Option chains contain more information than a single implied volatility: the volatility smile encodes a full risk-neutral terminal distribution, often with far fatter tails than lognormal. The article turns that into actionable percentiles by recovering density from call-price curvature (Breeden–Litzenberger), then publishing the 5–95% band.

Because second derivatives amplify bad quotes, the pipeline focuses on robustness: derive the forward from put-call parity (median near-ATM), convert to a Black-76 forward framework, fit a smooth IV curve in log-moneyness, reprice a dense strike grid, then finite-difference to density. It reports diagnostics instead of hiding flaws: negative probability mass, convexity violations, parity spread, and strike coverage.

Implementation is split cleanly: Python writes a small JSON feed per ETF; MQL5 reads it to draw lev...

👉 Read | Forum | @mql5dev

MetaTrader 5 tick volume is a per-bar quote-change counter. It reflects time-of-day effects, is incomplete mid-bar, and loses intra-bar timing. A more direct activity measure can be built from inter-quote durations.

The approach uses time between consecutive quote changes, removes the diurnal profile in fixed time bins with interpolation, then models remaining clustering via an Engle-Russell ACD recursion. The output is a per-quote activity ratio: 1 equals normal for that hour, 2 implies double the expected quote pace, 0.5 half.

Implementation details include folding same-millisecond bursts into one event and resetting after long gaps (default 5 minutes). Calibration is pure MQL5 and reports out-of-sample likelihood gains versus a rhythm-only baseline across EURUSD, GBPUSD, USDJPY, XAUUSD, US30, SPX500, with modest but statistically significant inc...

👉 Read | Docs | @mql5dev

Bridging “draw this level/pitchfork/fib” from natural language to deterministic EA logic hinges on one contract: a native chart object with a structured, unique name. Producer programs (indicators/scripts) create standard MT5 objects and encode direction, event type (TOUCH/BREAK/ZONE), role, and an ID in the object name so AIObjectTrader can recognize and evaluate them.

Two lifecycles are compared: one-shot scripts that place an HLINE or FIBOTIMES and exit, and a persistent indicator that maintains an OBJ_PITCHFORK while preserving a stable object identity during anchor updates.

AIObjectTrader scans chart objects, performs partial name recognition, dispatches supported geometry (HLINE/TREND/PITCHFORK/RECTANGLE), and tracks per-bar signals. Key boundaries remain: incomplete protocol validation, weak cross-restart identity/ownership, timing approximations, an...

👉 Read | Signals | @mql5dev

This stage extends the double-bottom toolkit from reviewing recorded examples to scanning chart history using reference-guided matching. A candidate must pass strict geometry (separation, depth, symmetry, neckline) and also resemble previously validated library samples.

The search normalizes the price path between the two lows into a fixed-length close/low profile, resampled to a set number of points and scaled by pattern depth. Similarity is computed via RMSE and converted to a bounded score; the mean of the top neighbors must exceed a configurable threshold. Results are labeled CANDIDATE/CONFIRMED/FAILED based on post-formation closes, not trade outcomes.

The EA scans completed bars on the attached symbol/timeframe, filters duplicates, renders selected detections on-chart, and exports REFERENCE_MATCHES.csv with settings, counts, structural metrics, and th...

👉 Read | AlgoBook | @mql5dev

S3CE-Net reframes streaming data as sparse, asynchronous events rather than uniform time samples, making it a strong fit for markets where the decisive signals are rare bursts: volume spikes, liquidity shocks, abrupt order-book shifts.

The pipeline bins events into fixed windows with polarity channels, then uses a lightweight spiking SEW-ResNet with residual connections to extract stable spatiotemporal features efficiently without losing the event-driven nature.

Its core upgrade is SSAM: causal spike-guided attention that links current events only to the past, adding learnable relative positional bias to preserve spatial relationships. A companion STFS sampling strategy trains on feature subsets to reduce memorization and improve regime robustness—useful for adaptive, real-time MT5 systems.

