Perceptron Distance MA: Automated MT5 trading with verified performance and discipline

Perceptron Distance MA review for 2026 provides a focused performance analysis of an MT5 expert advisor built by Roman Poshtar. This review highlights how the EA combines a perceptron recruitment algorithm with moving average distance features, presenting out-of-sample tests and live-account snapshots to show realized returns and drawdowns. The performance analysis in this review is grounded in verified trade logs and documented parameter sets rather than anecdotal results. What makes Perceptron Distance MA unique is the emphasis on feature engineering and selection before training the perceptron ensemble. Roman Poshtar describes the selection and construction of features in published articles, and the EA implements those principles by translating geometric and statistical indicators into binary inputs for a lightweight neural gate. The algorithm evaluates combinations of distance metrics and multi-timeframe trends to determine entry probability, then applies strict exit rules to lock profits and limit losses. The review and analysis show that the system behaves predictably across correlated FX pairs and major indices when volatility conforms to historical patterns. Risk management is explicit within the advisor: position sizing is volume-scaled to account equity and configurable maximum drawdown thresholds are enforced. Perceptron Distance MA typically reduces exposure during extended low-liquidity periods and increases selectivity when market structure is clear. Expected performance characteristics include moderate trade frequency, steady equity accumulation in favorable regimes, and occasional flat periods during range-bound conditions. The detailed review notes that parameter optimization and broker execution latency materially affect realized returns, and it recommends live-forward testing on MT5 before allocating significant capital.