Built in Python programming language, this model has been trained with many years of data from the Hong Kong Jockey Club. By churning through tens of thousands of runners and their myriad variables, the model learns to generalise and can then make predictions about the probabilities of horses’ positions in future races. As soon as the racecard for a future meeting is live, the pertinent data for every runner; the going, its form, jockey data, trainer data, etc. is scraped and fed into the model. Predictions are then generated and if certain probability thresholds are crossed then picks will be selected. There are roughly five or so picks per meeting which generates ~40 betting opportunities per month between Happy Valley and Sha Tin.

With extensive back-testing the model seems to generate consistent profit, the most promising with place pool betting. Frustratingly, I am based in Europe and don’t have the right to bet in the actual HKJC pools and exotics. I have to settle for each way bets at starting price (with fixed fraction place odds) through a UK botravelrummy.comokmaker. While not ideal oddswise, this still has positive ROI.

I hope you enjoy these free horse racing tips for Hong Kong.

Thanks for reading and best of luck!

James