Adewumi Babatunde Gbadebo / Profil

The article rebuilds a Gaussian Process forecaster in native MQL5 around market-state inputs and direct horizon log-return targets. It employs a Matern 3/2 kernel, symmetric Cholesky with adaptive jitter, standardized targets, and a log‑space grid search that avoids state side effects. The EA trades only when az‑score over observed variance clears a threshold and the edge beats the live spread. Calibration is verified in Strategy Tester.

Standard HMMs assume geometric, memoryless state durations, which poorly match real market phases. This piece implements a duration‑explicit Hidden Semi‑Markov Model natively in MQL5, with per‑state sojourn distributions and a residual‑time forward filter. Parameters are fit offline via EM and loaded through a compact JSON manifest. The EA for XAUUSD M5 uses expected remaining duration to gate entries and exits, helping hold trends while avoiding late entries near regime exhaustion.

Fixed exits ignore state changes while a trade is open. We implement a discrete-time competing-risks model entirely in MQL5, estimate cause-specific hazards for take-profit and stop-loss via Newton–Raphson on a person-period dataset with time-varying features, and turn cumulative incidence into a bar-by-bar hold-or-close rule, then test it against fixed take-profit/stop-loss with identical entry logic.