The On-Ball/Off-Ball Spectrum (HV Weekly: 4/24/2020)
Virginia on one end, Saint Mary's on the other.
Welcome back to the Hoop Vision Weekly.
Today, a special guest on a new episode of the Solving Basketball podcast, along with a corresponding deep dive into on-ball and off-ball play types, how offenses respond and shape to certain defensive coverages, some “pure man-to-man” defenses, and more.
And if you missed it, last week’s edition — “The Truth About Transfers” — was widely read and shared, and is now officially the most-viewed edition of the Hoop Vision Weekly since we launched this thing in December 2018. Thank you for your continued support.
Back in December, we kicked off our “All-Decade” series by looking at the statistical changes in style of play since 2010. Among other things, the data indicated that:
Over at The Athletic, Seth Partnow has recently launched a similar — and even deeper — look at the changes in NBA style of play. The three-part series (with more on the way) is an informative investigation of modern basketball and league-wide shifts.
Part 1: What data tells us about perceived changes to the game
Part 2: Variety is the spice of life
Naturally, we were eager to get Seth on a new episode of Solving Basketball. We discussed his findings, the value of play type data, and the similarities and differences between the NBA game and the college game.
Listen below, and read on for a more detailed breakdown around some of this work.
Listen: Apple or Spotify
The on-ball/off-ball defensive spectrum
In Seth’s work, he categorized the eight different play types into two groups: On-ball and off-ball.
The four on-ball play types all involve the ball handler generating the final outcome (shot, turnover, foul) of the play. Those four types: Isolations, post-ups, dribble handoffs, and ball screens where the handler calls his own number.
The four off-ball play types all involve a secondary player making the final outcome (shot turnover, foul) of the play. Those four types: Spot-ups, cuts, off-ball screens, and ball screens where the roller receives the ball.
In the NBA, the ratio between on-ball and off-ball play type usage — according to Seth’s work — consistently sits around 50/50 on a year-to-year basis:
The ratio of “on-ball” to “off-ball” scoring plays has remained remarkably consistent over this time period, with between 47 and 50 percent of half-court scoring plays coming from on-ball actions — isolations, post ups, or the ballhandler in a pick-and-roll or handoff situation.
In the NCAA, the ratio of scoring plays leans more heavily towards off-ball.
NCAA Average: 40% on-ball ——— 60% off-ball
During the podcast, our conversation mainly focused on the offensive side of the ball. However, the final topic discussed hit on the impact the defense has on opponent play types — specifically Virginia’s pack line.
There are many potential reasons why off-ball play types are more prominent in college. The most obvious place to start is that players are simply less skilled and dynamic with the ball.
But the priorities of college defensive schemes also play a clear role here. In the 2019-20 season, there were 17 teams to use at least 1,000 plays of zone defense. Those 17 defenses forced their opponent into the following on-ball to off-ball ratio.
Zone Defenses: 23% on-ball ——— 77% off-ball
Every defense — as we covered in Starting Five back in January — is some combination of man-to-man and zone principles.
Virginia’s pack line is a man-to-man that heavily prioritizes limiting on-ball opportunities. They aggressively hedge ball screens, double the team the post, and give early help one-pass-away. It’s a style of defense which seemingly goes against the numbers.
On average, off-ball play types (like spot-up shots) tend to be more efficient than on-ball play types. In theory, a defense should directly benefit from inducing isolations and post-ups. Yet the Tony Bennett style aggressively takes them away.
So how does the pack line defensive consistently remain one of the most efficient in college basketball?
Offenses get off-ball shot opportunities against Virginia, but they simply don’t make them. Bennett’s defense has ranked above the national average in opponent three-point percentage for nine seasons in a row — and in the top 30 for four seasons in a row.
The other end of the spectrum: Saint Mary’s
In the past, we have shown that Michigan’s defense is one of the “purest” man-to-man styles in the country. And their on-ball/off-ball ratio this season (even without Luke Yaklich on staff) backs that up.
Michigan Defense: 53% on-ball ——— 47% off-ball
But another man-to-man defense is actually more extreme than Michigan’s. Randy Bennett’s defense ranked number one in the country at inducing on-ball play type volume.
