ML BASED FILTER TO INCREASE THE HIT RATIO OF TECHNICAL TRADING SYSTEM
Problem: A broker had developed and automated his technical trading system and was running a managed trading product for his clients. Being a momentum system, the hit ratio was very low and many of his clients lost faith in the system and redeemed their funds. He tried many heuristic optimizations, but couldn’t increase the strategy’s hit ratio.
Solution: We built a Machine Learning based filter for his trading system. This system sat in front of his trading system, classified each signal based on the current market conditions and let only high probability signals pass through.
Results: Hit ratio of the trading system increased from 40% to 75%, without dampening returns.
Testimonial: The broker is now trying to learn Machine Learning himself.