Layup for you, 3 for me

February 4, 2016

By Stephen Shea

Modern NBA teams love the 3-point shot. (Well, most of them do.) This season, teams are averaging an all-time-high 23.7 3-point attempts per game.

Even in this 3-point-happy era, almost everyone would agree that offenses would rather get a good attempt at the hoop than a typical 3-point attempt.   Something like this would be nice.

Analytics have shown that NBA teams win through efficient basketball. As teams alternate possessions, each team should strive to get more efficient shots than they are giving up to their opponent. As a general rule, good offenses generate attempts at the hoop and good defenses prevent them.

But what if a team could be so effective at generating and knocking down quality 3-point attempts, and so stifling defensively, that they would happily trade their own 3-point attempt for an opponent’s shot at the rim?

In other words, what if a team could generate more points per 3-point shot than they gave up on opponents’ shots in the restricted area?

Even though opponent restricted area shots include wide-open dunks in transition, this is not an outrageous suggestion. In fact, it happens more often than many might realize.

Between the 1997-98 and 2014-15 season (18 seasons), it has happened 35 times, or about 6.6% of the time. (The NBA moved the 3-point line back to its original position in 97-98 after a 3-year trial at a shorter distance.)

Certain franchises have done it more often than others. The Spurs accomplished the feat 9 times in those 18 years (or 50% of the time). While in the same timeframe, Sacramento, Atlanta and Minnesota never came close.

Ok, maybe we’ve cheated a bit in using the past data to suggest that today’s teams should be capable of generating more points per 3-point shot than they give up on opponents’ restricted area attempts. Some of the teams that accomplished this feat in the past also took very few 3-pointers. For example, the 1999-00 Jazz did it, but they only took 10 3-pointers a game while giving up more than twice as many attempts at the rim. It’s easier to generate a higher 3P% if you only take the wide-open looks.

Remember that teams are averaging almost 24 3PA per game this season. In the previous 18 seasons, there have been 107 instances where a team has averaged at least 20 3PA per game. Among those 107 instances, 9 (8.4%) generated more points per 3-point shot than they gave up in the restricted area.

It is certainly possible for today’s teams to be more efficient on their own 3-point attempts than their opponents are from the restricted area.

We will see that it’s also very important.

Define the metric

Layup for you and 3 for me. When is that tradeoff acceptable to, or even desired by, a team? This occurs when the team is typically more efficient on 3-point attempts than their opponents are on restricted area FGA. To capture this phenomenon, first consider a team’s points per 3-point attempt minus their opponent’s points per restricted area FGA.

For example, the 2016 Lakers are scoring 0.945 points per 3-point attempt. They are giving up 1.266 points per attempt at the rim. The difference is -0.321 points per tradeoff.   Since no one likes a decimal, let’s multiply this difference by 100. We see that for every 100 times the Lakers exchange a 3PA for an opponent shot at the hoop, the Lakers are outscored by approximately 32 points. We’ll call this metric “3forME” and let negatives denote “outscored by.” The 2016 Lakers have a 3forME of -32.

It’s not that simple

So that this new metric does not get misinterpreted, let’s make several observations.

1. A high 3-point FG% is not just a result of putting better shooters on the floor. Teams must also be able to generate high-quality opportunities for those shooters. One very effective way to do this is through a drive and kick with a dangerous penetrator. When LeBron drives to the hoop, the defense collapses leaving players like Kevin Love open on the perimeter. One way to improve 3-point shooting is to be more effective driving to the hoop.

2. Rim protection has become a buzzword in basketball analytics. When we hear it, we think Rudy Gobert. Having Rudy Gobert roaming the paint will help a team’s interior defense, but big men in the middle are not the only contributors to good interior defense. In today’s game, having defensive positional versatility to switch screens helps keep dangerous penetrators like James Harden out of the lane. Teams need to defend the rim, but the first step towards that effort is keeping drivers on the perimeter.

3. 3forME does not account for free throws generated through offensive activity. Players are more likely to be fouled when driving to the hoop than launching from 3. For most players, free throws are an efficient shot. ….most players…

4. 3forME includes transition. Transition (which can stem from a defensive play) is a good way to generate open 3-point attempts. Transition defense is part of keeping an opponents’ restricted area FG% down. Breakaway dunks are an efficient shot…usually…

We will see that 3forME is incredibly predictive of team success, but this does not imply that a lineup of 4 Steve Kerrs and 1 Dikembe Mutombo is ideal. It’s more complicated than that.

