Nylon Calculus: Deep 3s, NBA range and college basketball
The college basketball season has been over for a while, and we’re really about to put it to bed — after the NBA Draft, almost no one will be thinking about or watching the 2017-18 season. We’ll fully be on to the possibilities of next year, with new players and rosters throughout the whole landscape. But there are still some things to learn by looking back, and how the Tournament was won this past year.
Plenty of ink has already been spilled on Villanova’s run to a second championship in three years. The Wildcats won each NCAA Tournament game by double digits, burying all comers under an avalanche of 3s — they took nearly 48.0 percent of their attempts from beyond the arc — and a tenacious switching defense. It was the highest 3-point attempt rate for a champion in tournament history — and second is Villanova’s championship team from just two years prior. And Michigan, their opponent in the title game, was also among the country’s leaders in 3-point attempt rate (43.2 percent of their shots came from 3).
It’s reflective of larger trends in college basketball, and the game itself. Attempts from beyond the arc have been slowly rising in college basketball since the line was introduced, and have spiked in the last four to five years. The average 3-point attempt rate last year was 37.5 percent, up from just 33.0 percent in 2013-14. Efficiency is also tracking upward, with the average 35.1 percent D-I 3-point success rate last year matching the highest of the last decade, per KenPom’s numbers.
This combination of historic efficiency and volume from 3 in college basketball — especially combined with the shorter 3-point line in college basketball — has led to discussions on moving the line back, perhaps to NBA range. This would theoretically help combat some of the recent trends. After all, it is (as a rule) harder to hit shots the further you move from the basket.
Available shot data for college basketball provides some credence for this theory.1
Based on the shot chart data available from the last two seasons, NCAA players are shooting about 36.3 percent on 3-pointers that are inside the NBA 3-point line. This slides to 34.6 percent on all 3s taken from beyond the NBA line.2
A more granular look at the data supports the potential efficacy of this fix. Take a look at this chart and table, tracking field goal percentage and eFG% in the NCAA as the location of the shot moves away from the basket.
In the NBA, the 3-point line follows an arc that’s 23.75 feet from the basket, sliding to 22 feet in the corners. Shots from this table certain to be from behind the NBA 3-point line start at a field goal percentage of 34.6 percent (between 24 and 25 feet from the basket), and consistently fall the further away you move.
Problem solved?
The potential pitfall here is that such a change has the potential benefit to teams that already push the spacing of their teams well beyond what’s required by the difference between a 2-pointer and a 3-pointer — a categorization that includes both national championship participants — while punishing those already using the shorter line more sparingly. This could especially be true in the short run, as everyone else adjusts.
As I noted while Villanova was running away from second-round opponent Alabama on the strength of a 3-point barrage, the Wildcats were also among the country’s leaders in floor spacing 3s, a pattern that continued through their national title. They led the country in pure number of attempts taken from beyond the NBA 3-point line as well as from beyond 25 feet by more than 100 attempts in each category. While the extra games were responsible for much of that gap, 76.0 percent of their total 3s came from beyond the NBA 3-point line, and an astounding 46.0 percent of their 3-point attempts came from 25 feet or more. Those ranked ninth and second, respectively, among teams in the country with an appreciable sample size in the available data.
Their opponent in the national championship game — the Michigan Wolverines — followed a similar pattern: John Beilein’s team took a ton of their 3s from well beyond the line. Michigan launched 74.7 percent (12th) of their 3s from beyond the NBA arc, and 38.0 percent (20th) of their 3s from beyond 25 feet.
Is it coincidence that a pair of top teams in truly long distance shooting met in the championship? Well, at least partly. The other leaders from last year included Marshall, North Carolina, Washington State, Marquette and Oregon. It’s obviously not a guaranteed harbinger of postseason success, but there are real on-court benefits to this type of spacing.
As Jesse Newell posited when he first took a look at this data in college basketball, the extra space provided by stepping a few feet behind the line can be very valuable to an offense. While the shots are generally less efficient taken on their own, it can help open up the rest of the offense, and force the defense to cover more space and make more difficult decisions.
One of the best ways we can mathematically measure spacing is the convex hull, which is essentially connecting a number of points (in basketball’s case, five) to create a shape with smallest area possible using those points. You can find a great explanation of this concept, and its use in NBA statistics, in this article on Nylon Calculus by the excellent @AcrossTheCourt, as he explored what the Houston Rockets could be gaining this past season by taking so many of their attempts from extremely deep.
In a simple demonstration provided therein, having one person step a few extra feet behind the 3-point line can create an extra ~200 square feet of space for opposing defenses to cover, and increase the distance between each offensive player by as much as five feet, even in relatively typical offensive scenarios. Jay Wright took that concept to its extreme this year — five of Villanova’s six most heavily used players took at least 60 shots from beyond 25 feet (4.25 feet, minimum, behind the college 3-point line).
That kind of spacing matters. There’s a reason Villanova’s been one of the most efficient teams of all time — for two years running — from inside the 3-point line. When everyone on the floor is a threat to shoot, especially from extreme range, there’s very little defenses can do to help inside. We saw the same phenomenon with the Rockets this past year. Their 3-point shooting and overall spacing enabled Chris Paul and James Harden the freedom to have two of the more prolific and efficient seasons in isolation ever. The top 20 teams, in terms of percentage of shots taken from NBA 3-point range, hit 63.2 percent of their shots at the rim last year. Everyone else averaged 60.5 percent. The general spread of 2-point percentage follows a similar pattern — the top hit 53.0 percent of their total twos, while the rest hit 49.5 percent of the same.
Obviously, there are other aspects to team offense — deep 3s and 2-point efficiency don’t tell the story. And I plan on digging further into the deeper team-wide effects as part of a future article on this topic. But let’s circle back to the unseen problems in moving the line back.
