Part 5 extends a dual WaveTrend crossover indicator into a full chart UI: a fast oscillator generates entry candidates, while a slower one optionally filters trades to reduce noise.

Signals now form “boxes” around crossover bars and wait for a price breakout before confirming direction. Boxes can auto-extend using average candle size, making the breakout logic adapt to volatility instead of fixed ranges.

Rendering moves beyond standard plots by using the MQL5 Canvas library. Gradient “fog” overlays visualize trend strength with smooth interpolation, while signals can be shown as triangles or labeled bubbles.

Risk tools are integrated directly into the indicator: dynamic TP/SL levels are computed from candle-size multipliers or percentage moves, then drawn as lines and summarized in an on-chart table, updated only on new signals for efficient redraws.

👉 Read | Signals | @mql5dev

#MQL5 #MT5 #Indicator

A Python trading simulator is being aligned with MetaTrader 5 semantics by implementing order_send as a single entry point for all trade operations. The request payload mirrors MqlTradeRequest fields and routes actions to either MT5 or a local tester mode.

In tester mode, pending orders, positions, and deals are stored in containers, with explicit handling for open, close, modify, delete, and history logging. Close requests validate bid/ask pricing rules and require opposing order types.

Request validation is centralized in a TradeValidators class: lot constraints, margin checks, entry price validity, stops level, SL/TP direction, max volume, max orders, and freeze level rules. A CTrade wrapper calls the simulator’s overridden order_send, enabling MT5-like trade APIs for strategy testing workflows.

👉 Read | Forum | @mql5dev

#MQL5 #MT5 #AlgoTrading

In the MetaTrader 5 beta build 5540, we have improved ONNX integration in MQL5:

• ONNX models now run significantly faster on graphics cards with CUDA support.
• New flags have been added for GPU management and logging.
• The library installation workflow has been changed: it is now installed only on the first launch of an ONNX-based program, rather than together with the platform.

In addition, we refined the rendering of text and analytical objects on charts using the Blend2D engine introduced in the previous update. Trading reports have also been improved, and the strategy tester has been made more robust.

The web version of the platform has received several improvements as well. When adjusting stop levels on the chart, you can now see an approximate profit or loss in monetary terms. The display of certain trading data has also been corrected.

Read more…

The article refactors an MT5 auto-optimization workflow to swap strategies cleanly, moving from SimpleVolumes to SimpleCandles while keeping library code independent from project code. The key change is separating “what varies per strategy” from the reusable project-creation and database plumbing.

A reliability bug is fixed in the optimization results database: concurrent tester agents can trigger “database locked” insert failures. The solution adds retry logic with an attempt counter and short random backoff to SQL execution paths, plus a shared flag to detect errors and decide whether to commit or roll back transactions.

Developer ergonomics improve via redesigned loop macros that preserve debuggability and by documenting how to extend strategy inputs (example: max spread gate) end-to-end: tester inputs, init-string serialization, constructor parsing, ...

👉 Read | Quotes | @mql5dev

#MQL5 #MT5 #AlgoTrading

Adaptive VWAP Institutional v1.0.0 is a VWAP indicator for MetaTrader 5 built with HFT and professional asset management constraints in mind, emphasizing consistent session logic and data-quality controls.

Asset class detection uses a multi-step verification flow (crypto, forex, metals, stocks, indices) to apply appropriate reset policies. Session handling includes DST-aware timezone calculations using Zeller’s Congruence, plus 17:00 New York rollover support for forex, gold, and energy.

Data conditioning adds median-volume sampling to reduce the impact of bad ticks and volume spikes. Session state persistence is handled via disk caching across restarts and timeframe changes. Core execution is optimized for O(n) behavior suitable for VPS deployments.

An on-chart diagnostic panel reports distance to VWAP (%), accumulated volume, bar count, and active tim...

👉 Read | Signals | @mql5dev

#MQL5 #MT5 #Indicator

A script is available for pulling cryptocurrency quotes from Binance and showing them in a custom symbol group inside MT5.

Enable WebRequest in Terminal Settings under Advisors and add https://api.binance.com to the allowed URLs. Copy BinanceQuotesDownloader into the terminal’s Scripts directory and run it from the Navigator.

Open View -> Symbols, then Custom -> Binance to select required instruments. Create charts via File -> New Chart -> Custom -> Binance.

Quote loading status is reported in the Experts tab. Once populated, the custom symbols can be used with indicators and standard chart analysis workflows.

👉 Read | NeuroBook | @mql5dev

#MQL5 #MT5 #script

A remote Python risk manager for MetaTrader 5 turns risk control into an always-on service, designed to protect accounts during volatility and human error. It focuses on the math behind drawdowns: deeper losses require disproportionately larger returns to recover, so limits must be enforced early.

The system connects via the MT5 Python API and checks account state every ~18 seconds. It tracks balance/equity drawdown, daily and weekly loss caps, and margin level, then can automatically close all positions and block further trading when thresholds are breached.

