 50-70% is good for entry)
- Minimal measurement filter — Set to 0 for ATR-auto calibration, or specify in pips
- Confluence detection — When an OB and FVG overlap, a composite POI_OB_FVG zone is created, which receives the very best confluence rating
2.4. Break of Construction (BOS) & Change of Character (CHoCH)
These are the structural shift occasions that outline pattern continuation and reversal:
- BOS Bullish (STRUCT_BOS_BULL) — Value breaks above a previous swing excessive, confirming pattern continuation upward
- BOS Bearish (STRUCT_BOS_BEAR) — Value breaks under a previous swing low, confirming downtrend continuation
- CHoCH Bullish (STRUCT_CHOCH_BULL) — First break above swing excessive after a downtrend, signaling potential reversal to bullish
- CHoCH Bearish (STRUCT_CHOCH_BEAR) — First break under swing low after an uptrend, signaling potential reversal to bearish
The engine makes use of configurable swing lookback (default: 50 bars) to determine swing factors. Every structural occasion is time-stamped and bar-indexed for the confluence scoring system to reference.
2.5. Liquidity Sweeps & Liquidity Swimming pools
Liquidity Swimming pools type when equal highs (EQH) or equal lows (EQL) cluster on the similar worth stage. The engine identifies these as magnets for institutional cease hunts:
- Contact rely monitoring — Extra touches = stronger liquidity pool
- Sweep detection — When worth pierces by means of a liquidity pool and reverses, a high-probability commerce setup is generated
- PATTERN_LIQ_SWEEP — One of many six entry sample sorts, triggered particularly on liquidity sweep occasions
As much as 30 Liquidity Swimming pools are tracked per image concurrently.
2.6. Premium & Low cost Zones
The H4 analyzer classifies the present worth into Premium (above 50% of the vary) or Low cost (under 50%) zones. That is crucial for filtering:
- Purchase indicators are solely accepted when worth is within the Low cost zone (under honest worth)
- Promote indicators are solely accepted when worth is within the Premium zone (above honest worth)
This institutional idea alone eliminates a big share of low-probability indicators.
3. The Reinforcement Studying Agent: Q-Studying Self-Optimization
Maybe probably the most technologically superior characteristic of the ium collection is the embedded Q-Studying Reinforcement Studying engine (CAdaptiveLearner class). This isn’t a advertising gimmick — it’s a full, persistent, self-improving system that learns from each commerce the EA takes.
3.1. How the RL Agent Works
The system operates on a traditional Q-Desk structure with discretized market states:
- Characteristic Extraction: Earlier than each commerce, the engine captures a 14-dimensional characteristic vector representing the entire market state
- State Discretization: Every characteristic is bucketed into discrete zones (e.g., RSI → oversold/impartial/overbought)
- Q-Desk Lookup: The characteristic vector is transformed to a string key and appeared up within the persistent Q-Desk
- Rating Modification: The Q-value for that state adjusts the confluence rating by -4 to +4 factors
- Commerce Execution: After the commerce closes, the result (win/loss + pips) updates the Q-Desk
- Persistence: The Q-Desk is saved to CSV after each commerce and loaded on EA restart
3.2. The 14-Dimensional Characteristic Area
Every commerce’s market context is captured as a FeatureKey consisting of:
| Characteristic | States | Description |
|---|---|---|
| h4_bias | 3 | Bullish / Impartial / Bearish macro pattern |
| h1_structure | 5 | None / BOS Bull / BOS Bear / CHoCH Bull / CHoCH Bear |
| sample | 6 | Engulfing / Pin Bar / Break&Retest / FVG Fill / Liq Sweep / POI Rejection |
| session | 4 | Asian / London / NY / London-NY Overlap |
| rsi_zone | 3 | Oversold (<30) / Impartial / Overbought (>70) |
| vol_zone | 3 | Low (<0.8x) / Regular / Excessive (>1.5x) |
| spread_zone | 3 | Tight (<33%) / Regular / Large (>66%) |
| adx_zone | 3 | Weak (<20) / Average / Sturdy (>40) |
| hour_block | 6 | 4-hour blocks (00-03, 04-07, …, 20-23) |
| day_of_week | 5 | Monday by means of Friday |
| h4_zone | 3 | Impartial / Premium / Low cost |
| poi_type | 4 | OB / FVG / OB+FVG / Liquidity |
| class | 4 | FX Main / FX Cross / Crypto / Index |
| divergence | 3 | None / Bullish / Bearish |
This creates a theoretical state house of 3 × 5 × 6 × 4 × 3 × 3 × 3 × 3 × 6 × 5 × 3 × 4 × 4 × 3 = ~37 million distinctive mixtures per image. In apply, the Q-Desk makes use of an LRU eviction technique capped at 2,000 entries, maintaining solely probably the most often encountered states.