The teams already taking all those 3s that are ruining the game? They’re very much prepared for the move. Villanova hit 39.2 percent of its 3s from NBA range this year, and 38.0 percent of its attempts from 25 feet or more. 3s from inside the arc were little more than layups for them, as they hit 43.4 percent of those attempts.
The top 20 teams in total NBA attempts this year hit 36.5 percent of their shots from beyond the arc, while canning 35.0 percent of their 3s from beyond 25 feet. The rest of the sample sat at just 34.7 percent on NBA 3s, and 32.6 percent from outside 25 feet
There’s obviously a drop in efficiency, but the teams that have already prepared to launch from these deep distances are still showing a rather large efficiency gap relative to the rest of the league. That gap (36.5 percent vs 34.7 percent) is about reflective of that between the 100th and 200th best teams in the country from 3-point range, and is about the same as the difference in 3-point percentage from inside the arc.
A deeper arc could have a portion of its intended effect — it’s certain that the number of 3s taken, and efficiency from beyond the new arc, would fall league-wide. But it seems likely that dramatically increasing the number of NBA 3s taken in the NCAA tomorrow would inordinately benefit the teams already taking extreme advantage of deep 3s. Their coaches, returning players and offensive systems would have a built-in comfort level with the depth of the new arc and know best how to exploit the extra space afforded by forcing defenders to cover players in those spaces. Teams that already struggle with volume and efficiency from behind the arc would see both numbers fall, with what was once an easier 3 now simply a very long mid-ranger.
It could be healthy for the diversity of game itself, once teams and personnel had time to adjust. But if you’re trying to immediately hinder postseason runs like last year’s Villanova and Michigan squads, you may want to reconsider (at least until your team gets some practice from shooting deeper).
—
Of course, the real reason many people might be interested in NBA 3s taken in the college game is to see how their favorite prospects did. And, of course, I’ve got that data right here!
A few words of caution, though:
- We’re dealing with very small sample sizes here. There’s a reason college free throw percentage is considered more stable/predictive as an indicator for shooting in the NBA. College seasons are only 30 games or so, and in a good year you’ll get 150-200 3-point shots from a prospect. Splitting it further into NBA 3s and deep 3s only reduces that sample, so drawing strong conclusions from these numbers probably isn’t in your best interest.
- I don’t have a ton of information on the translatability of these numbers, as I only have collated data for the last two years. Todd Schneider, who’s built the excellent BallR tool for NBA shots, did some similar work for NCAA shooting recently. While he didn’t go incredibly in depth — essentially comparing pro 3-point percentages to NBA players with more than 100 attempts from NBA 3-point range while in college — his analysis did find that NBA 3PT field goal percentage in college doesn’t seem especially predictive.
- I do have some additional information for shot charts that stretches back to 2013-14, but I need to convert & analyze it. Would like to run my own analysis on this, especially with very deep 3s, to see if anything concrete can be obtained.
So after running wild with confirmation bias for your favorite/hated players, dump all the salt on these numbers.
Trae guesses as to who led the NCAA in deep & NBA 3s last seas — damn it. It’s an obvious conclusion for anyone who spent time watching college basketball this season, but Oklahoma’s Trae Young lead the NCAA in 3s from well beyond the arc — by a significant margin.
By my numbers, Young took 236 of his 3-point attempts from beyond the NBA arc (hitting 35.6 percent), and a whopping 184 attempts from beyond 25 feet (hitting 33.1 percent). And this despite missing four games from my sample.3 Trae Young the Team would rank in the top 80 for number of attempts from beyond 25 feet, which is just nuts. The next closest player (Marshall’s Jon Elmore) took just 134 attempts from 25 feet or more, despite having the full sample of games available.
Here’s the top-25 for prolific deep shooters in college basketball last year:
Only Villanova, Marquette and Marshall have multiple representatives on the list, which jives with the team leaderboard above. But while Trae was one of the most prolific deep shooters in the league last year, he certainly wasn’t one of the most efficient. That title (on a reasonably high volume of over 50 attempts) goes to Shep Garner, from Penn State. He hit 52 of his 103 attempts from beyond 25 feet, though was strangely at ‘just’ 42.0 percent overall from NBA 3. The best NBA 3-point shooter in the NCAA last year was Cassius Winston (Michigan State), who canned 52 of his 107 attempts from beyond the pro arc.
The best and most prolific deep shooter over the last two years combined is probably Andrew Rowsey (Marquette), who hit 144 of his 316 attempts (45.6 percent) from NBA range, and 80 of his 175 attempts (45.7 percent) from 25 feet or more.
I’ve got 160 players from this year and 130 from last year who took more than 100 attempts from beyond the NBA line, so hopefully we’ll be able to test whether this stuff means anything a bit better around draft time next year.
And, for one last table, all the NBA 3-point attempts for this year’s prospects (NCAA-only), referencing the Stepien’s Big Board. The list is sorted from most NBA 3s attempted last year to least.
Next: How will we know if the tanking reforms work?
Footnotes
- All the data referenced in this article is the same as the data used to build the free NCAAB shot chart app I’ve published on the Stepien. It’s based on data scraped from publicly available sources that provide shot locations. Approximately 40.0 percent of all games are available, skewing heavily toward Power 6 conference teams. In building this app, I superimposed an NBA 3-point line on the NCAA court. This was used to determine 3s that landed from inside and beyond the NBA 3-point arc.
- It should be noted the average 3-point % in this sample skews slightly higher than the typical D-I average, mainly because the best teams tend to have more games available.
- TCU (1/13), Kansas (1/23), Texas Tech (2/13), and Kansas State (2/24)