Bet261 Prediction Engine

Autonomous AI Prediction Engine for Bet261 Virtual Spanish League.

Quick Start

Double-click to start:

• start_all.bat - Start Backend + Frontend together
• start_backend.bat - Start API only
• start_frontend.bat - Start Dashboard only
• start_orchestrator.bat - Start autonomous engine

Access:

• Dashboard: http://localhost:3000
• API: http://localhost:8000
• API Docs: http://localhost:8000/docs

Setup (First Time)

Backend

cd backend
py -m pip install python-dotenv pydantic pydantic-settings sqlalchemy playwright pandas numpy scikit-learn xgboost lightgbm scipy fastapi uvicorn python-multipart aiohttp httpx loguru python-dateutil joblib
py -m playwright install chromium
py init_db.py

Frontend

cd frontend
npm install

Overview

This system continuously monitors Bet261 virtual Spanish league matches, collects historical data, builds predictive models, and identifies value betting opportunities.

Architecture

The system consists of 14 integrated subsystems:

1. Browser Automation Engine - Playwright-based Chrome automation
2. Data Extraction Engine - Parses and deduplicates match data
3. SQLite Database - Stores teams, matches, predictions, bets
4. Feature Engineering Pipeline - Computes 50+ predictive features
5. Team Strength Models - ELO, Bayesian, Poisson ratings
6. Sequence Pattern Analysis - Markov chains, HMM
7. Machine Learning Ensemble - Logistic Regression, Random Forest, XGBoost, LightGBM
8. Monte Carlo Simulation - Match outcome probabilities
9. Odds Analysis - Value detection, Kelly criterion
10. Betting Decision Engine - Confidence thresholds, value requirements
11. Bankroll Management - Fractional Kelly, risk metrics
12. Profit Tracking - Bet history, ROI calculation
13. Season Lifecycle Manager - 38 matchdays, standings
14. Continuous Learning Engine - Model updates, weight adaptation

Project Structure

Bet261PredictionEngine/
├── start_all.bat           # Start all services
├── start_backend.bat       # Start API
├── start_frontend.bat      # Start Dashboard
├── start_orchestrator.bat   # Start autonomous engine
├── backend/
│   ├── app/
│   │   ├── api/           # FastAPI routes
│   │   ├── core/          # Config, database, orchestrator
│   │   ├── models/        # SQLAlchemy ORM models
│   │   └── services/      # Business logic modules
│   ├── init_db.py         # Initialize database
│   ├── requirements.txt
│   └── .env
├── frontend/
│   ├── src/
│   │   ├── pages/         # React components
│   │   └── services/      # API client
│   ├── package.json
│   └── vite.config.ts
├── data/                  # SQLite database
├── logs/                  # Application logs
└── models/                # Saved ML models

Configuration

Key settings in .env:

Setting	Default	Description
DATABASE_URL	sqlite:///./data/bet261_prediction.db	SQLite database path
INITIAL_BANKROLL	1000	Starting bankroll
MIN_ODDS	2.0	Minimum odds to bet
MIN_CONFIDENCE	0.65	Minimum prediction confidence
KELLY_FRACTION	0.25	Fractional Kelly multiplier
MAX_STAKE_PERCENT	0.05	Max bankroll per bet
RESULT_TRIGGER_SECONDS	58	Countdown trigger for scraping
MODEL_RETRAIN_INTERVAL	100	Matches between retraining

API Endpoints

Dashboard

• GET /api/dashboard/overview - Main dashboard data
• GET /api/dashboard/standings - League standings
• GET /api/dashboard/statistics - League statistics

Matches

• GET /api/matches/ - List matches
• GET /api/matches/upcoming - Upcoming matches with odds
• GET /api/matches/recent - Recent results

Predictions

• GET /api/predictions/ - List predictions
• GET /api/predictions/upcoming - Predictions for upcoming matches
• GET /api/predictions/accuracy/stats - Accuracy statistics

Betting

• GET /api/betting/ - Bet history
• GET /api/betting/pending - Pending bets
• GET /api/betting/summary - Betting summary
• GET /api/betting/bankroll/current - Current bankroll

Teams

• GET /api/teams/ - All teams
• GET /api/teams/elo-rankings - ELO rankings
• GET /api/teams/{id}/matches - Team match history

Spanish League Teams

The system tracks 20 teams:

• Real Madrid, Barcelona, Atlético Madrid, Athletic Bilbao
• Sevilla, Real Sociedad, Villarreal, Real Betis
• Valencia, Getafe, Rayo Vallecano, Osasuna
• Celta Vigo, Mallorca, Girona, Almería
• Cádiz, Las Palmas, Alavés, Espanyol

How It Works

1. Data Collection: Browser automation monitors Bet261 for upcoming matches and results
2. Feature Engineering: Each match gets 50+ computed features (form, H2H, strength, momentum)
3. Prediction: ML ensemble + Monte Carlo + ELO combine for final probabilities
4. Value Detection: Compare model probabilities to bookmaker odds
5. Betting Decision: Apply confidence/value thresholds, Kelly criterion
6. Continuous Learning: Update models after each match result

License

MIT