When you have all the information,
racing is no longer a game of chance
10 years of French horse racing history (1996-2005) structured and organized for quantitative analysis.
Vector Data Model
A perfect relational mesh connecting major entities for high-frequency extraction.
Table : Races
Racetrack, Weather, Prize, Distance
Table : Actors
Odds Spread, Shoeing, Handicap
Table : Yields
Human Entities
Horses, Jockeys, Trainers
An Unprecedented
Volume
A raw relational database. Every race, every actor, every yield has been captured, normalized, and indexed to enable instantaneous joins.
SELECT c.id, a.cheval_id, r.quarte_ordre
FROM courses c
INNER JOIN acteurs a ON c.id = a.course_id
INNER JOIN rapports r ON c.id = r.course_id
WHERE a.deferre = 'Des 4'
AND a.ecart_pmu_pmh > 10
AND c.meteo = 'Bon'
AND a.recul = TRUE;
Query
40+ Parameters
Having data is not enough. The Huggy system integrates querying algorithms designed to instantly cross-reference over 40 dynamic fields.
-
Extreme Granularity
Shoeing, blinkers, 25m handicap starts, weather, autostart type, sex, age, pedigree history.
-
Indexing Performance
Database integrity is ensured by an optimized architecture, preventing relational table corruption during heavy queries.
From Archive to
Prediction
The past models the future. By isolating Typical Profiles over 10 years, the engine deploys statistical algorithms to validate your strategies.
Bernoulli's Law
The system does not promise certainty; it calculates mathematical expectations. Using the binomial distribution, Huggy computes the exact probability of profiting over a sequence of N races.
Monte-Carlo Simulation
Generate thousands of random trajectories based on the historical variance of your queries. Identify the maximum Drawdown before risking real capital.
- > Yield Calculation
- > Standard deviation of losing streaks
- > Bankroll Optimization
A Goldmine for
Machine Learning
In the AI era, access to structured data is the most precious resource. This archive possesses two irreplaceable assets:
1. Organic Alpha
Current markets are saturated with algorithmic arbitrage. This database contains the Smart Money Spread captured before the bot invasion.
2. NL2SQL Corpus
With 40+ parameters per table, it's the perfect playground to train Reasoning Models to transform business questions into SQL joins.
Licensing & Acquisition
Native export in SQL (.sql) or CSV for Data Warehouse integration.
Analyst Segment
R&D Pack
- 2 years of sample data
- Full relational structure
- Limited PMU/PMH Spread
Historical Deep-Dive
Full Archive
- Full 1996-2005 Database
- PMU/PMH Spread Included
- Native SQL structured access
- 40+ Variables per race
Enterprise / Quants
Reasoning Engine
- Intellectual Property Transfer (Source Code)
- Bernoulli & Monte-Carlo Module
- Original Query Builder