In the inaugural episode of the Thinking Basketball podcast, I introduced a concept called “Scoring Value,” or ScoreVal for short. ScoreVal is an attempt to measure the value of a player’s scoring contributions in a single number, which helps to resolve the classic “volume versus efficiency” debate. But there’s a bit more to it than that.
ScoreVal takes the “scoring component” of a Box Plus-Minus model and breaks it out to derive a single, plus-minus style number for scoring impact. Unlike the original podcast, this version is adjusted to league averages to create a more even playing field for older players. The inputs are:
- relative scoring rate (points per possession)
- relative true shooting percentage (rTS)
- scoring turnovers (relative to league average)
- team rORTG (relative offensive rating)
- teammate outside shooting
Turnovers are an often overlooked component of scoring. In this case, we’re only interested in “scoring turnovers,” which are based on the percentage of a player’s offensive load that comes from his scoring attempts. The remainder of a player’s turnovers are consider “playmaking turnovers,” so if he passes a lot, it’s assumed he commits more turnovers while passing. If he shoots a lot, it’s assumed more of his turnovers come from scoring. This, of course, is not a perfect estimate, but it’s generally a reasonable one without manually judging whether a turnover came on a scoring attempt.1
For example, James Harden committed 60 percent of his turnovers last year on “bad passes,” per the play-by-play (according to basketball-reference). The model attributes 45 percent of his turnovers to playmaking. In 2006, Kobe Bryant committed 26 percent of his turnovers on “bad passes.” The model attributes 27 percent of his turnovers to “playmaking.” It’s usually in the ballpark, which is good enough in this case.
True Scoring percentage
If we take these “scoring turnovers” and fold them into shooting efficiency, we can produce a slightly clearer picture of why the model upgrades or downgrades certain seasons when compared to true shooting percentage. This is a simple adjustment: (1) Add scoring turnovers to true shot attempts as “missed attempts,” (2) redo the calculation (points divided by attempts divided by 2) and the result is a “true scoring percentage,” if you will. When applied to some of the big scoring seasons ever, we see the following relative changes from true shooting to “true scoring:”
- ’19 Harden -0.6%
- ’77 Abdul-Jabbar +0.2%
- ’16 Curry +0.2%
- ’14 Durant +0.4%
- ’13 James +0.9%
- ’84 Dantley +0.9%
- ’91 Miller +1.7%
- ’62 Chamberlain +2.0%
- ’06 Bryant +2.6%
- ’87 Jordan +2.6%
Harden is really the only mega scorer who actually loses ground when including turnovers. Remember, “scoring turnovers” are estimated based on the ratio of scoring to playmaking volume, so Harden isn’t necessarily being punished for passing more.2
ScoreVal’s value
As always, keep in mind that this one-number reduction isn’t perfect. In reality, each player’s “real” scoring value will vary slightly based on circumstance. Heck, even the idea of siphoning off scoring is a bit fuzzy since playmaking is such an integral counter-weight based on a defense’s reaction. This isn’t baseball, after all.
However, the model does add intelligence to what’s in the box score. So if you want to get an idea of “how good” a scoring season is based on just the available box score data, this is an upgrade over mentally curving volume and efficiency.3 And by calculating everything relative to league norms, there’s little (or no) need to adjust for era.
There are still other factors to consider though. “Fit” comes to mind first — for example, the presence of an elite shot-creating teammate can boost a finisher’s numbers. This happened in Phoenix when Steve Nash delivered Michelin-star room service to Amare Stoudemire for years. (Although, I still credit Amare for his scoring prowess and being an incredible finisher.) In that sense, role is always something to consider too, although most players aren’t saddled in desperate floor-raising situations for their entire career. By the way, the argument that certain players take a bunch of bail-out shots to end possessions doesn’t hold much water either, especially when considering a great scorer should be able to manufacture quality attempts earlier in the possession.4
Efficiency is really important
Perhaps the most noteworthy takeaway of this model is that efficiency is really important. Many fans have long valued monster volume scoring as the gold standard as long as the efficiency was passable, but ScoreVal objects to that idea.