Layup for you, 3 and title for me

It is hard to generate shots at the hoop. Defenses are designed to prevent players from driving and to double dangerous post players. By comparison, a team that spaces the floor with 3 or 4 effective 3-point shooters can get a reasonable 3-point attempt on almost any possession.

If a team can generate a positive 3forME, if they can be more efficient shooting 3s than their opponent is after working hard to generate a shot at the hoop, then that team will win games. If they can generate a positive 3forME that is better than their competitors, then they will win championships.

I hate to look too far into the past when trying to understand the present NBA. The game is evolving rapidly. The best strategies of 10 years ago are not necessarily optimal today.

Let’s focus on the last 4 seasons (2015-16 included). In that time frame, the average season 3forMe was -13.4. There were 7 seasons of a positive 3forME. Here is the top 5 (in reverse order).

5. 2013-14 San Antonio Spurs (3forME=2.83)

San Antonio had the highest 3forME that season. As a team, they shot a league-leading 39.7% from 3. Danny Green, Patrick Mills and Marco Belinelli each took at least 293 3s in the regular season and shot at least 41.5% on them.

Behind Tim Duncan and Kawhi Leonard, the Spurs led the Western Conference with a Defensive Rating of 102.4.

The group obliterated the Miami Heat in 5 games on their way to another NBA Championship for coach Gregg Popovich.

4. 2012-13 Miami Heat (3forME=3.33)

The Heat had the best 3forME in 2012-13, and in true 3forME fashion, won the NBA title.

The Heat led the Eastern Conference shooting 39.6% on threes. It helps when your top 5 players in 3-point attempts also all shoot at least 40.6%.

The team’s defense was solid. They had a DRtg of 103.7, which contributed to a Net Rating of 8.6. That Net Rating was second only to OKC, which was the only other 2013 team with a positive 3forME. The Thunder’s title dreams were crushed when Russell Westbrook went down to an injury in the opening playoff series.

3. 2014-15 Golden State Warriors (3forME=3.64)

The Warriors had the best 2015 3forME. They were the only positive 3forME team that season. (Atlanta was second at -3.51.) Following the trend of the previous 2 seasons, the league-leading 3forME team also won the NBA title.

2. 2015-16 San Antonio Spurs (3forME=5.13)

To have a 3forME of 5.13 in today’s NBA is absolutely remarkable. The Spurs aren’t taking many 3s, which makes generating a high 3forMe a bit easier, but even still, their score is amazing. It’s higher than any 3forME posted in the prior five seasons. I’d go on, but…

1. 2015-16 Golden State Warriors (3forME=13.09)

3forME has accurately predicted the NBA Champion in each of the previous 3 seasons, and the modern game has never seen a 3forME like this. There is only one team in the last 19 seasons that has posted a 3forME above 6.6. That was the 2001 Spurs, which had a defense anchored by Duncan and Robinson. They also only took about 13 3s a game. Golden state is averaging over 30 3-point attempts and shooting a ridiculous 43% on them. If it weren’t for the exceptional play of the 2016 Spurs, Golden State would be a HEAVY favorite to win the title.

It’s hard to describe how incredible this 3forME score for Golden state truly is. Remarkably, this performance by the Warriors leads to an under-appreciation for what the Spurs are accomplishing.

Does it help to know that Toronto is third in 2016 with a 3forME of -5.24 or that 21 of the 30 teams this season have a 3forME below -10?

Conclusions

It is difficult to generate a positive 3forME in today’s NBA, but it is not impossible. Through sound roster and lineup construction and through advanced offensive and defensive game plans, teams are generating very impressive 3forME scores.

It can’t be emphasized enough that 3forME scores reflect more than the talent on the roster. They are a product of the system as well. Golden State and San Antonio have talent, but there are other teams with talent that are not nearly as successful.

Exceptionally poor 3forME reflects more than inadequate talent. The Timberwolves have had a 3forME between -26.74 and -36.24 in each of the last 4 seasons. That suggests a systematic failure. Heck, the 76ers at least have a 3forME of -20.

Considering the current trends in the NBA, I predict that in the future, a positive 3forME will be more than a great predictor of a championship. It will be a prerequisite.

Appendix: Complete 3forME data for last 4 seasons