Implementation emphasizes reliability: a PyQt5 GUI for configuration and live metrics, multithreading to keep monitoring responsive, SQLite for state/history, autosave for restart safety, plus weekly HTML reports for post-trade analysis and iteration across multiple accounts.

👉 Read | CodeBase | @mql5dev

#MQL5 #MT5 #RiskMgmt

A risk sizing indicator calculates position risk as a percentage of account equity and returns an acceptable lot size for the configured limits.

Configuration requires two inputs: target risk percentage and stop distance in pips. Based on these values, the tool determines the trade volume that keeps the maximum loss within the selected risk threshold for the instrument.

This supports consistent position sizing across varying stop sizes and helps standardize risk management when market conditions change.

👉 Read | Quotes | @mql5dev

#MQL5 #MT5 #Indicator

This article implements Larry Williams’ swing-based volatility in an MT5 Expert Advisor, replacing single-bar range metrics with two explicit three-day swing measurements. The EA takes the larger absolute swing distance as the active volatility proxy, capturing recent market expansion independent of direction.

At each new bar open, fixed breakout levels are projected from the session open using configurable swing percentages. Trades trigger only after price crosses these levels, using market orders (no pendings), with SL sized from the same swing and TP derived via a reward multiple of defined risk.

The code is structured for reuse: new-bar detection, a dedicated struct for stored levels, precise normalization, trade-direction controls, and manual or balance-risked position sizing, while enforcing one open position at a time.

👉 Read | AppStore | @mql5dev

#MQL5 #MT5 #EA

Part 12 upgrades an MQL5 correlation matrix dashboard from a static grid into an interactive tool driven by chart mouse events. The panel can be dragged, minimized to a header-only view, and provides hover feedback on buttons and timeframe selectors to reduce misclicks during live analysis.

The dashboard adds practical controls: cycle symbol ordering by average absolute correlation (original/ascending/descending), switch between correlation coefficients and p-values, and toggle light/dark themes with immediate recoloring of cells, headers, borders, and legend.

Implementation centers on new enums, inputs, and global state for hover/drag/minimize, plus functions that reorder matrices safely, rebuild UI blocks, and attach tooltips showing symbol pair, method, timeframe, bars, correlation, and p-value.

👉 Read | CodeBase | @mql5dev

#MQL5 #MT5 #AlgoTrading

Part 36 of the MQL5 series shifts from basic API connectivity to request security for exchange endpoints such as Binance.

Core focus is signing requests with HMAC-SHA256. The signature proves authenticity and integrity: the server recomputes the same value from request parameters plus a shared secret key and rejects mismatches.

SHA256 provides deterministic 256-bit hashing with strong change detection, but it does not identify the sender. HMAC adds the secret key so only authorized clients can generate valid signatures.

In MQL5, signing is performed locally before WebRequest. Inputs are typically a query string including a timestamp, plus the secret key. Both must be converted to byte arrays (commonly UTF-8). CryptEncode is used to produce the binary signature, which is then sent as a request parameter.

👉 Read | VPS | @mql5dev

#MQL5 #MT5 #HMAC

Central Force Optimization (CFO) applies deterministic kinematics to numerical optimization. Candidate solutions act as probes; fitness defines mass, and only better probes attract worse ones. Acceleration is aggregated from pairwise forces using mass difference and distance exponents, then positions are updated via a kinematic step with boundary reflection.

A reference implementation defines S_CFO_Agent with coordinate and acceleration arrays, plus fitness initialization. C_AO_CFO manages parameters (g, alpha, beta, population size, repositioning factor, optional noise), initialization, initial distribution, acceleration computation, and position updates across epochs.

Key operational points include one-way attraction, frep decay across epochs, distance-squared checks for numerical stability, and deterministic outcomes for a fixed initial probe la...

👉 Read | Calendar | @mql5dev

#MQL5 #MT5 #AlgoTrading

Source text is missing.

Send the text to convert, and include any platform constraint (X/Twitter, LinkedIn, Mastodon) plus a target length if different from 800 characters.

👉 Read | Signals | @mql5dev

#MQL5 #MT5 Please paste the post text (or upload an image of it). I need the content to generate 2 relevant hashtags.

An MQL5 panel for MetaTrader 5 was built to estimate “fair” FX rates from a currency matrix and to flag deviations versus live quotes.

The model derives cross-rates via multiple triangular paths, applies iterative convergence with weighted updates, and uses an arbitrage threshold to filter spread-driven noise. Output includes deviation signals, candidate triangles, and a live-updating matrix view.

Limits are noted upfront: broker quote coverage, latency, and execution constraints make it non-competitive with co-located hedge-fund infrastructure. Current status is demo-only, with a robot under development and validation pending on real trading conditions.

👉 Read | NeuroBook | @mql5dev

#MQL5 #MT5 #AlgoTrading

A Python-based Strategy Tester is extended toward MetaTrader 5 parity by adding sequential tick/bar playback, account and position monitoring, and reportable performance analytics.

