When you look at average possession length (APL) data, the thing that stands out is that slow defenses tend to be very good. Not slow in terms of the player’s speed, but slow in terms of forcing opponents to take a long time with the basketball.

Syracuse, Florida, and Arizona currently have the three slowest defenses in the country. Clemson, San Diego State, and Virginia rank fourth, sixth, and seventh, respectively, and all three have an adjusted defensive efficiency that ranks in the top ten. (The converse, that fast offenses are good, is not true nearly as often.) We know what makes defenses good, but what makes them slow?

I looked at all the stats I track and found the ones that best predicted defensive APL. Here those stats in order of significance.

1) Defensive free throw rate. A lower defensive free throw rate tends to yield longer possessions. One thing to note here is that it takes some interpretation to determine cause and effect. It is obvious that fouls will end possessions prematurely, but I would guess the bigger issue in this case is that fouls often result from the offense attacking the basket and teams that force long possessions are able to prevent their opponent from getting easy scores.

2) Defensive two-point percentage. A lower defensive two-point percentage tends to yield longer possessions. The lack of being able to get easy shots forces the offense to take more time with the ball. Block percentage are correlates well to defensive APL, but not quite as well as overall two-point percentage.

3) Defensive three-point attempt percentage. More opponent three-point attempts tend to yield longer possessions. If you’re not getting easy twos, you’ll tend to shoot more threes. Teams that aggressively defend the three point line can still be great defensively, but they won’t force long possessions.

4) Offensive steal rate. Lower offensive steal rate tends to yield longer possessions. This one’s interesting, but it makes sense. If your offense gives up more steals, the opposing team will get more transition opportunities and thus shorter possessions. This is the only offensive characteristic that correlates well to defensive APL

5) Defensive turnover percentage. Higher defensive turnover percentage tends to yield longer possessions. This is where things get interesting because the relationship is probably the opposite of what you expected. I’m just going to be a simpleton and propose that longer possessions give a team more opportunities to make a mistake.

6) Defensive rebound percentage. Higher defensive rebound percentage tends to yield longer possessions. This one is also counterintuitive. In an individual possession, an offensive rebound makes for longer possessions. Over the long haul, though, longer shots make for worse offensive rebounding opportunities, and I’d guess that’s what we are seeing here.

7) Defensive three-point percentage. Lower defensive three-point percentage tends to yield longer possessions. While I generally avoid talking about defensive three-point percentage as if it’s a real thing, there does seem to be one way for the D to exert its influence on opponents three-point accuracy and that’s by forcing longer possessions. I’d guess this is more an effect of longer possessions than a cause. Teams are taking more bad threes late in the shot clock as a result of not being able to get an easier shot earlier.

If you can do all the things listed here, you’ll probably have a good defense. I think mainly you want to be at the first two, however. If you are good at preventing points in the paint, then teams will take bad mid-range twos and threes later in the shot clock. Your three-point defense will suddenly look like you have control over the perimeter. Life will be good.

These characteristics aren’t foolproof though. Teams get oodles of easy 2’s and free throws against Denver, but for whatever reason, it takes them a while to do it. Texas defends the two well but team shoot quickly against them. There are still some mysteries about what makes defensive possessions long, but at the heart of it is usually an inability to get easy twos.