Time Is Running Out – Early Offensive Initiation Matters
By Seth Partnow
Jan 17, 2015; Houston, TX, USA; Golden State Warriors head coach Steve Kerr during the second half against the Houston Rockets at Toyota Center. The Warriors defeated the Rockets 131-106. Mandatory Credit: Troy Taormina-USA TODAY Sports
"Be quick but don’t hurry– John Wooden"
Time is an often overlooked commodity in basketball. Time to get off a shot before the defense closes out. Practice time for a team to jell. Playing time for a young prospect to learn the ins and outs of the professional game. Even at a more micro level, the NBA game is dominated by time, in particular, the shot clock. Twenty-four seconds seems like a long time. It’s almost half a minute. You’ve gotten this far in the article and you still have about 6 seconds left to get up a good shot! Those 24 seconds signifying the outer limits of the length of a possession intrinsically factors into any reasonable analysis of the game. A shot which is considered wasteful or selfish with 21 seconds left on the 24 is vital with only one tick remaining. Maximizing those seconds is crucial to an offense, while denying their productive use is one of the hallmarks of good defense.
Of course, teams almost never have 24 to actually run an offense. The majority of NBA possessions start after a shooting attempt by the opposition — a made or missed field goal[2. For ’tis a Make or Miss League.] or a trip to the free throw line. After completing the defensive possession, the (now) offensive team must advance the ball and initiate the play. This takes time. In fact, it takes a lot of time. According to SportVU data[4. Thru games of March 1, 2015], the average time taken by an NBA team to initiate it’s offense[3. Using a quick and dirty definition of initiation = time when the ball first crosses the arc on the attacking half of the floor] following one of these scoring attempts is around 9.2 seconds.[3. This analysis purposefully ignores plays starting from a dead ball, like a timeout, or coming after a steal.] In general, this is probably an overestimate, as any functional offense includes some off-ball preparatory screening and cutting at the start of the play. But even if that figure is exaggerated by 50%, over a quarter of the shot clock might as well not exist.
While the average is 9.2 seconds from gaining possession to first crossing the arc, there is decent variation among teams with the Warriors being the fastest (8.4 seconds on average) with Toronto bringing up the rear (10.4 seconds). Without similar data from last year, it’s impossible to say if that Golden State number is faster than their 2013-14 effort, but the Dubs do lead the league in pace, averaging around 2.4 more possessions per 48 minute game than a year ago[4. Per Basketball-Reference.com], and have jumped from 8th to 1st in terms of field goals attempted in the first 8 seconds of the shot clock[5. Per SportVU shot log data.] This increased tempo appears to have some explanatory power for the Warriors offensive improvement from 12th to 2nd.
For starters, there is an unsurprising, almost definitional correlation between getting offense initiated earlier and getting more shots early in the shot clock:
While there appears to be very little relationship between speed of offense initiation and a team’s efficiency on those early attempts, every team in the NBA is more efficient early in the shot clock than at any other time, even removing high value offensive rebound putbacks from the equation[7. See the last example here for controlling for these on the defensive side of the ball.], by an average of 8.2% in eFG%[8. Using eFG rather than true shooting for sake of simplicity — shot clock times of shooting fouls are imperfectly captured in play-by-play data and a huge pain to pull from public SportVU data.]. In other words, much like getting more assisted shots would be good for any offense, generating more early looks will help any squad to a degree.[9. With the caveat that early shots tend to be taken because they are open and/or being allocated to better players, and calling for “more early shots” somewhat assumes no significant reduction in the quality of those shots.]
There also appear to be some more downstream effects, in that speed of offense initiation and the efficiency of shots taken in the middle 8 seconds of the shot clock are somewhat correlated:
The correlation is not particularly strong, and could easily represent statistical noise. On the other hand, it’s also plausible that getting into offensive continuity earlier forces the defense to defend more potential shots per possessions, each one providing another chance for a mistake to be made. Under this interpretation, each extra second of offense represents something of a lottery ticket that can be cashed in for a high quality shot in the right circumstances, and as many a gambling commercial will tell you, you have to be in it to win it.
As a final note, one of the earliest pieces of coaching I can remember receiving was to not spend time sulking after an opposing made shot. Any time the team would pause, our coach would yell “Don’t stand around. Get it [the ball] out, get in in and LET’S GO!” The value of this advice is somewhat apparent, through March 1, the NBA as a whole was averaging 11.3 seconds to initiate the offense following an opposition made FG, as compared to only 7.3 seconds from a miss. In fact, Golden State is the only team in the league averaging under 10 seconds, so perhaps Steve Kerr got the same bit of coaching and has passed it along as another way for the Warriors to push the tempo and always put the opposing defense under pressure.