👉 Read | Forum | @mql5dev

MQL5 Algo Forge runs Forgejo behind MQL5.com auth. The version endpoint is public and reports Forgejo 16.x with a +gitea-* compatibility suffix. Feature availability depends on instance settings; checks were revalidated through 12 Oct 2026.

CI (Actions) works with real runners. Webhooks and the package registry work. Branch/tag protection, Git LFS, mirrors, and migration are enabled. SSH clone and platform-wide code search are disabled; per-repo code search works.

MetaQuotes additions include “Open with MetaEditor” via metaeditor5://checkout?repo= and a README Translate control for 15 languages, changing how documentation should be structured.

Releases enforce attachment limits and block .mq5/.mqh/.ex5/.set uploads. Workarounds: zip artifacts or publish binaries in the generic package registry and link them as external assets.

👉 Read | Calendar | @mql5dev

VelaMayor MTF is a multi-timeframe visualization indicator for MetaTrader 5, aimed at workflows based on multi-timeframe analysis, SMC, price action, and ICT.

It renders the last X candles from a selected higher timeframe (H1/H4/D1) directly on the active execution chart (M1/M5/M15), reducing the need to switch charts while keeping higher-timeframe structure visible.

Rendering supports independent border and fill colors for candle bodies, improving contrast across light and dark themes. Periods are drawn as separate boxes limited to the candles that formed within each timeframe interval, avoiding full-block stretches and preserving spacing between periods.

Wicks are centered and detached to maintain clean boundaries. Candle 0 updates tick-by-tick. Auto-recovery uses chart events and a timer to restore objects after timeframe changes or reconnects. O...

👉 Read | AlgoBook | @mql5dev

Margin Risk Simulator estimates where the current symbol would hit the broker’s margin call and stop out levels, and answers two scaling questions: the margin level after adding X lots, and the maximum extra volume that keeps margin above a chosen safety level. It is analysis-only and never sends, modifies, or closes orders.

Runs as an Expert Advisor to access OrderCalcProfit and OrderCalcMargin; Algo Trading does not need to be enabled. It draws margin call and stop out price lines in both directions, supports netting and hedging, and reads broker thresholds and modes (percent or money) from account properties.

What-if mode adds hypothetical volume at the current price, recalculates levels, and shows dashed lines for the updated thresholds. Max extra volume is rounded to broker steps/limits and requires non-negative free margin. Pending limit/stop orders b...

👉 Read | CodeBase | @mql5dev

GridPro Panel is an open-source grid trading EA for MetaTrader 5 with an on-chart control panel and full MQL5 source. All entry, averaging, and exit rules are visible in code rather than hidden in a compiled file. It targets hedging accounts and will refuse to run on netting margin mode.

The grid supports both sides or a single direction, managing buys and sells as separate baskets. First entry can be immediate or gated by a selectable signal (MA cross, RSI, Bollinger return, Stochastic). Subsequent levels follow fixed-point steps or ATR-based steps, with optional step expansion. Lot sizing supports fixed, multiplier, or arithmetic increase, with caps and optional balance scaling.

Basket exits include money TP, points-from-average TP, trailing money TP, and optional basket SL. Risk controls cover account drawdown stop, daily loss/profit limits, max lots, mar...

👉 Read | AlgoBook | @mql5dev

Indicator update focused on session structure and repeatable statistics.

Asian, London, and New York session high/low boxes are plotted in broker server time, including optional 50% midlines and configurable start/end inputs. Overnight sessions are supported and the tool runs on any symbol and timeframe.

ICT killzones can be shaded in the background, with default windows set to London 07:00–10:00 and New York 12:00–15:00 (server time). Liquidity sweep markers flag cases where price wicks through the Asian high/low and closes back inside the range.

Breakout alerts are provided via popup and push notifications, limited to once per session per side to avoid repeated triggering. A compact statistics panel reports average session range in pips over a configurable number of days.

Operation requires M1 history for accurate box building; allow data to down...

👉 Read | VPS | @mql5dev

Buffer Inspector Kb is a utility script for determining which indicator buffer contains a usable signal when wiring indicators into Expert Advisors or alert logic.

