rizalfarhan.
Back to Study Cases
R&D•2024•5 min read

Algorithmic Trading – XAUUSD (Gold) Expert Advisor

LSTMPythonZeroMQWebSocketsMT5Protobuf

Approach

LSTM Deep Learning models for time-series forecasting with ZeroMQ/WebSockets for low-latency execution in the Python-to-MT5 pipeline.

Technical Insight

In high-frequency trading, Slippage is the primary enemy. Optimize the Python-to-MT5 pipeline by using Protocol Buffers (Protobuf) instead of JSON for data serialization to reduce payload size and decrease execution latency by up to 40%.

Key Learnings

  • Protobuf reduces payload size and latency vs JSON in trading pipelines
  • LSTM models suit time-series forecasting for XAUUSD
  • ZeroMQ/WebSockets enable fast execution with minimal slippage