There have been 577 seasons with a scoring rate above 20 points per 75 on below-average true shooting. Among those seasons, the lowest efficiency to produce a positive ScoreVal comes in at around 4 percent below league average (-4% rTS). The worst efficiency for a season with a ScoreVal of at least +0.5 is around -2% rTS.5 Elgin Baylor posted a +0.7 in 1965 on true shooting 1.6 percent below league average (-1.6% rTS). And the lowest efficiency season to top +1 in ScoreVal is -1% rTS. In other words, it’s hard to provide value as a volume scorer without being efficient!
This holds true for the highest volume scorers too. There have been 32 seasons with a scoring rate above 30 points per 75. Of the 12 least efficient seasons — where their true shooting was within 3 points of league average (+3% rTS or worse) — not one cracked the top-100 ScoreVal seasons on record. Some of these seasons are idolized with religious fervor, yet they aren’t as likely to provide value for a winning team as more efficient counterparts with slightly lower volume.
Though we have to be careful here. Many of these lower-efficiency players were valuable overall offensive weapons by just about any measurement, but that reintroduces the idea of playmaking. Yes, volume scoring at average efficiencies can pave the road for setting up teammates, but when we look at the quality of the scoring output alone, it falls short of more efficient seasons in terms of estimated value. A few extra points don’t offset the 20 or 25 times a game the more efficient player attacks.
Being a great scorer is really about generating (and converting) high-efficiency attempts over and over. Sure, some shots are easy for one player and hard for another — Shaq and Steph Curry would both be dreadful if they traded shot profiles — but the ability to manufacture attempts that lead to “efficient” points on-demand is a key to winning in basketball. It’s also why some regular season scoring numbers can be dented in the playoffs, but that’s for another day…
The best scoring seasons
Below are the 50 best scoring seasons since the shot clock, per ScoreVal.6 Click here for the entire list.7
| Player | Season | Tm | MP | Pts/75 | relative TS% | relative Scoring TOV% | relative True Scoring% | ScoreVal |
|---|---|---|---|---|---|---|---|---|
| Stephen Curry | 2015-16 | GSW | 2700 | 31.9 | 12.8% | -1.4% | 13.0% | 3.3 |
| Kareem Abdul-Jabbar | 1970-71 | MIL | 3288 | 24.9 | 10.6% | -1.8% | 11.4% | 3.3 |
| Kareem Abdul-Jabbar | 1971-72 | MIL | 3583 | 25.7 | 9.9% | -2.0% | 10.9% | 3.2 |
| Amar'e Stoudemire | 2007-08 | PHO | 2677 | 27.7 | 11.6% | -1.1% | 12.2% | 3.0 |
| Wilt Chamberlain | 1961-62 | PHW | 3882 | 28.0 | 5.7% | -2.3% | 7.7% | 2.8 |
| Michael Jordan | 1990-91 | CHI | 3034 | 32.0 | 7.1% | -3.7% | 9.6% | 2.8 |
| Adrian Dantley | 1983-84 | UTA | 2984 | 27.8 | 10.9% | -1.6% | 11.8% | 2.7 |
| Kevin Durant | 2012-13 | OKC | 3119 | 28.2 | 11.2% | -0.4% | 10.9% | 2.7 |
| Bob McAdoo | 1974-75 | BUF | 3539 | 27.4 | 6.7% | -1.7% | 7.9% | 2.7 |
