April 7, 2009
Effective Pace
Correcting for Confounding Factors

by Kevin Pelton

When you think of slow-paced NBA teams, odds are you think of veteran defense-minded squads like the San Antonio Spurs or the Detroit Pistons of recent vintage. From this perspective, the Portland Trail Blazers are an enigma. As we've discussed at BP, Portland is the league's second-youngest team in terms of playing age. The Blazers are also far more effective on the offensive end of the floor, ranking second in the league in Offensive Rating and 15th in Defensive Rating. Nonetheless, Portland has played at the NBA's slowest pace in terms of possessions, 86.5 per game.

There is no question the Blazers play a deliberate style, but are they the league's slowest team? Earlier this year, TrueHoop's Henry Abbott--a confirmed Blazermaniac--raised the possibility that the numbers might be misleading.

"[W]hen either team gets a lot of offensive rebounds it makes both teams appear to be slower," wrote Abbott. "Similarly, teams that commit and force very few turnovers will appear to be faster, although neither of those things really tells you anything about the speed at which they play. Those trends persist. I suspect that when you look at listings of team pace, you will see some pace inflation from teams that frequently turn the ball over, or don't get or give up a lot of possession-extending offensive rebounds."

In particular, offensive rebounds are a factor for Portland. The Blazers lead the league by grabbing 32.7 percent of their own available misses, meaning lengthy possessions that slow the game down. How can we correct for this issue? We could count what the APBRmetrics community calls "plays," where an offensive rebound starts a new play though it continues a single possession. However, this would cause the opposite problem, since many second chances turn into extremely quick scores. I've suggested averaging the two measures, but this is a messy solution that does nothing to account for the issue of turnovers.

Instead, I decided to take a different look at the concept of pace--elapsed time. Roland Beech of 82games.com was able to pull for me data showing how often the first action of a team's possession resulted in a shot attempt, a turnover or a foul, and the average time for each action.

As it turns out, the difference in time for turnovers and field-goal attempts isn't enormous. At the league level, the average (first) field-goal attempt comes 13.8 seconds after the start of a possession, the average turnover after 12.2 seconds. Fouls, intriguingly, tend to come much earlier--8.6 seconds after the start of a possession. Naturally, because of quick putbacks, subsequent actions are all much faster, topping out at a 6.5-second average for multiple shot attempts.

To correct for the different distribution of each team's possessions, I used the league percentage of field-goal attempts, turnovers and fouls, multiplying this by how long on average the given team took on each type of play. Effectively, this neutralizes the stylistic differences between teams in terms of offensive rebounding, turnovers and getting to the free-throw line, and serves as a weighted average of their elapsed time on each type of play. The resulting "effective pace" tells some interesting stories. Let's start with the fastest and slowest teams overall:

FASTEST TEAMS

Team            EffPace

Golden State      11.5
L.A. Lakers       11.5
Cleveland         11.7
Indiana           11.7
Utah              11.7

SLOWEST TEAMS

Team            EffPace

Toronto           12.3
San Antonio       12.2
Atlanta           12.2
Orlando           12.2
Portland          12.2

For the most part, these rankings are similar to traditional pace rankings. There are a couple of surprises, most notably Cleveland. A slowdown team in terms of possessions, the Cavaliers shoot forward by this measure. In part, this might have something to do with the way the data is tracked, as Cleveland comes out as an up-tempo squad even before adjusting the possession distribution. At the other end, Toronto drops all the way to the bottom of the league in terms of pace, while the Blazers are out of the cellar.

The more revealing information might be provided by the ability to split pace into its offensive and defensive components. We tend to think of the offense as dictating the speed of the game much more than the defense does, and indeed the fact that the standard deviation of offensive paces is higher than that for defensive paces (0.30 vs. 0.19) suggests that offense "controls" this aspect of the game somewhat more. Still, teams do vary in how quickly they force their opponents to play. Let's take a look at the rankings strictly at that end of the floor:

FASTEST DEFENSES

Team            EffPace

Golden State      11.5
Cleveland         11.5
Utah              11.7
Houston           11.7
L.A. Clippers     11.7

SLOWEST DEFENSES

Team            EffPace

Milwaukee         12.2
Portland          12.2
Atlanta           12.1
New Orleans       12.1
San Antonio       12.1

This sheds some additional light on why the Blazers seem to play much faster than the numbers show. As it turns out, Portland's offense isn't all that sluggish after accounting for offensive rebounds. Seven teams have a slower effective pace on offense than the Blazers do, but only Milwaukee spends more time at the defensive end of the floor. This might be subtle enough to escape observation--it's not something I've noticed in watching Portland all season.

Conventional wisdom would probably hold that a fast defense is a bad defense, because it means opponents are scoring quickly. One of the criticisms of the Seven Seconds or Less Suns was their unwillingness to expend defensive energy so they could get the ball back more quickly. While that might be true of the Don Nelson Warriors, it doesn't seem to apply to either this year's Phoenix team (11.9 seconds, right at average) or Mike D'Antoni's new Knicks squad (12.0 seconds, on the slow side).

Additionally, the presence of two of the league's four top defenses amongst the fastest-paced squads seems to cast doubt on the importance of forcing opponents deep into the shot clock.

The difficulty in tracking the data makes it unlikely that some form of effective pace will ever replace standard possessions per game as how we measure teams' pace of play. We can still learn from the lessons of effective pace. Most important amongst those is this: Keep in mind that how many possessions a team averages per game isn't always an exact depiction of how the team plays. There are tendencies that can render a team's calculated pace somewhat misleading. Don't be fooled.

Kevin Pelton is an author of Basketball Prospectus. You can contact Kevin by clicking here or click here to see Kevin's other articles.