3.3. Q-Worth Replace System
The Q-value for every state combines win price deviation with common pips efficiency:
confidence = min(1.0, total_trades / 20.0) wr_signal = (win_rate – 0.5) × 2.0 // vary: -1.0 to +1.0 pips_signal = +0.3 if avg_pips > 0, -0.3 if < 0 raw_q = (wr_signal + pips_signal) × confidence × max_modifier Q_new = Q_old × (1 – learning_rate) + raw_q × learning_rate
Key traits:
- Studying price: Configurable 0.05 to 0.5 (default: 0.15) — controls adaptation pace
- Minimal knowledge: Requires 3+ samples per state and 30+ complete trades earlier than activating
- Per-symbol studying: Every image has impartial commerce counters; studying prompts per-symbol after minimal trades threshold
- EMA averaging: Common pips use exponential shifting common (80/20 previous/new) to stop single-trade domination
3.4. Adaptive Threat Sizing
Past rating modification, the RL agent additionally controls place sizing by means of an adaptive danger multiplier (0.5x to 2.0x):
multiplier = 0.5 + (win_rate – 0.3) × 3.0 // linear scale multiplier = clamp(0.5, 2.0) // security bounds multiplier = 1.0 + (multiplier – 1.0) × confidence // dampen by knowledge confidence
This implies the EA routinely sizes up on high-probability setups and sizes down on traditionally weak situations — a Kelly Criterion-inspired strategy with out the mathematical complexity.
3.5. Persistent Studying Throughout Periods
Every EA saves its Q-Desk and commerce log to separate CSV recordsdata within the MT5 Widespread listing:
- Stratum_QTable.csv / Stratum_TradeLog.csv
- Goldium_QTable.csv / Goldium_TradeLog.csv
- Indexium_QTable.csv / Indexium_TradeLog.csv
- Satoshium_QTable.csv / Satoshium_TradeLog.csv
This implies studying persists throughout EA restarts, MT5 restarts, and even system reboots. The commerce log data the total 14-dimensional characteristic context for each commerce, enabling post-hoc evaluation and technique refinement.
4. The Confluence Scoring System
Each potential commerce sign should go by means of a multi-layered scoring system that assigns factors primarily based on what number of impartial elements align. That is the engine’s main high quality gate.
4.1. Rating Parts (Most: ~21+ factors)
| Part | Factors | Situation |
|---|---|---|
| H4 Pattern Alignment | +3 | Commerce route matches H4 EMA bias |
| H1 Construction Affirmation | +3 | BOS or CHoCH confirms route |
| POI High quality | +1 to +3 | OB (+1), FVG (+1), OB+FVG confluence (+3) |
| Entry Sample | +1 to +2 | Sample kind detected (Engulfing, Pin Bar, and so forth.) |
| Momentum Affirmation | +1 to +3 | RSI + Stochastic + MACD alignment |
| Quantity Affirmation | +1 | Quantity ratio above threshold (default: 1.2x) |
| Divergence | +1 to +2 | Common or hidden divergence detected |
| Session High quality | +1 | Buying and selling throughout optimum session hours |
| Premium/Low cost Zone | +1 | Purchase in low cost, promote in premium |
| Forex/Asset Energy | +1 to +3 | Energy divergence confirms route |
| RL Q-Worth Modifier | -4 to +4 | Historic efficiency of this actual market state |
| Newsium Rating Bonus | -2 to +2 | Financial information alignment with commerce route |
A sign should attain the minimal confluence threshold to be thought of legitimate. This threshold varies by Buying and selling Mode (see Part 8).
