The New Offensive Meta

January 1, 0001 - 8 minutes

library(dplyr)
## Warning: package 'dplyr' was built under R version 3.6.1
## 
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

The path to where we are now

With the rise of analytics, teams have been continuously searching for tricks to optimize their offense. The Golden State Warriors demonstrated the effectiveness of the run-and-gun offense dominated by three-pointers. In recent years, the mid-range shot has fallen out of fashion in favor of “higher return” shots. While sometimes these analytics driven approaches lead to powerful offense, they can also be misguided if not properly used. Well-meaning people like Dean Oliver, and his four factors, have led people astray with wild guesses at what is important. According to Oliver an offense is defined by shooting (40%), turnovers (25%), rebounding (20%) and free throws (15%). In a previous blog, I have shown that statistically the importance is more like shooting (84%), rebounding(8%) and turnovers (7%). In my example, I consider free throws to be a part of shooting. Looking at some of the teams this season who have overachieved on offense, evidence points to the supremacy of shooting but consensus has not completely caught up to the reality. Additionally, the idea that a team can simply maximize shooting sounds much more unrealistic than saying a team can maximize offensive rebounding or ball security. However, there is clear evidence that playstyle and shot selection can greatly increase shooting (measured in points per shot).

Translating playstyle to improved offense

The key to a succesful offense is discovering styles of play that maximize the important parts of the offense instead of ones that minimize “unimportant” parts. To make this idea clearer we can look at two teams who had poor offensive rebounding; BYU (335th) who maximized another aspect of their offense at the expense of offensive rebounding and Wyoming (348th) who instead minimized offensive rebounding with no trade-off.

Recently, the importance of the offensive rebound has come into question. Wyoming, perhaps inspired by this news, employed an offense that pathologically avoided offensive rebounds. A typical Wyoming possession was in no ways unique up until the point of the shot where all of their players would run away from the basket, essentially avoiding any chance of an offensive rebound. You would assume this would translate to a strong defense by avoiding having to defend in transition but their defense was pedestration (164th in the nation). They also had shooting which ranked 242nd in the nation meaning they essentially minimized their offense rebounding ability and did not maximizing anything. For reference to how a typical Wyoming possession looks, see the video below (Notice the times where people were in possession for rebounds but instead sprinted up the court).

Succesful usage of playstyle to maximize output

Now that we have Wyoming as an example of what not to do, we can look at the success stories such as BYU. In this section we will look at some systems that succesfully used playstyle to maximize their shooting, and subsequently their offensive success in the hopes of figuring out the formula for this success. I will first focus on the surprises of the season, non-major, non-traditional powerhouse teams who finished in the top 50 in offense in the 2019-2020 season (BYU, Richmond, Utah State, Davidson, Akron, San Francisco, North Texas).

Generalizing playstyles

The goal of this article is to provide a general overview of what worked offensively last season. Future articles will be dedicated to individual teams to try to dissect their success. The table below shows each of the overachieving teams organized by their overall offense rankings.

off = readRDS("Off_Rankings_2019_2020.rds") %>% arrange(Offense.Rank) %>% dplyr::select(Team, `PPS Rank`, `ORB Rank`, `TOV Rank`, `Offense Rank` = Offense.Rank) %>% filter(Team %in% paste0(c("BYU", "Richmond", "Utah State", "Davidson", "Akron", "San Francisco", "North Texas"), "_2019"))  %>% mutate(Team = gsub("_.*", "", Team))

knitr::kable(head(off,10), "html", booktabs = T) %>%
  kableExtra::kable_styling(position = "center")
Team PPS Rank ORB Rank TOV Rank Offense Rank
BYU 2 335 9 7
Davidson 19 257 61 25
Akron 16 192 110 32
San Francisco 56 43 137 37
Richmond 25 334 17 43
Utah State 65 69 109 44
North Texas 18 158 178 46

Here, PPS (Points per shot) is a stand-in for shooting. This table makes it abundantly clear that there are two types of offense at play here. The vast majoity are poor offensive rebounding teams with excellent shooting. San Francisco and Utah State are more well-balanced teams who excelled in all categories to varying degrees. Turnovers are much more variable among the teams.