Configuration is moved into JSON to mirror MT5 settings (symbols, timeframe, date range, modelling, deposit/leverage). The tester validates required keys and datatypes to prevent silent misconfiguration.

Four modelling modes are implemented: real broker ticks, synthetic ticks generated from M1 bars, new-bar execution (fast, least granular), and 1-minute OHLC (speed/accuracy compromise). A unified OnTick loop calls the EA callback per tick or per bar-open, while updating orders, positions, and account state.

The tester logs orders and deals, then produces an HTML report with core metrics (profit, win rates, drawdown, factors) and plots balance/equity curves, enabling practical st...

👉 Read | NeuroBook | @mql5dev

#MQL5 #MT5 #Strategy

Market Memory Zones standardizes price-level detection using objective zone types instead of subjective support/resistance markup.

Three core classifications are used: displacement zones from ATR-multiple impulsive candles with low overlap, structure transition zones from CHoCH-style swing breaks, and liquidity sweep origin zones from stop-run candles followed by reversal. Optional unfilled/inefficiency zones cover gap-like imbalances.

Implementation is object-driven: rectangles rendered in-chart, no indicator plots. Configuration includes detection timeframe, lookback, ATR multiplier, per-zone toggles, optional volume filter, max active zones, opacity, and forward extension.

Core design elements include a ZONE_TYPE enum, a MemoryZone struct (range, time, type, strength, mitigation checks), lifecycle hooks for init/deinit, new-bar processing, and zo...

👉 Read | Docs | @mql5dev

#MQL5 #MT5 #Indicator

Cointegration and correlation can persist long enough to trade, but relationships decay. Most failures are progressive and can be monitored with rolling tools such as RWEC and IS/OOS ADF.

Some breaks are abrupt. A single event can flip hedge ratios and correlations across FX, commodities, and equities. In econometrics this is a structural break: regression parameters change in a way the model does not anticipate.

Example: NVDA/INTC after the Sep 18, 2025 partnership announcement. Correlation moved from about -0.71 to +0.75, invalidating any portfolio assuming a stable beta. RWEC can flag instability ahead of the event, but it cannot confirm permanence.

Chow tests are designed for confirmation when a candidate breakpoint is known. They compare pre-break, post-break, and full-sample regressions; large F-stat and near-zero p-value indicate a regime c...

👉 Read | VPS | @mql5dev

#MQL5 #MT5 #AlgoTrading

GSM++ reframes graph learning as a sequence problem: tokenize the graph, encode local neighborhoods, then model global dependencies. The key idea is hierarchical tokenization that preserves topology and time while shortening sequences, with optional Mix of Tokenization to select the best view per node.

The framework clarifies when architectures win: transformers handle connectivity and global context efficiently, RNNs fit localized structure and streaming, and hybrids balance both; poor ordering can cause information loss in the middle of sequences.

The MQL5 implementation adapts this to candlestick data by skipping similarity clustering and using trainable bar tokenizers. Four token types (node, edge, subgraph, unit sequences) are fused via attention pooling in a reusable CNeuronMoT module, improving feature selection for trading signals.

👉 Read | NeuroBook | @mql5dev

#MQL5 #MT5 #AlgoTrading

A simple hotkey setup can cycle through symbols in the Market Watch list.

Comma (,) selects the previous symbol. Period (.) selects the next symbol. The handler typically maps these keys to keycodes 188 and 190, often implemented as switch cases such as case 188 and case 190.

To change the defaults, look up the desired key’s keycode and replace the numeric values in those cases. Ensure the chosen keys do not conflict with existing platform shortcuts and are consistently handled for keydown events.

👉 Read | VPS | @mql5dev

#MQL4 #MT4 #script

Faster delivery rarely comes from typing speed. It comes from fewer bugs, predictable architecture, and a repeatable workflow.

Coder and developer roles differ. Developers own design, planning, implementation, and maintenance. Coders typically implement scoped changes without full system context.

Key levers: enforce consistent naming and structure to reduce rework; debug quickly using formatting tools, targeted logging, and controlled changes; use libraries as designed instead of re-implementing utilities; split complex logic into small functions to keep code readable.

Operational habits matter. Minimize distractions, read documentation instead of copy-pasting, and apply spaced repetition to retain critical patterns.

MetaEditor setup and keyboard shortcuts reduce friction, including compile, navigation, search, and debug controls.

👉 Read | AlgoBook | @mql5dev

#MQL5 #MT5 #AlgoTrading

Momentum deviation can be applied to build deviation bands around a momentum series, forming an indicator conceptually comparable to Bollinger Bands but based on momentum statistics rather than price volatility.

The bands are derived from momentum with a deviation measure that expands and contracts as momentum dispersion changes. This makes the envelope responsive to shifts in acceleration and deceleration.

Operationally, it can be used in a similar manner to Bollinger Bands: monitoring momentum moves toward or beyond the outer bands for potential mean-reversion setups, and watching sustained runs near a band as a sign of persistent directional pressure.

👉 Read | Quotes | @mql5dev

#MQL5 #MT5 #Indicator