| Charles Barkley | 1987-88 | PHI | 3170 | 26.8 | 12.7% | 0.6% | 11.8% | 2.7 |
| Kevin Durant | 2013-14 | OKC | 3122 | 31.4 | 9.4% | -1.4% | 9.8% | 2.7 |
| Peja Stojakovic | 2003-04 | SAC | 3264 | 23.3 | 10.8% | -2.9% | 12.3% | 2.6 |
| Kevin McHale | 1986-87 | BOS | 3060 | 23.9 | 11.7% | -1.4% | 12.5% | 2.6 |
| Amar'e Stoudemire | 2004-05 | PHO | 2889 | 27.0 | 8.8% | -1.4% | 9.7% | 2.6 |
| LeBron James | 2012-13 | MIA | 2877 | 28.1 | 10.5% | -1.9% | 11.4% | 2.6 |
| Kiki Vandeweghe | 1983-84 | DEN | 2734 | 27.4 | 7.5% | -4.5% | 10.5% | 2.6 |
| Shaquille O'Neal | 1998-99 | LAL | 1705 | 29.7 | 7.3% | -2.1% | 8.2% | 2.6 |
| Bob McAdoo | 1973-74 | BUF | 3185 | 23.4 | 9.1% | -1.3% | 9.8% | 2.6 |
| Kiki Vandeweghe | 1982-83 | DEN | 2909 | 24.1 | 8.2% | -3.9% | 10.9% | 2.6 |
| Karl Malone | 1989-90 | UTA | 3122 | 30.4 | 8.9% | 0.1% | 9.2% | 2.5 |
| Kareem Abdul-Jabbar | 1976-77 | LAL | 3016 | 24.5 | 9.7% | -1.0% | 9.9% | 2.5 |
| Shaquille O'Neal | 1993-94 | ORL | 3224 | 27.9 | 7.7% | -1.7% | 8.5% | 2.5 |
| LeBron James | 2013-14 | MIA | 2902 | 28.4 | 10.8% | -0.7% | 10.7% | 2.5 |
| Shaquille O'Neal | 1997-98 | LAL | 2175 | 30.1 | 6.3% | -1.7% | 7.6% | 2.5 |
| Wilt Chamberlain | 1962-63 | SFW | 3806 | 27.4 | 5.8% | -1.9% | 7.4% | 2.5 |
| Kareem Abdul-Jabbar | 1972-73 | MIL | 3254 | 23.9 | 8.2% | -1.8% | 9.3% | 2.5 |
| Michael Jordan | 1987-88 | CHI | 3311 | 32.7 | 6.5% | -3.1% | 8.7% | 2.5 |
| Kevin Durant | 2015-16 | OKC | 2578 | 29.3 | 9.3% | -0.3% | 9.0% | 2.5 |
| Adrian Dantley | 1979-80 | UTA | 2674 | 26.2 | 10.4% | -0.7% | 10.6% | 2.5 |
| Kevin McHale | 1987-88 | BOS | 2390 | 22.3 | 11.8% | -1.5% | 12.5% | 2.5 |
| Michael Jordan | 1989-90 | CHI | 3197 | 32.0 | 6.9% | -3.3% | 9.5% | 2.5 |
| Shaquille O'Neal | 2001-02 | LAL | 2422 | 29.4 | 7.0% | -1.9% | 8.3% | 2.4 |
| Stephen Curry | 2017-18 | GSW | 1631 | 29.9 | 11.9% | -1.0% | 11.8% | 2.4 |
| Neil Johnston | 1956-57 | PHW | 2531 | 20.5 | 9.5% | -1.5% | 9.9% | 2.4 |
| Shaquille O'Neal | 2002-03 | LAL | 2535 | 28.3 | 8.3% | -1.5% | 9.3% | 2.4 |
| Dan Issel | 1981-82 | DEN | 2472 | 24.6 | 7.0% | -2.8% | 8.7% | 2.4 |
| Charles Barkley | 1988-89 | PHI | 3088 | 24.1 | 11.6% | -0.6% | 11.6% | 2.4 |
| Charles Barkley | 1989-90 | PHI | 3085 | 24.1 | 12.4% | -0.1% | 12.4% | 2.4 |
| Kevin Durant | 2016-17 | GSW | 2070 | 27.1 | 9.8% | -2.0% | 11.0% | 2.4 |
| Kareem Abdul-Jabbar | 1982-83 | LAL | 2554 | 23.4 | 8.8% | -1.6% | 9.9% | 2.4 |
| Neil Johnston | 1955-56 | PHW | 2594 | 19.7 | 9.7% | -1.4% | 10.4% | 2.4 |
| Larry Bird | 1987-88 | BOS | 2965 | 28.2 | 7.0% | -3.5% | 9.3% | 2.4 |
| Charles Barkley | 1990-91 | PHI | 2498 | 27.8 | 10.1% | -1.1% | 10.6% | 2.4 |
| Adrian Dantley | 1985-86 | UTA | 2744 | 28.8 | 8.8% | -1.9% | 10.1% | 2.3 |
| Kenny Sears | 1958-59 | NYK | 2498 | 18.3 | 13.3% | 0.0% | 12.6% | 2.3 |
| Dirk Nowitzki | 2005-06 | DAL | 3089 | 28.6 | 5.4% | -3.7% | 8.0% | 2.3 |
| Shaquille O'Neal | 1994-95 | ORL | 2923 | 30.0 | 4.5% | -2.6% | 6.6% | 2.3 |
| Michael Jordan | 1988-89 | CHI | 3255 | 30.0 | 7.7% | -2.5% | 9.4% | 2.3 |