Saint Mary's Defense: 58% on-ball ——— 42% off-ball
The defensive philosophy for Saint Mary’s is most apparent in ball screen defense. Their coverages can differ, but the Gaels try to be as passive as possible — essentially guarding the ball screen two-on-two.
(Hoop Vision readers might remember: We coincidentally tried a very similar ball screen philosophy at New Mexico State when playing against Saint Mary’s.)
 Under the screen — sometimes even on shooters
The simplest way to induce the ball handler into taking a shot during a ball screen is to just go under. Usually, it’s a coverage reserved for poor shooters. Going under a strong shooter isn’t exactly a recipe for success.
However, Saint Mary’s has done just that. In 2019 WCC championship game, the Gaels went under on Zach Norvell — Gonzaga’s best shooter off the bounce — all game long in the upset win.
This season in the WCC semifinals, Saint Mary’s tried the same thing against the best three-point shooting team in the country in BYU.
Going under shooters like Zach Norvell or Jake Toolson sure seems suboptimal, but it’s important to remember that Gonzaga and BYU were two of the best offenses in the country — the defense has to pick its poison one way or another.
The Saint Mary’s logic — at least in my estimation — is to live with some off the bounce threes rather than needing to provide help which ultimately results in catch-and-shoot threes. And for whatever it’s worth, the Gaels had success in both games.
 Drop coverage — guarding two-on-two
In the WCC quarterfinals, Saint Mary’s used drop coverage against Pepperdine. In the clip below, the player guarding the screener (Jock Perry) drops into the paint while the player guarding the handler (Tommy Kuhse) fights to get back even.
Notice how the other three defenders not involved in the screen are simply staying at home on shooters. In an ideal world, Kuhse and Perry are the only two Gaels’ players with ball screen responsibilities.
In the play above, Pepperdine’s Colbey Ross attempts to keep Kuhse on his hip — but this is the type of result that Saint Mary’s will live with.
In part due to the ball screen coverage, Ross finished the double OT game with 43 points, 29 shots, and 11 free throw attempts.
 Guarding two-on-two gone wrong
Not providing any additional help on ball screens is of course easier said than done. The better the ball handler is off the bounce and the better the screener is on the roll or pop, the harder it is for the defense to play two-on-two.
The following two clips show Gonzaga guards getting downhill quickly against Saint Mary’s passive ball screen coverage.
Not only does the speed of Gonzaga’s guards — Admon Gilder and Ryan Woolridge — matter here, but so does the type of ball screen. Both are examples of “throw-and-chase” screens — with Petrusev starting out on the perimeter instead of the block.
It’s harder to execute coverage against the throw-and-chase variety. And after the miscoverage, the Saint Mary’s defense isn’t designed to give any extra help.
Hoop Vision Basketball Analytics Online Courses
In last week’s newsletter, I put out an announcement to college coaches for analytics consulting services. I’ve worked with a variety of different teams and coaching staffs to increase their ability to implement data into different aspects of a college program.
In the process of putting that out, we also received responses from people outside of just college basketball staffs — and without university athletic budgets. In light of that interest, we have begun exploring the idea of Hoop Vision online courses.
These courses — like all things Hoop Vision — would focus on the intersection between the analytics side of basketball and the coaching side of basketball.
If you are a fan looking to understand the game deeper, an aspiring coach looking to break into the industry, or a current coach looking to expand your value to your current program — I think Hoop Vision is uniquely positioned to help.
For those interested, we have put together a brief survey at the link below. Filling it out — which (obviously) doesn’t represent a commitment of any kind on your end — will enable us to gauge the overall interest level and desired course curriculum.
Link: Hoop Vision Basketball Analytics Course Form
We’ll have more information on the HV online courses in the near future. And in the meantime, you can reach out with any questions to firstname.lastname@example.org.
This newsletter is entirely supported by its readers. Thank you for your support of Hoop Vision and allowing us to continue to grow.
Subscribe to Hoop Vision Plus ($10/month or $100/year) for instant access to all 40 of our team scouting breakdowns. Plus access to our full archive of HV+ exclusive articles.