← NBA predictor Today’s predictions →

NBA team dashboard · Free · Updated daily

How every team is really playing.

Offense and defense per 100 possessions, standings, splits and the model’s pre-game calls for all 30 teams. Click any team for its full profile.

Conference

Loading…

Offense vs defense

Points scored (→) and allowed (↑) per 100 possessions. Up and to the right is better; lines mark the league average.

Standings & ratings

Sort by any column. Click a team to open its profile.

# Team

How the numbers work. Possessions are estimated from the box score (0.96 × (FGA + TOV + 0.44 FTA − OREB)). ORTG and DRTG are points scored and allowed per 100 of a team’s possessions over the season; Net is the difference; Pace is possessions per 48 minutes. The model’s pre-game call uses only games played before tip-off, from each team’s fifth game on. Last season, the best net rating belonged to the Oklahoma City Thunder (+11.4).

Udemy course · by Aykut Onat

Want to build a predictor like this?

Learn it step by step, from raw NBA data to a deployed app. The course builds the first version of this model, the foundation this page grew from.

  • Collect and clean real NBA game data
  • Engineer features and train XGBoost
  • Deploy your own app on Hugging Face

45 lectures · 2h 54m · Beginner-friendly Python

Enroll on Udemy
See the curriculum
  1. Introduction & installing the tools
  2. Data collection & exploration
  3. Feature engineering
  4. Modeling
  5. Testing the model
  6. Visualization with Tableau
  7. User interface & publishing