Discussing the article: "Leak-Free Multi-Timeframe Engine with Closed-Bar Reads in MQL5"

[MetaQuotes]

Check out the new article: Leak-Free Multi-Timeframe Engine with Closed-Bar Reads in MQL5.

The article presents two systematic pitfalls in MQL5 multi‑timeframe work: indicator handle leaks that exhausted resources and repainting from reading the forming bar (index 0). It introduces MTFEngine.mqh, a unified include that creates and tracks handles in one place and defaults all reads to closed bars (index 1). A D1–H4–H1 example shows how this approach keeps signals technically correct and consistent with charts.

Writing multi-timeframe (MTF) logic in MQL5 is simple in theory: create indicator handles in OnInit(), read values from the last closed bar (shift = 1), and release handles in OnDeinit(). In practice these rules live only “in the developer's head.” A single iMA() call inside OnTick() or a CopyBuffer(…,0,…) typed by habit is enough to introduce two silent but dangerous failures: unreleased indicator instances that gradually exhausted terminal resources, and “phantom” signals that fire on still‑forming bar values and vanish after the bar closes. The symptoms are subtle — growing mt5.exe memory, Journal lines like "IndicatorCreate failed", or a strategy that produces more signals in live trading than it did in a closed‑bar backtest — so many authors miss the cause.

This article presents a structural fix: MTFEngine.mqh. Rather than rely on developer discipline, the engine makes correct behavior the default. All indicator handles are created only during initialization and managed internally; ReadBuffer() defaults to bar_shift = 1 and offers safe array reads for crossover checks; IsReady() checks both buffer contents and HTF bar timestamps to avoid stale HTF reads; and a single ReleaseAll() cleans up every handle. The result is a small, explicit API that prevents the two common failure modes before they can appear.

Author: Ushana Kevin Iorkumbul