From the Crow's Nest - NBA Finals (06/10/16)

By Sean Pyritz @srpyritz
Last week, I took an in-depth look at a simple way the Cavaliers might stabilize their offense after suffering defeat in game one. While Coach Lue may say he wants to push the pace in order to re-engineer the ball movement that stumped the Eastern Conference, I suspect he wishes to stick a fork in the LeBron isolation post up game as well. Given LeBron's timid jumper and insane turnover rate, combined with the disappearing shot clock when trying to shake free of Iguodala, it is hard to fault Lue. But enough about last week, today I want to dive into some interesting numbers as we head into game four. What follows involves looking at rarely explored situations through a lens not much used in modern NBA analytics media.

About a years ago, I heard Spurs head coach Gregg Popovich talk about how fouling and sending people to the free throw line hurts your offense because you are forced to play against set defense. This makes sense in theory, but I was interested in seeing if there were numbers to back up the intuition. I've been charting games, possession by possession, for a little over a year now, with the primary intention of tracking the sets and actions ran by each team – I am a weirdo, no need to worry. While I had been making special note of AFTs – After Free Throw possession – it recently occurred to me that teams do not face set defenses only after free throws, but after any number of what I call “dead-ball” situations. I use the term loosely to include anything that isn't what I call “live-ball” situations – after a missed field goal attempt or steal. Made shots, timeouts, sideline out of bounds would all constitute predecessors to dead-ball situations as I have defined it. In this way, we can approximate that a team is likely facing set defense after dead-ball scenarios and unset defense after live-balls.

In order to assess how the Warriors and Cavs have done in these situations so far in the Finals, we will call upon some help from one of the founding fathers of modern basketball analytics, Dean Oliver. One of the main ideas of his seminal work Basketball on Paper is that possession based statistics should be used to measure efficiency because they account for differences in pace between teams and minutes played per game for players. The concept of offensive rating – points a team scores per 100 possessions – stems from his ideas. Along with offensive and defensive rating, when evaluating a team's performance, floor percentage is an interesting and seemingly forgotten stat. Floor percentage is the number of scoring possessions – any time a team comes down the floor and scores at least one point – divided by the total possessions. This provides a sense for how a team scores, with a high floor percentage reflective of lots of free throws and/or high percentage shots at the rim, and a low percentage perhaps reflective of three pointers or perhaps lots of turnovers. Where offensive rating points to efficiency, floor percentage describes the process of reaching said efficiency.

As a compliment these concepts, I came up with a statistic that bridges the gap between offensive rating and floor percentage, which I creatively named “Points per Scoring Possession.” The only change from offensive rating is you take the number of points and divide by scoring possessions instead of total possessions. It is simply a measure of how many points a team can be expected to score if they do in fact score on a given possession. Thanks to Basketball-Reference and Oliver's techniques at estimating statistics from simple box score numbers, below I have charted the Points per Scoring Possession (PPSP) against the floor percentage for each team this past regular season. The offensive rating of each team is captured both in the color of the shapes and the shapes themselves. As you can see, there is a clear division between the above average and below average offenses. You can envision almost a 45 degree line cutting the graph in half from top right to bottom left where the further you are to the right, the more efficient your offense.

If you look at the leaders in three pointers this season, it follows logically that they perfectly correspond with the top teams in terms of points per scoring possession shown above. What is unusual however, is the extraordinarily high floor percentage from the Warriors – the league leader in three pointers. Golden State is having its cake and eating it too from long distance. Cleveland has attempted to model itself after the Warriors to an extent, nearly matching them in floor percentage, but remain the wannabe version of one of the greatest offenses in the history of the league.

Now that we've got a lens to view possessions through, we can bring it back to dead-ball and live-ball situations in the Finals. The table below gives an assortment of statistics for each game and the overall totals in the competitive portions of the Finals after dead-ball and live-ball situations. Whether looking at the overall or each game individually, you can see a clear difference between dead-ball and live-ball situations. Counterintuitively, Cleveland is scoring less efficiently through three games after live-balls than after dead-balls. They are also scoring less frequently in live-ball scenarios. Golden State's transition defense has to be given a lot of credit for this discrepancy. Watching this series, Golden State is clearly making an effort to balance the floor and get back in transition out of respect to the Cavaliers sharpshooting.

Trying to decipher the dramatic turnaround in game three, the Warriors lack of toughness is certainly captured in Cleveland's astounding 63% floor percentage in dead-ball scenarios – an absolutely disastrous effort from Golden State defensively on the level of some of Cleveland's weaker Eastern Conference foes. Interestingly, the Warriors actually scored the most points per scoring possession to this point in the series in game three. While Cavaliers fans definitely feel confident heading into game four, it is a bit disconcerting that through three games, the Cavs seem unable to deter the high-potency of the Warriors offense. A return to form in terms of floor percentage by the Warriors spells trouble for the Cavaliers.

Taking a look at the series as a whole, it is noteworthy that the floor percentage of each team has settled right near their season averages and while the Warriors are still bringing a familiar punch with each score, the Cavs aren't displaying their same power. The dampening of the PPSP by Cleveland is clear evidence of the impact the Warriors defense is having on scoring proportion of the Cavalier's offense. If Cleveland is going to win this series, they will have to reverse the scoring mix on the Warriors like they did in game three. I invite you to watch tonight's games through the lenses of these topics, we will visit these ideas again I assure you.