BRIEF: Our client commissioned us to develop a robust automated trading strategy for The Hong Kong Jockey Club betting pools.
CHALLENGE: Due to the plethora of betting syndicates and teams operating within Hong Kong, the HKJC markets have become very efficient and hence difficult to profit from.
SOLUTION: As nearly all teams still use traditional statistical techniques such as regression, to gain an edge, our solution is primarily based on advance deep learning. Specifically CNN and LSTM/DNN architectures combined with an ensemble model. Our solution also includes an intelligent portfolio optimizer.
RESULTS: We attained a 2% yield before rebates equating to a circa 100% annualized return on investment for our client.