After being attached to a chart with indicators already loaded, it enumerates every buffer of every indicator, including calculation and color-index buffers that are not visible in the Data Window. It classifies each buffer as arrows, line, color index, empty, and prints the buffer numbers on-chart near the latest arrow marks. A full report is written on the chart and to the Experts journal.

The Data Window only lists drawn buffers, so hidden color buffers can shift the effective numbering used by iCustom. The script reads buffers via CopyBuffer, so the displayed indices match what an EA must use. It reads closed bars only and warns when visible arrows are likely drawn as objects, which cann...

👉 Read | Freelance | @mql5dev

Risk logic in FixedMACrossover_Part6.mq5 has grown into a 1,000+ line file where roughly 100 lines cover the MA crossover and the rest is safety machinery. Reuse across EAs has required copying large blocks, with validation and configuration spread across init code.

A refactor moves risk management into four headers under MQL5\Include\Risk Management\: shared types/conversions plus three classes by scope: trade (CTradeRisk), basket (CBasketRisk), account (CPortfolioRisk). Trade and basket tiers use config structs with defaults and validation; account keeps setter-style limits disabled until set. EAs include only needed tiers.

Behavior changes are limited: trade-tier checks run before basket/account gates, symbol volume specs with max < min are rejected, logs/comments differ, and inputs validate before indicator handles.

The shared file centralizes a moneyA...

👉 Read | Calendar | @mql5dev

Part 2 extends Prop Firm Guard with a persisted daily session baseline, preventing balance reference drift after EA restart or reattachment. A session requires explicit trader confirmation before the starting balance becomes authoritative, and the same confirmed values must restore for the same account and session day.

A daily session model is added via enums for lifecycle state and load results, plus an SDailySession structure stored in SQLite. The session key is derived from a configurable reset hour and minute in broker server time, producing a stable session-date key across rollovers.

SQLite gains a daily_sessions table keyed by account, server, and sessionDate, with consistency checks and UPSERT-based updates. Startup restores an existing session or creates an awaiting-confirmation session without auto-accepting current balance.

Money handling adds ac...

👉 Read | Forum | @mql5dev

S3CE-Net reframes market data as an event stream: sparse, meaningful spikes instead of fixed-interval bars. Built on SEW-ResNet, it adds SSAM spatiotemporal attention to connect causal event dependencies and STFS to adapt features across time and space, aiming to separate structure from noise and reduce reaction to insignificant fluctuations.

The practical focus is implementing SSAM for real-time use in MetaTrader 5 via OpenCL. A CNeuronSSAM layer wraps GPU kernels with neural “plumbing”: Q/K/V projection, multi-head attention output, residual path, and normalization, plus an intra-frame relative position bias to emphasize local relationships (useful for price levels, volume clusters, order-flow segments).

Key engineering choices target determinism and speed: static submodules avoid dynamic allocation, buffers are preplanned for OpenCL, temporary sco...

👉 Read | NeuroBook | @mql5dev

Multi-timeframe scanners and high-frequency EA rules implicitly run hundreds of hypothesis tests, so “5% significance” guarantees a steady stream of false signals even with zero edge. The article reframes this as an online multiple-testing problem where decisions are irrevocable and the number of tests is unknown.

It ports online false discovery rate control into pure MQL5 using LORD3, LORD++ and SAFFRON. These methods manage an alpha-wealth budget that is spent on each test and partially refilled by accepted signals, preventing endless overtrading on noise while keeping better power than Bonferroni-style gating.

A key implementation point is p-value validity in a live stream. The design uses rule “firings” (not every bar), enforces a strict evidence/outcome window separation, and generates causal p-values via a stationary block bootstrap to preser...

👉 Read | CodeBase | @mql5dev

A multi-symbol EA often keeps per-symbol globals and grows through duplication: separate bar times, magic numbers, failure flags, position checks, and event wiring. Adding instruments increases coupling between symbol state and MT5 event logic, which makes lifecycle paths fragile and hard to extend.