4.2. Sign Lifecycle
Every sign progresses by means of outlined states: PENDING → ACTIVE → TP1_HIT → TP2_HIT → TP3_HIT or SL_HIT, with EXPIRED and CANCELLED as terminal states. The sign carries full metadata together with entry/SL/TP costs, sample kind, HTF bias, confluence rating, ticket numbers for every partial place, and adaptive studying keys.
5. OCM Forex Energy Filter (Stratum & Satoshium)
The OCM (Optimized Forex Matrix) Engine is an embedded forex energy calculator that analyzes relative energy throughout 8 main currencies utilizing 28 foreign exchange pairs — all calculated internally with none exterior indicators.
5.1. How OCM Works
Stratum EA and Satoshium EA each use the OCM engine (CCurrencyStrength class):
- ATR-Normalized Change Calculation: For every of the 28 normal foreign exchange pairs, the engine calculates worth change over the lookback interval, normalized by ATR to make all pairs comparable
- Forex Decomposition: Every pair’s motion is decomposed into base forex energy (optimistic contribution) and quote forex weak point (detrimental contribution)
- Averaging: Every forex’s uncooked energy is averaged throughout all pairs it seems in (sometimes 7 pairs per forex)
- Normalization: Uncooked values are rescaled to a 0-100 vary the place 50 is impartial
The 8 tracked currencies are: EUR, GBP, USD, JPY, AUD, NZD, CAD, CHF
5.2. Buying and selling Sign Integration
The OCM filter contributes to indicators in two methods:
- Rating Bonus: When the bottom forex is powerful (≥62) AND the quote forex is weak (≤38), the sign receives as much as +3 bonus factors. A partial match (just one situation met) offers half bonus.
- Course Block: When forex energy strongly opposes the commerce route (e.g., shopping for a pair the place the bottom could be very weak AND the quote could be very sturdy), the sign is blocked completely.
5.3. Newsium Enhancement
When Newsium EA is working alongside, the OCM engine blends financial information scores into forex energy calculations with a 15% weight issue. This creates a hybrid technical-fundamental energy studying.
6. DXY Greenback Energy Filter (Goldium)
Goldium EA replaces the OCM engine with a purpose-built CDXYStrengthFilter that focuses solely on USD energy — as a result of gold’s main driver is the greenback.
6.1. DXY Proxy Development
Moderately than requiring the precise DXY index, Goldium constructs an artificial USD energy studying from 4 main USD pairs (EURUSD, GBPUSD, USDJPY, USDCHF) utilizing ATR-normalized worth adjustments. This strategy works with any dealer, no matter whether or not they supply DXY as a tradeable instrument.
6.2. Gold-Particular Logic
- Inverse correlation: Sturdy USD → bearish for gold (promote bias), Weak USD → bullish for gold (purchase bias)
- Rating bonus: As much as configurable bonus factors when USD energy confirms commerce route
- Course blocking: Blocks purchase indicators when USD is powerful (above threshold), blocks promote indicators when USD is weak
- Newsium enhancement: When Newsium EA is lively, USD information scores are blended into DXY energy with 20% weight (EnhanceDXYWithNews perform)
6.3. Asian Vary System
Goldium additionally features a CAsianRange module that captures the Asian session vary (00:00-07:00 GMT) and makes use of it as a reference for London/NY session breakout detection — a way well-known amongst gold scalpers for figuring out the primary institutional transfer of the day.