| Kareem Abdul-Jabbar | 1980-81 | LAL | 2976 | 24.7 | 8.2% | -1.5% | 9.2% | 2.3 |
| Jerry West | 1964-65 | LAL | 3066 | 23.0 | 9.3% | -1.9% | 10.2% | 2.2 |
Among that group, Curry’s 2016 is the most efficient when incorporating turnovers. It’s the only season in NBA history to win the scoring and efficiency titles, a mind-bending feat that should be talked about like the moon landing.8 Heck, there have only been 20 seasons in NBA history of any volume to top Curry’s “true scoring” efficiency that year (with a minimum of 1,000 minutes played). Reggie Miller did it in 1991 while tallying 22.7 points per 75, which was 7 points (per 75) ahead of the league average scoring rate. Curry was 16 points ahead of the average scoring rate in 2016. Most other players at such rarified efficiency are around or below the league average scoring rate.
Kareem’s early ’70s seasons are impressive in their own right, although the splintering of talent to the ABA takes some of the shine away. Amare was great, but he also had Nash. Wilt’s 50-point season and Jordan’s 1991 (not his explosive 1987) are also worthy candidates as the most valuable scoring year ever.
But for me, the gap in efficiency is too large to ignore. I buy what ScoreVal is selling: Stephen Curry’s 2016 scoring season is the most impactful in NBA history.
- I’ve built a slightly more precise “bad pass estimator,” but the increased accuracy barely makes a difference in ScoreVal for almost every player. For some players it is occasionally worth one or two tenths of a point.
- For example, Magic Johnson lost a percentage point in relative “true scoring” compared to relative true shooting in 1984 because he was turnover prone. But, as Magic’s turnovers declined, he actually gained efficiency ground in the second half of his career when including these scoring turnovers.
- When we have better data in the future, these numbers will certainly change a bit.
- For example, James Harden led the league in 2019 with 3.9 attempts per game in the final 4 seconds of the shot clock. He took about 2 more last-second shots than the typical offensive star — around double! — and completely removing those 2 extra shots would only raise his eFG from 54.1 percent to 54.7 percent.
- Bob Love in 1975 and Mike Mitchell in 1981.
- If you’re wondering, the worst scoring season with at least 1000 minutes played was Will Perdue’s 2000 campaign, in which he scored 7 points per 75 possessions on 39 percent true shooting, for a -2.3 ScoreVal.
- I’ve mentioned Jacob Goldstein’s “Efficere” before, which is a different approach to a similar topic. In that case, the question is more about “how much value does a player add, given the shots that he takes?” Efficere also has 2016 Steph Curry No. 1 All-Time, and by a mile over the next-best seasons (2013 and 2014 LeBron James).
- Jerry West also led the league in scoring rate and efficiency in 1965, but he didn’t win the scoring title that year.
Comments are closed.