AI-Powered Harness Racing Betting Assistant
Budget: $750 – $1,500 AUD
Harness Racing Data Analysis & Betting Assistant (Windows App) Websites used are Www.harness.org.au Www.tab.com.au https://www.racingqueensland.com.au/industry/harness/harness-sectionals https://www.harness.org.au/form-index.cfm
I’m looking to build a Windows-based desktop application that helps users analyse harness racing data and manage their personal betting activity more efficiently.
a data-driven racing assistant that lets users: • Import and view harness racing form data (scraped from public sources like harness.org.au)
• Calculate AI-based price estimates for each horse based on recent form, sectional times, driver changes, and other factors
• Manually input and compare their own prices for each runner
• Track meetings via a built-in calendar interface
• Log in to their TAB account (optional) to manage betting activity
• Keep detailed betting history, profit & loss, and reporting dashboards Key features:
• Desktop GUI using PyQt or another Windows framework
• SQLite backend to store user accounts, horse form, and betting history
• Form scraper module (harness.org.au)
• Odds comparison interface (with tab.com.au)
• Smart filters based on AI vs user-defined prices
• P&L tracking tab This is designed for private use only by individual users who already have their own TAB accounts. No scraping of private data or unauthorized API use is involved.
1. AI Horse Pricing System (Form-Based Prediction Model) Input Features: Scraped from harness.org.au • Last 5 starts (positions, margins, competition class) • Gear changes (blinkers, hopples, etc.) • Driver upgrades/downgrades (ranking of driver based on historical win rate or performance) • Sectionals (last 400m/800m/overall times) • Track condition (fast, wet) • Barrier draw AI Model Suggestions: • Model Type: Gradient Boosted Trees (like XGBoost) or a lightweight neural network (TensorFlow or PyTorch). • Training Data: Historical race results (1-2 years), features extracted above. • Output: Predicted winning probability (converted to fair odds). ⸻ (TAB Integration) Tasks: • Poll TAB odds from tab.com.au as soon as markets open (6-8AM). • Compare AI fair price vs TAB odds. • Bet if odds ≥ value threshold (e.g., 15% edge). • Log all bets for P&L. API Access: • TAB does not have a public betting API. You’d either need: • Automation (e.g. Selenium or Playwright) to mimic a human logging in and placing a bet (risky, fragile). • Partner access from TAB for full integration (requires approval). ⸻
3. Calendar View + Form Display Data: • Scrape from harness.org.au/racing/fields: • Daily meetings • Race times • Horse form • Store in a small database (PostgreSQL or SQLite). • Display with a calendar UI (use JavaScript frontend framework like React, or a Python dashboard like Streamlit/Dash). ⸻
4. Profit & Loss Tracker Functionality: • Record bets placed (horse, odds, stake, result). • Auto-log robot-placed bets. • Manual override/input for cash bets. • Display: • Total profit • ROI • Win % • Daily/weekly/monthly breakdown
I’m looking to build a Windows-based desktop application that helps users analyse harness racing data and manage their personal betting activity more efficiently.
a data-driven racing assistant that lets users: • Import and view harness racing form data (scraped from public sources like harness.org.au)
• Calculate AI-based price estimates for each horse based on recent form, sectional times, driver changes, and other factors
• Manually input and compare their own prices for each runner
• Track meetings via a built-in calendar interface
• Log in to their TAB account (optional) to manage betting activity
• Keep detailed betting history, profit & loss, and reporting dashboards Key features:
• Desktop GUI using PyQt or another Windows framework
• SQLite backend to store user accounts, horse form, and betting history
• Form scraper module (harness.org.au)
• Odds comparison interface (with tab.com.au)
• Smart filters based on AI vs user-defined prices
• P&L tracking tab This is designed for private use only by individual users who already have their own TAB accounts. No scraping of private data or unauthorized API use is involved.
1. AI Horse Pricing System (Form-Based Prediction Model) Input Features: Scraped from harness.org.au • Last 5 starts (positions, margins, competition class) • Gear changes (blinkers, hopples, etc.) • Driver upgrades/downgrades (ranking of driver based on historical win rate or performance) • Sectionals (last 400m/800m/overall times) • Track condition (fast, wet) • Barrier draw AI Model Suggestions: • Model Type: Gradient Boosted Trees (like XGBoost) or a lightweight neural network (TensorFlow or PyTorch). • Training Data: Historical race results (1-2 years), features extracted above. • Output: Predicted winning probability (converted to fair odds). ⸻ (TAB Integration) Tasks: • Poll TAB odds from tab.com.au as soon as markets open (6-8AM). • Compare AI fair price vs TAB odds. • Bet if odds ≥ value threshold (e.g., 15% edge). • Log all bets for P&L. API Access: • TAB does not have a public betting API. You’d either need: • Automation (e.g. Selenium or Playwright) to mimic a human logging in and placing a bet (risky, fragile). • Partner access from TAB for full integration (requires approval). ⸻
3. Calendar View + Form Display Data: • Scrape from harness.org.au/racing/fields: • Daily meetings • Race times • Horse form • Store in a small database (PostgreSQL or SQLite). • Display with a calendar UI (use JavaScript frontend framework like React, or a Python dashboard like Streamlit/Dash). ⸻
4. Profit & Loss Tracker Functionality: • Record bets placed (horse, odds, stake, result). • Auto-log robot-placed bets. • Manual override/input for cash bets. • Display: • Total profit • ROI • Win % • Daily/weekly/monthly breakdown
Related categories:
PHP
Web Scraping
Software Architecture
AI (Artificial Intelligence) HW/SW
Desktop Application