7. VWAP Confluence & Index Correlation Guard (Indexium)
Indexium EA introduces two distinctive techniques designed particularly for inventory index habits:
7.1. VWAP (Quantity-Weighted Common Value) Engine
Institutional merchants rely closely on VWAP as a benchmark for order execution. Indexium calculates intraday VWAP for every image and makes use of it as a directional filter:
- Value above VWAP + Bullish sign → Rating bonus (confirming institutional shopping for stress)
- Value under VWAP + Bearish sign → Rating bonus (confirming institutional promoting stress)
- Value above VWAP + Bearish sign → Rating penalty (counter-institutional commerce)
- Value under VWAP + Bullish sign → Rating penalty (counter-institutional commerce)
7.2. Index Correlation Guard
Since US indices (NAS100, US30, SPX500) are extremely correlated, Indexium implements a CIndexCorrelation module that:
- Teams correlated indices — US indices type one group, European indices one other
- Limits publicity per group — Configurable most same-direction positions per correlation group (default: 3)
- SPX Sentiment Indicator — Makes use of SPX500 as a market-wide sentiment gauge to affect indicators on different indices
7.3. Hole Detection Engine
Inventory indices often hole at session opens (in contrast to foreign exchange). Indexium features a hole detection algorithm that identifies opening gaps and generates gap-fill commerce setups when situations align with the SMC framework.
7.4. Session-Native Design
Every index has its personal optimum buying and selling session:
- NAS100/US30/SPX500: US session (13:30-20:00 GMT) as main, with pre-market exercise
- DAX40: European session (07:00-15:30 GMT)
- UK100: London session (08:00-16:30 GMT)
Indexium applies session high quality scoring that provides bonus factors throughout probably the most liquid hours for every particular index.
8. Bitcoin-Native Structure (Satoshium)
Satoshium EA is constructed from the bottom up for Bitcoin’s distinctive market microstructure — it’s not a foreign exchange EA tailored for crypto.
8.1. 24/7 Session Administration
Not like foreign exchange which closes on weekends, Bitcoin trades constantly. Satoshium addresses this with:
- BTC Session Filter Modes:
- ALL — Commerce 24/7 with out restriction
- US_ONLY — Solely commerce throughout US hours (13:30-20:00 GMT) when institutional quantity peaks
- ACTIVE — Commerce 07:00-20:00 GMT, skipping the low-liquidity Asian session
8.2. Weekend Safety System
Bitcoin weekends usually see low liquidity, flash crashes, or manipulation. Satoshium features a three-mode weekend safety:
- AUTO — Detects whether or not the dealer offers weekend quotes and adapts routinely
- ON — All the time shut or keep away from trades earlier than weekend (Friday shut)
- OFF — Full 24/7 buying and selling together with weekends
8.3. BTC-Calibrated Threat Parameters
Bitcoin volatility is orders of magnitude increased than foreign exchange. Satoshium makes use of asset-specific calibration:
- Cease Loss vary: 500-15,000 factors (vs. 5-50 pips for foreign exchange)
- Pip worth: 100 factors = 1 pip (BTCUSD with 2-digit precision)
- Max unfold tolerance: 5,000 factors (typical BTCUSD unfold is ~1,700 factors)
- Image options: Routinely detects BTCUSD, BTCUSDm, BITCOIN, XBTUSD, BTC/USD, BTC_USD throughout any dealer
9. Newsium: The Financial Intelligence Engine
Newsium EA is the basic evaluation spine of the complete ium collection. It operates as an impartial EA that reads the MT5 built-in Financial Calendar and publishes processed intelligence to the opposite EAs by way of World Variables.
9.1. Information Evaluation Pipeline
- Calendar Loading: Reads occasions from configurable lookback (default: 4 hours previous) to lookahead (default: 24 hours future)
- Impression Classification: Maps MT5 calendar significance to a few ranges: HIGH, MEDIUM, LOW
- Deviation Calculation: Computes Precise vs. Forecast deviation as a share, with particular dealing with for small values (like 0.1% price selections) to stop share explosion
- Shock Issue: Combines forecast shock (70% weight) with pattern shock vs. earlier worth (30% weight)
- Time Decay: Current occasions have full weight; older occasions decay linearly over configurable decay interval (default: 120 minutes)
- Impression Scoring: Closing rating = shock × impact_weight × decay_factor
9.2. Clever Occasion Classification
Newsium understands that not all information is created equal. It classifies occasions into classes with particular dealing with:
- Inverse Occasions: Occasions the place increased values are detrimental for the forex (unemployment, jobless claims, commerce deficit, funds deficit). These get their scores inverted routinely.
