Daily Fantasy Baseball Simulation
Current• Objective: Develop an efficient, accurate Daily Fantasy MLB Simulator and Optimizer.Key Achievements:• 67% winning rate in 'cash' contests (top 43-44% lineups).• 2.4% placement in top 1% of 'tournament' contests with top-heavy payouts.Core Components:1. Data Handling & Transformation:• ETL for MLB Statcast Data.• Daily scraping and API requests for statistics, lineups, weather, fantasy prices, umpires.2. Advanced Feature Engineering:3. Machine Learning Models:• Pitcher/Batter Skill Modeling: Baseline skills in neutral conditions.• Batter vs. Pitcher Outcomes: Factors like handedness, batted ball profiles, pitch types etc.• Pitching Change Probability: Based on game state and pitcher stats.• Pitcher Selection Modeling: Likelihood of selection based on game state, and pitcher stats.• Umpire Model: Strikes above/below expected used to adjust outcome rates.• Weather Impact Model: Effects weather using current and historical weather data.• Pitch Type Classifier: Unbiased classification from Statcast data.• Expected Pitch Value: Expected outcome quantification from Statcast data.• Park Factors: Influence of different parks on outcomes.4. Simulation & Optimization:• Monte Carlo simulations• Lineup optimizers within constraints.• Cash Game: Maximize average points.• Tournament Strategy: Balance player correlations, ownership for ROI.