Harnessing large-scale genetics, environmental and market data to help optimise peak performance in thoroughbred horses.
M2M Data Limited is a Hong Kong–based data science firm specialised in the analysis of large-scale equine performance data. From our offices in Lee Garden, we combine quantitative research, genetics, and advanced computation to help clients better understand the drivers of peak performance in thoroughbred horses.
Our clients include professional ownership groups, syndicates, advisors and financial institutions seeking robust, evidence-based insight into equine performance risk and opportunity.
We do not provide consumer services or public wagering products. Our focus is strictly on data, models and decision-support tools.
To construct transparent, repeatable, data-driven views of equine performance that can support long-term planning, risk management and capital allocation.
We analyse how “the crowd” historically prices and ranks horses, then measure where those expectations have been systematically over- or under-optimistic.
By blending public expectations with our proprietary datasets (genetics, training and race conditions), we generate refined probability estimates for a wide range of race outcomes.
Our models are built to be robust over thousands of races, not just individual events – focusing on statistical convergence, risk control and repeatability over time.
Race-Level Data (15+ years) – Historical results, sectional times, fields, margins, and benchmark performance across major jurisdictions.
Contextual Factors – Course configuration, distance, going/track condition, weather, draw, field size and race type.
Horse, Jockey & Trainer Profiles – Longitudinal win/loss records, “hotness” indicators, course and distance preferences, training patterns and layoff metrics.
Proprietary Genetic Signals – Bloodlines, classification markers and identified genetic attributes associated with speed, stamina and recovery.
Market & Sentiment Indicators – Aggregated measures of how public expectations have evolved pre-race, enabling us to isolate cognitive and emotional biases without referencing any individual participant.
Form momentum (“hotness”) over multiple race and time windows
Distance and going preference
Course affinity
Jockey and trainer effectiveness over different horizons
Genetic “boosts” and constraints relevant to particular race types
These features feed into our multi-factor models, allowing us to understand why a horse performs the way it does, not just whether it wins.
High-frequency data ingestion and cleaning pipelines
Cloud-based compute with containerised model deployment
Probabilistic and econometric models designed for stability across large samples
Robust back-testing framework with variance and drawdown analysis
Our research programme focuses on out-of-sample validation and risk-aware performance measurement, rather than headline numbers.
Historical tests on major race markets show that our enhanced probability estimates can generate materially improved risk-adjusted outcomes over ~1,750+ race samples.
We provide institutional partners with:
Scenario analysis and stress-testing under different market conditions
Expected return ranges and probability distributions
Time-to-convergence estimates (how many races are required for the edge to reliably express itself)
All performance information is delivered for analytical and educational purposes to support internal decision processes. It is not a public offer, recommendation or invitation to any form of retail participation.
Names anonymous; credentials real. The following are examples of the expertise behind M2M Data Limited.
BSc in Agricultural Biochemistry and PhD in Molecular Genetics from leading UK universities
Former professor at a Royal-chartered agricultural university
Founder of an equine genetics research company focused on performance traits in thoroughbreds
MEng in Computer Systems and Electrical Engineering
Over a decade of experience as a senior software and systems engineer at one of the world’s largest quantitative sports-data organisations
Specialist in building high-availability, low-latency data platforms for real-time decision support
MS in Theoretical Physics and PhD in Particle Physics
Research background at leading European and Japanese institutions, including work at a major international physics laboratory
Former consultant to a global market-making firm, bringing institutional-grade quantitative methods to equine performance modelling
BSc in Computer Science
Academic research in probability theory and econometric models applied to horse-racing datasets
Focuses on model implementation, feature engineering and continuous performance monitoring
Together, the team blends genetics, physics-grade mathematics and large-scale software engineering to create rigorous, testable models of equine performance.
M2M Data Limited is committed to operating to standards that institutional counterparties – including banks and regulated financial institutions – can be comfortable with.
No Retail Products – We do not target or service the general public.
No Consumer Wagering Services – Our work is focused on data analytics, research and institutional-level decision support only.
Data Governance – Strong controls around data privacy, security and access; clear separation between raw data, models and client-facing outputs.
Risk & Model Governance – Documented model assumptions, periodic reviews, independent validation and conservative interpretation of results.
Ethical Use of Data – Commitment to animal welfare, responsible use of information and alignment with regulatory best practices in all jurisdictions in which our clients operate.
Lee Garden District, Hong Kong
Email: info@m2mgroup.io
Email: partners@m2mgroup.io