- Inflation Occasions: CPI, PPI, PCE — Increased inflation is handled as currency-bullish (hawkish price expectations entice capital)
- Charge Occasions: FOMC, ECB, BOE price selections — Instantly impacts all asset lessons in another way
- Employment Occasions: NFP, ADP, employment change — Sturdy employment = forex bullish
- Sentiment Occasions: PMI, ISM, client confidence, ZEW — Main indicators
- GDP Occasions: Direct financial progress measurement
9.3. Asset-Particular Bias Calculations
Newsium generates specialised bias scores for every asset class:
Forex Scores (for Stratum):
- Aggregated information scores per forex on a -10 to +10 scale utilizing sigmoid normalization
- Duplicate occasion filtering (same-time releases inside 2 minutes are de-duplicated, maintaining solely the highest-impact occasion)
- Pre-news filter flag when high-impact occasions strategy
Gold Bias (for Goldium):
- USD Inverse Rating (40% weight) — Sturdy USD = bearish gold
- Inflation Rating (25% weight) — Increased inflation = bullish gold (hedge demand)
- Charge Rating (20% weight) — Increased charges = bearish gold (alternative value)
- Threat Sentiment Rating (15% weight) — A number of currencies bearish = risk-off = bullish gold (secure haven)
Index Biases (for Indexium):
- Financial Rating (60% weight) — GDP, employment, retail knowledge
- Sentiment Rating (40% weight) — PMI, confidence surveys
- Particular dealing with: Charge will increase → bearish for shares (increased borrowing prices), Excessive inflation → bearish for shares (price hike fears)
- Particular person bias for every index: NAS100, DJ30, SPX500, GER40, NIKKEI225, UK100
BTC Bias (for Satoshium):
- USD Inverse Rating (35% weight) — Sturdy USD sometimes pressures BTC
- Threat-On Rating (35% weight) — A number of bullish currencies = risk-on = BTC bullish
- Liquidity Rating (30% weight) — Dovish (price cuts) = BTC bullish, Hawkish (price hikes) = BTC bearish
9.4. Inter-EA Communication Protocol
Newsium publishes knowledge by way of MT5 World Variables with the prefix “NEWS_”:
NEWS_USD_SCORE → Forex mixture rating (-10 to +10) NEWS_USD_BIAS → Directional bias (-3 to +3) NEWS_USD_FILTER → Pre-news commerce filter lively (1.0 or 0.0) NEWS_USD_HIGH_MIN → Minutes to subsequent high-impact occasion NEWS_XAU_BIAS → Gold-specific total bias NEWS_BTC_BIAS → Bitcoin-specific total bias NEWS_NAS100_BIAS → Index-specific bias NEWS_LAST_UPDATE → Timestamp for staleness detection
Every ium EA checks for Newsium availability by searching for the NEWS_LAST_UPDATE variable and verifying it’s latest (inside 120 seconds). If Newsium isn’t working, all EAs gracefully degrade — information options are merely disabled with no impression on core SMC performance.
9.5. Commerce Impression Features
When built-in, Newsium impacts buying and selling selections by means of a number of channels:
- Rating Adjustment: Information differential between base and quote currencies modifies the confluence rating (ApplyNewsToSignalScore)
- Pre-Information Commerce Block: Blocks new trades X minutes earlier than high-impact occasions (default: 5 minutes)
- Lot Discount: Reduces place measurement close to high-impact occasions (default: 50% discount inside quarter-hour)
- SL Tightening: When information goes towards an open place, cease loss is tightened by configurable share (default: 20%)
- SL Widening: Optionally widens cease loss close to high-impact occasions to stop untimely stop-outs from information spikes
10. Six Entry Sample Varieties
After the macro pattern, construction, POI zone, and momentum filters are all aligned, the ultimate entry set off requires one among six worth motion patterns detected on the entry timeframe (M1 or M5, configurable):
| Sample | Detection Logic |
|---|---|
| Engulfing | Present candle physique totally engulfs prior candle physique. Configurable minimal protection ratio (default: 0.6x) |
| Pin Bar | Wick-to-body ratio exceeds threshold (default: 1.5x). Lengthy wick displaying rejection from POI zone. |
| Break & Retest | Value breaks a key stage and retests it as help/resistance inside configurable lookback (default: 10 bars) |
| FVG Fill | Value returns to fill a Truthful Worth Hole at 50-70% fill stage, making a high-probability re-entry level |
| Liquidity Sweep | Value sweeps past a liquidity pool (cease hunt) and reverses with sturdy momentum |
| POI Rejection | Sturdy rejection candle from inside an lively POI zone with momentum affirmation |
All sample detection parameters are configurable, permitting merchants to tune sensitivity from aggressive (unfastened patterns) to conservative (strict affirmation required).