A scalable structure is an owning portfolio manager plus independent strategy nodes. Each CStrategyNode encapsulates one symbol-timeframe identity, including completed-bar gating, trading state, diagnostics, and permissions. CPortfolioManager routes OnTimer and trade-transaction events without touching node internals.

CArrayObj with FreeMode(true) provides an explicit ownership contract. Allocate an unowned node, Initialize() it, then Add() becomes the commit point where the collection assumes deletion responsibility. Failures before Add() must delete locally; Delete...

👉 Read | Forum | @mql5dev

QCL Smoothed Heikin Ashi for MT4 renders calculated candles directly in the main chart window, using a two-stage smoothing pipeline. The first pass smooths price inputs, then a second pass smooths the derived Heikin Ashi OHLC, allowing separate periods and methods for each stage.

The implementation supports SMA, EMA, SMMA, and LWMA via a custom MovingAverage routine, maintaining intermediate arrays for base, HA, and final smoothed values. Only four indicator buffers are registered for the output OHLC; all other arrays remain internal.

Candles are drawn with per-bar chart objects and color-coded for bullish, bearish, or neutral states. Output values are synthetic and should not be treated as executable prices; the active bar can repaint until close.

👉 Read | Calendar | @mql5dev

This scalper EA is designed for M1 execution and depends on consistently tight spreads and fast order routing. An ECN environment with low commissions and minimal slippage is typically required to keep transaction costs predictable.

Hosting should be handled on a VPS located close to the broker’s trading servers, targeting up to 1 ms latency. Stable connectivity, low jitter, and uninterrupted uptime are critical, as brief delays can materially affect fills, stop execution, and overall results.

👉 Read | Signals | @mql5dev

This article builds an MQL5 library for agent-based markets that lets strategies be tested on price series with known ground truth, avoiding the ambiguity of optimizing on real EURUSD history where “edge” can be pure search luck.

At its core is the Minority Game: agents pick buy/sell from fixed strategy tables and only learn which table to trust. Market quality is measured with volatility per agent (vs coin-flip baseline) and state-conditional predictability, then validated against the model’s published phase transition.

To make the synthetic series resemble real assets, the library adds a “sit out” option (grand canonical variant), producing variable participation, fat tails, and volatility clustering. A dedicated stylized-facts analyzer outputs assertable statistics (kurtosis, Ljung-Box, Hill tail index, variance ratio), enabling reproducible che...

👉 Read | Docs | @mql5dev

StopDistance is an indicator that measures per-bar volatility as (High − Low) in symbol points and visualizes it in a separate window. It adds an SMA or EMA of the range and a computed stop distance equal to average range multiplied by a configurable factor.

The goal is operational: quantify how far a symbol typically moves within one bar on the current timeframe, using the same unit used by stop-loss fields and many EAs. Unlike ATR, gaps between bars are excluded, so the metric reflects in-bar movement only.

Display logic is straightforward: blue histogram for each bar’s range, red line for the average over Period bars, and a green dashed line for Average × StopFactor. Stops set below the average often sit inside normal bar noise.

For automation, three buffers are exposed for iCustom/CopyBuffer: range (0), average (1), suggested stop (2), all in ...

👉 Read | VPS | @mql5dev

MQL5 can’t catch many critical VM faults (zero-divide, out-of-range, null access), so robust EAs need prevention, not recovery after the fact. This framework adds an exception-like discipline using precondition guards, typed error objects, cooperative TRY/CATCH macros, and consistent propagation.

Core pieces: CException as a structured error container (type, code, severity, context, timestamp), CResult<T> for explicit success/failure returns, CExceptionStack for manual call traces, and CLogger for durable terminal+file diagnostics.

Safety is enforced via GUARD_* (return Fail early) and CHECK_* (signal a THROW inside TRY). number<T> and array<T> wrappers validate division and indexing before execution. Practical benefit: predictable control flow and debuggable failures in complex trading systems—if wrappers/guards are used consistently.

👉 Read | Signals | @mql5dev