🕷️️ Job Radar • SCRAPING

freelancer.com 2026-04-25 🟡

🔹 Horse Racing Outcome Prediction
👤 Client: 🇦🇺 Queenscliff, Australia Member since 2026-04-25
💰 Price: $15 / hr Average bid
🚩 Problem: Generate accurate win/place probabilities and identify value bets by comparing model outputs against live market prices.
📦 Existing: Not specified

Specifications:

[Target] Predictive Modeling: Train ensemble models (XGBoost, LightGBM) or deep learning architectures for probability output.
[Method] Feature Engineering: Develop complex features including sectional time normalization, WFA adjustments, and Beyer-style speed figures.
[UI/UX] Not applicable
[Stack] Python, Pandas, SQL, Betfair API, XGBoost, LightGBM, TensorFlow/Keras
[Security] Ensure data privacy and secure API integrations.
[Format] JSON for model outputs, CSV for feature engineering scripts

Workflow:

1. Data Ingestion: Scrape or ingest historical and real-time race data via APIs/web scraping.
2. Feature Engineering: Normalize sectional times, adjust weights, calculate Beyer-style speed figures.
3. Model Training & Tuning: Train ensemble models using XGBoost/LightGBM or deep learning architectures.
4. Leakage Control: Implement temporal splits to prevent look-ahead bias.
5. Automated Pipeline: Develop end-to-end pipeline for real-time feature generation and automated bet execution.
6. Backtesting: Run walk-forward backtests against historical data to validate +EV strategy.