11. WR Boosters: Win Charge Enhancement Programs
The WR (Win Charge) Booster module is a set of filters designed to eradicate statistically weak trades:
11.1. Loss Cooldown
After 3 consecutive losses on any image, the EA enters a cooldown interval (default: quarter-hour) the place no new trades are opened for that image. This prevents tilt-trading throughout antagonistic situations.
11.2. Anti-Correlation Filter
Prevents focus danger by limiting same-direction publicity:
- Max Identical-Course Open Trades: Default 5 (throughout all symbols)
- Max Per Forex Group: Default 3 (e.g., not more than 3 lengthy positions on USD-quote pairs)
11.3. Multi-TF RSI Confluence
Optionally available extra filter that requires RSI alignment throughout each M5 and M15 timeframes earlier than permitting entry.
11.4. Paper-First Commerce Filter
This distinctive system permits the EA to trace indicators as “paper trades” first earlier than committing actual capital:
- FILTER_PAPER_FIRST — New symbols begin as paper-only; as soon as they obtain a configurable minimal win price (default: 40%) over minimal trades (default: 10), they’re routinely promoted to stay buying and selling
- FILTER_ALL_LIVE — Skip paper part, commerce the whole lot stay instantly
- FILTER_ALL_PAPER — Pure sign monitoring mode, no actual trades
12. Buying and selling Modes: One EA, A number of Personalities
Every ium EA provides pre-configured buying and selling modes that alter a number of parameters concurrently for various danger appetites:
Stratum EA Modes:
| Mode | Entry TF | Min Rating | Min R:R | Character |
|---|---|---|---|---|
| Conservative | M15 | Excessive | 1.5:1 | Fewer indicators, increased high quality, wider stops |
| Balanced | M5 | Medium | 1.0:1 | Greatest mixture of frequency and high quality |
| Aggressive | M1 | Low | 0.8:1 | Most indicators, scalping strategy |
| Customized | Person-defined | Person-defined | Person-defined | Full handbook management |
Goldium EA Modes:
| Mode | Character |
|---|---|
| Fortress (M15) | Most safety, gold’s excessive volatility calls for persistence |
| Customary (M5) | Balanced strategy for normal gold buying and selling |
| Blitz (M1) | Aggressive scalping throughout high-volatility classes |
| Customized | Full handbook management over all parameters |
Indexium EA Modes:
| Mode | Character |
|---|---|
| Institutional (M15) | Mimics institutional order circulation — affected person, high-conviction |
| Customary (M5) | Balanced intraday index buying and selling |
| Momentum (M1) | Captures fast momentum strikes throughout session opens |
| Customized | Full handbook management |
13. Dealer-Agnostic Image Mapper
One of the sensible engineering options is the CSymbolMapper class. Each dealer names their symbols in another way:
- EURUSD, EURUSDm, EURUSD.i, EURUSD.r, EURUSD_SB, EURUSDpro, EURUSD.ecn
- BTCUSD, BTCUSDm, BITCOIN, XBTUSD, BTC/USD
- NAS100, USTEC, US100, NASDAQ, NAS100.money, USTECH, NDX100
- XAUUSD, GOLD, XAUUSDm, XAUUSD.i
The Image Mapper routinely:
- Tries the usual base title first
- Scans dealer’s image listing for suffix/prefix variations
- Makes an attempt different title tables (BTCUSD → BITCOIN, NAS100 → USTEC, and so forth.)
