Blog
Deep dives into NBA analytics, machine learning and the science of basketball prediction
Regular Season Recap: What the Model Got Right, Wrong, and Why
760 predictions. One full regular season. Here is an honest breakdown of XGBoost_TOP17_v1.4_Platt — tier by tier, team by team, and what it means going into the playoffs.
Regular Season Recap: What the Model Got Right, Wrong, and Why
760 predictions. One full regular season. Here is an honest breakdown of XGBoost_TOP17_v1.4_Platt — tier by tier, team by team, and what it means going into the playoffs.
3 Months In: The Model Converged — Here's What the Data Actually Shows
After 695 live predictions across a full NBA season stretch, our live accuracy has converged to backtest levels. The calibration picture has also changed — and not in the direction you'd expect.
Understanding Basketball's Four Factors: A Deep Dive on the Metrics That Actually Win Games
Dean Oliver's Four Factors framework explains why teams win with remarkable precision. Here's how each metric works, what the weights actually look like, and how our XGBoost_Top17 model uses them.
I Analyzed 5,000 NBA Games - Here's What Actually Predicts Wins
What I learned building a prediction model that contradicts everything you think you know about basketball
16 Days Live: Why Our Honest 64% Beats Claimed 80%
After 116 predictions over 16 days, here's our brutally honest performance breakdown and why calibration matters more than raw accuracy.