Besides scoring, the major contribution to a basketball offense is playmaking. Or, more specifically, the ability to draw extra defenders away from their assignments. As a measure of how well a player does this, assists leave something to be desired; they are only tallied when a pass is made to a player who scores, regardless of how helpful the pass was. Which is why we need a way to detect who creates open shots for teammates by drawing extra defensive pressure. For that, I use something I call “Opportunities Created.”
The Method
One way to do this is to track any time a second defender leaves his man in order to help defend an offensive player. This can be voluntary defensive strategy or it can be the result of the first defender being beaten off the dribble.
Such events are actually fairly easy to keep track of. Here’s a quick example from a game:
(Yes, I believe all NBA replays should be watched in Italian.) Note the first play of the video, when the ball is fed into Tim Duncan in the post. At about 0:21, help comes to double-team Duncan. This leaves Michael Finley open for a 3. The unguarded shot — regardless of the result — was created by Duncan drawing defensive help away from Finley. That is an opportunity created (OC) by Tim Duncan.
At 13:30 of the video, Dirk Nowitzki drives on Tim Duncan and forces Ginobili to help. The ball is swung around until Jerry Stackhouse ends up with a jumper (despite a nice close-out effort by Finley). That is also an OC by Nowitzki, who originally broke down the Spurs defense, despite it not registering as an assist.
Every OC has to end in some kind of attempt, even if it is a fouled attempt. If there is never a shot (or foul), there can never be an OC. (What exactly would the player have created, then?) OC’s can be registered when a player draws defensive pressure and the following occurs:
- An open attempt
- An open “hockey” attempt (extra passing, as in the Nowitzki play above)
- A foul at the rim on a layup attempt (created by the scrambling of the defense)
- An offensive rebound putback (created because the rebounder’s defender was forced to help)
The Results
As we would expect, guards create much more than bigs. They have the ball in their hands a lot, drive and dish a lot, and are often the defensive focus of pick and roll action. (It has become popular to “jump,” or trap the pick and roll to prevent the dribbling guard from penetrating or taking an open jump shot.) In last year’s playoffs, here is the positional breakdown of OC’s:
The breakdown by position is similar to data tracked this year as well.
So how well do assists correlate to OC’s? Overall, for the 133 players who logged at least 150 possessions in last year’s playoffs, the average error was 0.86 assists (with a standard deviation of 1.81 assists). Of those 133 qualifying players, the following had the largest discrepancy between assists and OC’s per 100 possesions (number of OC’s per 100 in parentheses):
- Rajon Rondo 6.8 (5.6 OC’s per 100)
- Ronny Price 5.8 (1.2 OC’s)
- Jason Kidd 4.7 (4.7 OC’s)
- Luke Walton 4.0 (4.0 OC’s)
- Jamrio Moon 3.9 (0.0 OC’s)
and the following players have the largest discrepancy between OC’s and assists (meaning they create more than assists would suggest):
- Brandon Roy 5.3 (8.7 OC’s)
- Nowitzki 4.0 (8.3 OC’s)
- Ginobili 3.3 (12.4 OC’s)
- Westbrook 2.8 (11.7 OC’s)
- Reddick 2.4 (6.3 OC’s)
I will post the complete leaders from last year’s playoffs in a follow up post.
The Discussion
There is a fairly strong correlation with assists (R=0.83). However, the error rate in certain players is enormous, which was the impetus for the stat in the first place; we want to know who’s creating opportunities, not simply who is passing to good players.
It might be worthwhile to simply track double-teams, but there isn’t always an attempt of some kind after a double team. Sometimes, the ball is re-entered into the post and a second or third double team can come on the same possession. If the player turns the ball over, nothing positive came from the double team.
Future consideration should be given to the defensive version of this metric: “OC’s Against.” Those occur every time a defender has a teammate help him in his assignment. If a player never needs defensive help, it means that he is never responsible for the offensive “power plays” (4 on 3 on the rest of the court) that occur because of an OC.
Another useful follow up exercise would be to look for “ball-stoppers.” Most NBA players and teams are coached well enough to take advantage of the power play provided when two defenders commit to one offensive player. Every once in a while, a player will allow the defense to recover and lose that advantage by holding the ball or not shooting when he should shoot, allowing a scrambling defense to recover. It’s fairly easy to spot, and seems worth tracking in the future.
great stuff elgee.
Looking forward to the follow up with the stat leaders.
[…] 22, 2011 by ElGee In a prior post, I introduced the concept of Opportunities Created (OC) as a measure of […]
[…] a prior post, I introduced the concept of Opportunities Created (OC) as a measure of […]
[…] passing and creation isn’t factored into any a player’s “usage.” Using Opportunities Created, a measure of offensive creation, it’s possible to estimate a player’s contribution, or “offensive load:” […]
[…] down to 30% from 3-point land this season. He rarely earns trips to the free throw line and rarely creates opportunities for […]
[…] fouls drawn. This allows a fairly accurate estimate of FD using FTA. However, as is the case with Opportunities Created and assists, it is the outliers who are often the most interesting. Someone like Dwight Howard shoots far less […]
[…] errors essentially create a power play — similar to the power play we saw in opportunities created — that spikes the opponent’s offensive efficiency by nearly 50%. In the simplest […]
[…] And the correlations are even stronger when including Help Needed, the defensive counterpart to Opportunities Created. Without further ado, here are the marginal values used for EV. First, the defensive […]
[…] Opportunities Created (OC) […]
[…] league — is indicative of just how much he does for Chicago on that end. He’s 2nd in Opportunities Created and 14th in assists per game, so it’s not just a shooting festival. Let’s give Rose a […]