- Permits handbook override by way of customized mapping enter (“BTCUSD=BTCUSDm,NAS100=USTEC”)
- Units asset-specific parameters: pip worth, pip digits, min/max SL, unfold limits
This implies each ium EA works on just about any dealer with out handbook configuration — from IC Markets to Exness, Pepperstone to FBS.
14. Threat Administration & Commerce Execution
14.1. ATR-Primarily based Dynamic Cease Loss
Cease loss is calculated from the POI zone boundaries mixed with ATR-based buffer, then clamped to asset-specific min/max ranges. This ensures stops are all the time acceptable for present volatility situations.
14.2. Triple Take-Revenue System
Positions are cut up into three parts with configurable ratios (default: 40% / 40% / 20%):
- TP1: Conservative first goal — locks in revenue early
- TP2: Customary swing goal — captures the principle transfer
- TP3: Prolonged goal — captures runners with trailing cease
14.3. Break-Even & Trailing Cease
After TP1 is hit, the remaining place is moved to break-even with a configurable buffer (default: 1.5 pips to cowl unfold and fee). An auto trailing cease then protects income as worth advances.
14.4. Lot Sizing Modes
- Fastened Lot: Easy mounted lot per commerce
- Threat-Primarily based: Share of fairness in danger per commerce (default: 1%), with lot measurement calculated from SL distance
- Adaptive: Threat-based with RL multiplier overlay (0.5x to 2.0x primarily based on Q-value)
14.5. Security Controls
- Most positions per image (default: 1)
- Most day by day trades (0 = limitless)
- Magic quantity isolation — every EA has distinctive magic numbers to stop cross-interference
- Unfold filter — blocks trades when unfold exceeds asset-specific most
- Do-Not-Commerce hours — configurable buying and selling blackout durations
15. The Canvas Dashboard
Each ium EA incorporates a real-time Canvas-based dashboard rendered immediately on the chart utilizing MT5’s Canvas library. The dashboard offers at-a-glance info without having exterior monitoring instruments.
15.1. Three Measurement Modes
- Small (Compact) — Important info solely, minimal display actual property
- Regular — Full info together with all symbols, POI zones, sign standing
- Massive (Full) — Prolonged view with open commerce particulars, Q-learning standing, session timers
15.2. Dashboard Sections
- Header: EA title, model, mode, account information
- Image Grid: Per-symbol bias, construction, POI standing, sign rating, commerce standing
- Open Trades: Reside P&L, SL/TP ranges, break-even standing
- Statistics: Win price, complete pips, common R:R, trades right now
- Adaptive Studying Standing: Q-Desk measurement, complete trades, studying state
- Forex/Asset Energy: Visible energy bars for all currencies (Stratum) or USD energy (Goldium)
- Session Timer: Present session, time to subsequent session boundary
- Newsium Standing: Connection indicator, upcoming occasions, information bias
15.3. Interactive Options
- Drag & Drop: Dashboard may be repositioned on the chart
- Click on Interplay: Click on on image rows for detailed info
- UI Scale: Configurable scaling for high-DPI / 4K screens (100%, 125%, 150%)
16. Alert & Notification System
The alert system (CAlertManager) helps a number of notification channels:
- MT5 Popup Alerts — On-screen notifications with sound
- Push Notifications — Delivered to the MetaTrader cellular app
- E mail Alerts — For merchants who monitor by way of e mail
- Telegram Integration — Newsium EA helps direct Telegram bot messaging with customized Bot Token and Chat ID
Alert cooldown prevents notification spam (configurable, default: 60 seconds between alerts).
