When the Warriors and Trail Blazers tip off Game 1 of their first round playoff series on Sunday, we’ll see two of the most potent offensive teams the NBA has to offer in action. Golden State had the most efficient offense in the league over the course of the regular season, averaging 113.2 points per 100 possessions, while Portland ranked 11th in that category (107.8). However, the difference between the two offenses was extremely marginal from mid-February on, as the Warriors and Trail Blazers posted the fourth and fifth best offensive ratings in the NBA after the All-Star break and were separated by just 0.3 points per 100 possessions over that span. All four regular season matchups between the two teams (which Golden State swept 4-0) took place before the end of January, however, and a considerable portion of the Warriors’ success in those games was due to the kind of shots they took.
Using artificial intelligence and SportVu player-tracking data from thousands of NBA games, MOCAP Analytics has developed a model that estimates the probability of making any attempted shot or free throw. The model takes into account specific information about the shooter and defenders, along with many features (e.g., shot location, distances to defenders, actions prior to the shot). Using the probabilities of the shots taken in a game, MOCAP can measure the quality of ‘shot taking’ by counting how much each team or player ‘should have scored’ on these shots (i.e., expected score). In addition, by comparing actual scores with expected scores, MOCAP estimates the ‘shot making’ performance (or luck, depending on your point of view) in the scoring output for any player or team.
Through the graphics above, we can see that in their four regular season matchups, not only did the Warriors attempt higher quality shots (1.18 expected points per shooting play) than the Trail Blazers (1.06), they also outperformed as shot makers, scoring 0.08 more points than expected per shooting play. Great shot selection and good shot making had plenty to do with the Warriors outscoring the Trail Blazers by an average score of 125-106.
Here we can see examples of some great shot taking by Kevin Durant, who averaged an impressive 1.25 expected points per shooting play in the season series:
And here are some examples of great shot making by Stephen Curry, who scored 0.17 more points than expected on his shooting plays in the season series:
As you can see, it doesn’t hurt to have multiple world-class shooters on the roster, but by taking higher quality shots, and making more shots than expected, the Warriors put themselves in great position to win. Next let’s take a look at the activity leading up to some of these shooting plays.
Here we see the shot tracks of Kevin Durant and C.J. McCollum from the final – and closest – regular season matchup between the two teams, a 113-111 Warriors victory on January 29th. These charts display player movement on shooting plays leading up to their attempts. The white streaks show the off-ball action, while the orange streaks show on-ball action. The green circles at the end of the streaks indicate made shots, while the red crosses signal missed shots.
As shown in the shot tracks graphics, both players are versatile and can play on or off the ball. Durant attacked the paint and dribbled into several shots quickly after crossing halfcourt. According to NBA.com/stats, Durant shot 64.2 percent during the season on shots resulting from drives, the best efficiency in the league of any player who attempted at least 2.3 drives per game this season. McCollum’s shot tracks reveal several shooting plays where he sprints off screens, a signature of Portland’s offense. It’s no surprise McCollum ran 2.59 miles per game this season (second most in the NBA according to NBA.com/stats).
Of course, all of this data is based of the regular season series between the Warriors and Trail Blazers, and who knows what’s in store in the first round series ahead. After all, it will have been more than two-and-a-half months since the last time they faced each other when Game 1 tips off on Sunday, and both sides have certainly made adjustments since then. Still, by gaining insight into some of the reasons behind the Warriors’ success in the regular season series, we now have a better idea of how Golden State may plan to attack Portland, and vice versa. The game will always be played on the court, but the numbers help us better understand why what we’re seeing is happening.
Keep an eye out for more Dubs Deep Dives with the help of MOCAP Analytics throughout the postseason.