Why strong shooting teams are also poor rebounders

Perhaps the most interesting teams on this list are those with high points per shot and poor offensive rebounding. The fact that they all are poor at offensive rebounding does not point to a personnel deficiency but more to the fact that playstyles that promote high percentage shooting are not conducive to offensive rebounding. More importantly, it demonstrates that playstyle can lead to maximization of shooting despite personnel (This is not diminish the ability of the players themselves). In fact, this type of stat-line was common on many of the fringe top-50 low-major teams for example South Dakota who managed to finish 34th in points per shot and 68th in overall offense despite coming from the traditionally weak Summit League. South Dakota simultaneously finished 336th in the nation in offensive rebounding.

Wyoming (and common sense) has shown that poor offensive rebounding does not somehow cause strong shooting. Watching game tape tells a more nuanced story. Of the five teams with high shooting and poor offensive rebounding, four play a screen-heavy constant motion offense. These offenses are defined by a lack of a true center and constantly cleared lanes. North Texas in particular exemplifies this style. In the video below can see their use of screens and constant movement to create easy layups and open threes. North Texas players only occupy the lane for short periods of time to set up an attack on the rim and quickly clear out.

These cleared lanes are where the lack of offensive rebounds come from. Players are not actively avoiding offensive rebounds, like Wyoming, but they are simply not in position. Shooting is prioritized and the best shooting comes when spacing is the priority and the guards have the option to drive. Clogged lanes allow for higher offensive rebounding but prevent driving and limit offensive options (See Virginia), particularly against zone defenses where it is difficult to pass the ball to the inside man. Emprical evidence shows this style of play is optimal and that it is possible for a team without elite recruits to outperform elite-level recruiting teams using deliberate playstyles that maximize high percentage shots.

Davidson is the outlier in this group. They play a more Saint Mary’s style of offense. This offense is similar to the aforementioned motion offenses but they employ a more traditional center. However, instead of clogging lanes, the center uses positioning to clear out the lanes and open up the opportunity to drive to the hoop. Generally, the center in this system is constantly moving around or set up in a manner that forces defenders to make a decision on who to guard during a drive. Just like the previously mentioned offenses, the movement of the center sacrifices their positioning for rebounds. The play bookmarked below demonstrates a strong example of this. You can see how the defender clears the lane and opens up options for the guards.

Other paths to offensive success

Utah State and San Francisco differ from the other overachieving teams in both the distribution of their statistics and their playstyles. Both teams had a more balanced style finishing just outside the top 50 in shooting with strong offensive rebounding (San Francisco: 43rd, Utah State: 69th). They also limited turnovers (San Francisco: 137th, Utah State: 109th) but not to the extent of the aforementioned teams (Excluding North Texas). They show that there are multiple ways to success. Despite the similar statlines, San Francisco and Utah State are not stylistically similar.

San Francisco takes the typical WCC-style screen-heavy, all eyes on the ball handler, approach. A good way to spot a strong team is to watch each player individually and determine how often they are making plays that improve their teams chance at scoring, whether it be getting open, setting screens or clearing out defenders. On a player-to-player level the Don’s have a cohesion that most teams lack. Additionally, they are very disciplined on the offensive boards. Once a shot goes up they are actively getting into position, allowing for their strong offensive rebounding. In the case of Utah State, their success comes from utilizing size. I would not hazard to guess that they are the tallest team in the nation, boasting three 7-footers with their smallest player being 6’3. In total, there are only 2 players on the entire Utah State roster below 6’5. Accordingly, Utah State prefers to play heavily in the paint. They demonstrate how deliberate recruiting and playing to your strengths can produce strong results.