17. Session & Time Intelligence
Each EA features a complete session consciousness system:
- Automated GMT Offset Detection: When set to “auto” (99), the EA detects the dealer’s GMT offset routinely
- Session Home windows: London (07:00-15:00 GMT), New York (12:00-20:00 GMT), London-NY Overlap (12:00-15:00 GMT), Asian (00:00-07:00 GMT)
- Session Scoring: Trades throughout optimum classes obtain bonus confluence factors
- Do-Not-Commerce Hours: Configurable blackout durations (e.g., rollover hours 23:00-03:00 GMT)
- Pre-Shut Safety: Optionally available place closing earlier than DNT durations start
Indexium EA extends this with per-index session optimization, whereas Satoshium EA provides Bitcoin-specific 24/7 session modes and weekend safety.
18. What Makes the ium Collection Completely different
Let’s be sincere — the MQL5 Market has hundreds of Professional Advisors. Here’s what units the ium collection aside:
- Asset Specialization: Every EA is purpose-built, not tailored. Goldium isn’t Stratum with gold parameters — it has a totally totally different energy engine (DXY vs OCM), totally different session logic, and gold-tuned SMC parameters.
- Actual Reinforcement Studying: The Q-Desk RL system isn’t a advertising declare. It’s a full implementation with persistent studying, per-symbol monitoring, adaptive danger sizing, and 14-dimensional state house. Each commerce makes the EA smarter.
- Basic + Technical Fusion: No different EA collection on MQL5 Market provides a companion information engine (Newsium) that gives real-time elementary bias scoring with asset-specific modeling (gold inflation/price evaluation, BTC liquidity scoring, index financial well being monitoring).
- Zero Exterior Dependencies: No DLLs, no Python, no API keys, no cloud providers. Each line of code runs natively in MQL5. This implies full MQL5 Market compatibility, no safety dangers, and assured portability throughout brokers and VPS suppliers.
- Skilled-Grade Threat Administration: Triple TP system, break-even with unfold buffer, trailing stops, loss cooldowns, anti-correlation filters, paper-first validation, adaptive place sizing — these are the danger controls utilized by skilled fund managers, not retail retail merchants.
- 10,000+ Strains of Clear Logic: This isn’t a black field. Each resolution level is logged, each rating element is seen on the dashboard, and each commerce file is saved with full context for post-analysis.
19. The right way to Deploy the Full Collection
For max effectiveness, deploy all 5 EAs concurrently:
- Newsium EA — Connect to any chart first. It runs on a timer (each 10 seconds) and publishes knowledge globally.
- Stratum EA — Connect to any chart. It screens a number of FX pairs internally.
- Goldium EA — Connect to XAUUSD chart (auto-detects dealer image title).
- Indexium EA — Connect to any chart. It screens NAS100, US30, SPX500, DAX40, UK100 internally.
- Satoshium EA — Connect to BTCUSD chart (auto-detects dealer image title).
All EAs will routinely detect Newsium’s presence and combine financial intelligence. Every EA makes use of separate Magic Numbers, separate Q-Tables, and separate statistics recordsdata — there isn’t a cross-interference.
20. Really useful Settings by Account Measurement
| Account Measurement | Mode | Threat% | Lot Mode | RL Agent | Newsium |
|---|---|---|---|---|---|
| $500 – $1,000 | Conservative | 0.5% | Threat-based | Allow (amassing knowledge part) | Really useful |
| $1,000 – $5,000 | Balanced | 1.0% | Threat-based | Allow | Really useful |
| $5,000 – $25,000 | Balanced or Customized | 0.5-1.0% | Threat-based + Adaptive | Allow with adaptive danger | Extremely beneficial |
| $25,000+ | Customized | 0.25-0.5% | Adaptive | Full RL suite lively | Important |
Conclusion
The ium Buying and selling Collection represents a basically totally different strategy to algorithmic buying and selling on MetaTrader 5. As an alternative of 1 generic EA attempting to commerce the whole lot, we constructed 5 specialised engines — every understanding its market at a local stage — unified by a shared SMC framework, a self-improving RL agent, and an financial intelligence spine.
Each commerce the system takes makes it smarter. Each market situation it encounters is recorded and discovered from. And each financial occasion is analyzed and built-in into the choice course of in real-time.
This isn’t set-and-forget software program. It is a skilled buying and selling system that evolves with the markets.
Sources:
Copyright 2025-2026, Kaan Caliskan. All rights reserved.