Metals Daily is a leader in precious metals research and pricing, providing real-time data, news, and thorough analysis. They recognised AI's potential to transform the commodities market and partnered with Mercury Labs. Their goal was to develop a predictive model skilled at identifying price movements in stable commodities. This successful project demonstrated both the power of advanced analytics and the potential of AI in managing dynamic markets.
Challenges:
Crafting a cutting-edge commodities trading algorithm that capitalises on quantitative strategies for low volatility commodities markets.Sourcing and preprocessing historical data for modelling purposes.Design an algorithm that's not only predictive but also adaptive, adjusting its strategies based on real-time market changes.Incorporating technical indicators to generate actionable trading signals.Our Approach:
Explored strategies such as Moving Average Convergence/Divergence, Bollinger Bands, Stop-and-Reverse, and Hilbert-Trendline Moving Averages.Conceived a Trading Strategy Picker, meticulously analysing strategies against historical data to pinpoint the optimal blend of indicators that maximise returns while minimising risks.Rolled out a Sliding Window Player, simulating real-time trading using the chosen strategy over a designated testing period.Improved the Trading Strategy Selector, adding regular review features to keep it responsive to market changes.Outcomes:
57.68% win rate exceeding buy-and-hold benchmarks.Launched a strong backtesting framework, enabling Metals Daily to make well-informed trading choices.