Late Clock Offense is Hard Work

Apr 19, 2014; Toronto, Ontario, CAN; Toronto Raptors guard DeMar DeRozan (10) shoots with the 24-second shot clock not functioning during the game against the Brooklyn Nets in game one during the first round of the 2014 NBA Playoffs at Air Canada Centre. The Nets beat the Raptors 94-87. Mandatory Credit: Tom Szczerbowski-USA TODAY Sports
Apr 19, 2014; Toronto, Ontario, CAN; Toronto Raptors guard DeMar DeRozan (10) shoots with the 24-second shot clock not functioning during the game against the Brooklyn Nets in game one during the first round of the 2014 NBA Playoffs at Air Canada Centre. The Nets beat the Raptors 94-87. Mandatory Credit: Tom Szczerbowski-USA TODAY Sports /
facebooktwitterreddit
Apr 19, 2014; Toronto, Ontario, CAN; Toronto Raptors guard DeMar DeRozan (10) shoots with the 24-second shot clock not functioning during the game against the Brooklyn Nets in game one during the first round of the 2014 NBA Playoffs at Air Canada Centre. The Nets beat the Raptors 94-87. Mandatory Credit: Tom Szczerbowski-USA TODAY Sports
Apr 19, 2014; Toronto, Ontario, CAN; Toronto Raptors guard DeMar DeRozan (10) shoots with the 24-second shot clock not functioning during the game against the Brooklyn Nets in game one during the first round of the 2014 NBA Playoffs at Air Canada Centre. The Nets beat the Raptors 94-87. Mandatory Credit: Tom Szczerbowski-USA TODAY Sports /

I’ve recently spent a lot of time looking at those situations where good defense leads to early, highly efficient offense. But what about the other end of the shot clock? It’s been established that the later in the 24 a shot is taken, the more the average accuracy drops. The question is, why is late clock offense so much harder?

My initial supposition was due to time pressure, teams and players were taking a tougher mix of shots. As the clock expires, any shot is better than no shot at all[1. Some prayers get answered. Offensive rebounds are good things as well.]. And there is something to this. Using a a quick and dirty measure of shot quality[2. Expected Effective Field Goal Percentage (XeFG%) which accounts for shot distance and defensive pressure as measured by proximity.], there is a decline as the shot clock progresses. However, the decline in observed efficiency is far steeper[2. The dip in early shots is largely the result of “tip in” attempts, which tend to be surprisingly low accuracy considering they are shots coming from near enough to the rim to be, you know, tip ins.]:

xefobsefgtime1
xefobsefgtime1 /

As the “expected” efficiency is based on league average for all shots, there has to be some correction on the back end of the shot clock to balance out the edge on early shots. However, the size of that correction and the steepness of the drop off late in the clock is remarkable. Some of this decline is due to other factors. A slightly higher proportion of shots are “self-created” later in the clock than earlier. However, the drop affects teammate created and self-created[2. Using 2.5 seconds of possession prior to shooting as a proxy for self-created shots, as discussed here.]  opportunities roughly equally:

So it’s not just offensive players taking matters into their own hands more. Something else is going on here which isn’t being totally captured by existing[1. At least in the public domain.] data. Or, more likely, a combination of other factors.

  1. Closest defender distance is a highly imperfect measure of defensive pressure. SportVU does not capture hand position[2. Hand down, man down, I thought I heard someone say once.], and public data doesn’t reflect the speed of a closeout. The latter is important because as a defender, you know you can really sell out to run the shooter off a spot up shot if there isn’t time for the offensive player to do anything productive should you fly by him.
  2. On the flipside, things that might cause a player to not take a shot under normal conditions, like being off-balance, on the move or bobbling the catch, no longer dissuade when the shot simply must go up.
  3. There might simply be a psychological component of rushing the shot knowing time is short.

These factors aren’t mutually exclusive, and probably build on each other[4. One other possibility is the identify of players taking these shots, but that’s a broader topic worthy of its own post.]. But the data is clear, for reasons not totally determined, scoring against shot-clock pressure is more difficult. The “pace” in pace-and-space isn’t just about shooting early, but also about not being forced to shoot late.