Important Announcement:
We would like to inform you that the Gold AI Trading Research project has officially ceased further development.
This decision follows the system’s trading performance over the past three years of forward testing, which did not meet expected outcomes,
and due to a lack of continued support.
Research & Development of automated gold trading system
Project Summary
The Gold AI Trading Project was born by an investor group at the Malaysia Trading Event in 2017 who wanted to build a highly efficient automated gold trading system.
The goal of this trading system is that “it has to apply an AI technology” that can learn and automatically adjust trading strategies according to changing market conditions.
- Use AI (Artificial Intelligence) and Machine Learning technologies
- For use with MetaTrader 5 Platform
- Tested with statistics dating back more than five years
- It is estimated that this project will take approximately seven years to complete.
Current Status
- Machine Learning processes: 57%
- Estimated completion time: January 2026
Project Objectives
1. Build a highly efficient automated gold trading system
- Must pass a test with at least 5-year historical statistics
- Must use detailed test data at the level of 100% Real Tick on MetaTrader 5 Platform
2. Must be stable in the long term
- Expected average annual profit: 10%
- Maximal drawdown: less than 25%
3. Must be adaptable
- Must use AI technology to learn from the actual market and adjust the trading strategy by itself.
Processes of Building the Trading System
The trading system is built by applying complex algorithms based on AI and Machine Learning concepts.
1. Machine Learning
- Machine Learning is applied by letting a robot learn trading strategies by itself.
- Developers will specify conditions, variables, and targets, then let the robot test conditions and variables until it finds the strategies that can achieve the target.
- For that process, the robot has to go through a huge number of variables and conditions. The estimated minimum is approximately 100,000 different variables.
2. High-performance Servers
- Machine learning processes are run automatically every second on high-performance cloud servers. We choose cloud servers in Singapore that uses high-frequency compute technology that can help increase processing speed.
- *Update: At present, we move to the data center in London to reduce latency.
System Testing Reports of Beta Version
- Download Strategy Tester Report
Forward Testing & Machine Learning
- Machine Learning Processes: 57%
Key Questions
1. When will this trading system be available?
- It is currently not possible to provide an exact release date
- As of now, the project is approximately 57% through the machine learning process.
- The system is expected to be available in January 2026.
2. What are the conditions for requesting to use this trading system?
- The trading system is not available for sale.
- However, it is offered free of charge through the following platforms: Copy Trading, PAMM, and Social Trading.
- Access is granted to investors who open a trading account through the recommended broker as listed on our website.