Nylon Calculus: Golden State and league-adjusted statistics

Apr 8, 2017; Oakland, CA, USA; Golden State Warriors forward Kevin Durant (35) walks onto the court after a timeout against the New Orleans Pelicans in the second quarter at Oracle Arena. Mandatory Credit: Cary Edmondson-USA TODAY Sports
Apr 8, 2017; Oakland, CA, USA; Golden State Warriors forward Kevin Durant (35) walks onto the court after a timeout against the New Orleans Pelicans in the second quarter at Oracle Arena. Mandatory Credit: Cary Edmondson-USA TODAY Sports /
facebooktwitterreddit

With the addition of Kevin Durant, the Golden State Warriors came into this season with the potential to be most efficient offense ever. The starting point for whether that was realistic was by adjusting their points per 100 possessions during the 2016 season, and statistics vital to that mark for league averages. We’re overdue for a follow-up on how this season’s Warriors performed historically with Durant (for the most part), but it only seemed fair to also publish the same adjustments to statistics used in the first post for every team since 1947.

To start, accounting for league trends this season was especially important. At 108.8 points per 100 possessions, league-average scoring offensive efficiency was the highest Basketball-Reference has ever calculated, which goes back to 1974. Teams were so efficient in part because they attempted the highest portion of their shots from beyond the arc (31.6 percent), but also set new records in free throw accuracy and taking care of the ball. There were an average of only 12.7 turnovers per 100 plays, topping last season’s then-lowest mark of 13.2. That helped off-set a similar trend in offensive rebounding, down to a success rate of 23.3 percent after the previous record-low of 23.8 was recorded last season. Basketball was also played at the fastest pace since 1993 at an average of 96.4 possessions per 48 minutes. This was not your ordinary season.

The way I adjusted for league averages was simple. I divided a team’s mark by the league average that season, multiplying by 100, and adding a “+” to the name of the statistic measured. This was what Andrew Lynch used for offensive and defensive efficiency in 2014, creating ORtg+ and DRtg+.

Read More: Shoot your shot — the hotspots for every playoff team

To go back to the 2016 Warriors, they scored 114.5 points per 100 possessions when the league average was 106.4. That’s good for an ORtg+ of 107.6. Their raw offensive rating was 12th-best since 1974, but their ORtg+ was third. To branch off their ORtg+, it was so high because their eFG+, which tweaked their effective field goal shooting, was the highest since the 1971 Bucks. To go off their eFG+, it was so high because they were the best 3-point shooting team with a normal 3-point line by raw 3-point percentage, and with a minimum of 1,000 attempts, 3P%+.

Hopefully you get the drift. These adjustments can be made efficiently with Basketball-Reference’s Team Season Finder, league averages, and season summaries, among other methods. Of course, dividing team statistics by league averages is only one way to tweak raw numbers, but that’s what I stuck with here. Outside of 3-point percentage, I adjusted advanced-ish statistics like net ratings, offensive and defensive efficiency, true and effective field goal percentages, team and opponent three-point and free throw rates, rebound rates, turnover rates, and pace.

As an example of how these adjustments look all at once, I compared the last two seasons from Golden State. While this season’s squad was a notch below 2016’s by win percentage, they had the advantage in point differential and Net+, which divided raw offensive efficiency by raw defensive efficiency. On offense, they scored about one point more per 100 possessions than last season, but by ORtg+ they performed slightly worse. The table below includes their raw numbers, league averages, team adjustments for league averages, and where those adjustments rank since they were first able to be calculated on Basketball-Reference. With the exception of 3s debuting in 1980, shooting percentages date back to 1947 while opponent statistics go back to 1971 and possession statistics go back to 1974.

Because of the scoring environment this season, the ORtg+ of this season’s Warriors ranked 23rd since 1974. If league-average offensive efficiency was the same as last season, their ORtg+ would’ve been 108.6 and second-best, but still 0.3 points from the 2004 Dallas Mavericks’ mark of 108.9.

The most noticeable change in Golden State’s raw numbers was their 3-point percentage. While they ranked third this season, it wasn’t record-breaking like their performance in 2016. This year, they’re 54th by raw 3-point percentage and 145th by 3P%+, but that went up to 55th out of 566 teams if we filter out those with less than 1,000 attempts. One reason they fell back to earth this season were the drops in percentages from Stephen Curry and Draymond Green, including the late-February to early-March slump when Curry shot 24 percent. They also lost the spot-up shooting from Harrison Barnes, Brandon Rush, and Marreese Speights, but that’s obviously fine when it means acquiring Durant.

More advanced reasons for the dip in team 3-point percentage exist, but that’s not the point of this post. The hard-hitting analysis here is that being historically great over 82 games is difficult. Regardless, Golden State’s adjusted shooting percentages were still elite after finishing first in accuracy around the rim and the second straight season on 2-pointers outside 15 feet. However, because little things matter, their free throw rate dipped and a higher percentage of their plays were turnovers down the stretch. Note that a low TOV%+ was good on offense and bad on defense.

Despite the “drop” in scoring, the Warriors had a higher Net+ than last season because of a similar defensive rating. Below is a look at how their defensive metrics looked after adjusting for league averages:

Overall, Golden State is just the third team since 1974 to be a top-100 team in DRtg+ and Pace+, joining their 2015 squad and the 2003 Sacramento Kings. The largest leap from last season was turnovers forced, where Zaza Pachulia and Green’s steals per 100 possessions were career-highs. But the most unusual uptick was David West. The 36-year-old averaged over two steals and two blocks per 100 possessions, only the seventh time a player his age and his height or taller accomplished such a feat, minimum 800 minutes.

A concern should’ve been how many 3-point attempts the Warriors conceded, but opponents missed at a near-historic level. When looking at only teams with at least 1,000 allowed 3-point attempts, Golden State ranked seventh in opponent 3P%+. Weirdly, the best mark belongs not to a Spurs squad or Tom Thibodeau’s Bulls, but the 11th-best defense of 1995 in the Sacramento Kings. Golden State also locked up from three to 16 feet. That accounted for only 20 percent of all allowed shots, but it chipped away at  overall shooting percentages.

Hopefully the adjusted statistics made sense. Unfortunately, the Warriors’ example spoiled some of the fun in that while this season produced a historically dominant team by Net+ and raw point differential, there’s no historically great offense or defense. But hopefully a couple of dashboards further down in this post, which feature adjustments and rankings since 1947, were still interesting.

There are a few things to note about each dashboard. The default settings include only seasons with estimated pace, but that can be changed by checking ‘All’ for the pace era filter or filtering years. In the first dashboard with adjusted statistics, the ‘Stat Rank’ column doesn’t change if you sort from worst to best, probably fine since the 14th-worst offense since 1974 sounds better than 1,150th-ranked, but it made sorting weird. You must click on the name of a stat first and then sorting options appear, when typically only hovering a statistic does the trick. There’s also a ‘Total Tms’ column which shows only the total teams in the current filter. Lastly, an asterisk means playoff appearance as seen on Basketball-Reference, but it’s also another filter.

The second dashboard should appeal to those interested purely in rankings rather than the actual adjusted statistics themselves. For example, the 1996 Chicago Bulls were the only team since 1974 with a top 50 ORtg+ and DRtg+ while the 2015 and 2017 Golden State Warriors were two of the five teams to rank in the top 100 in both.

Hopefully these dashboards are easy to navigate and add more context to performances not just from this season’s teams, but also the past. The Warriors and Rockets, among others, broke new ground in 3-point rate this season, but before them were the Orlando Magic stretching the floor around Dwight Howard, the Boston Celtics in the early-2000s with Antoine Walker, and “The Bomb Squad” from the 1989 Knicks.

You can also view the dashboards here, which include more filters because of additional width and height.

All statistics and adjustments were according to totals at Basketball-Reference.