2016-17 NBA Preview: Golden State Warriors

May God help us all.

“I am become Death, the destroyer of worlds.” – Steve Kerr

2015-16 in review

73 wins. Their first loss was the 25th game, which just a day after a double-overtime win. At one point they were 48-4. Yet they lost in the finals, and in retrospect they were far from invincible. Stephen Curry’s injury limited their power. While they broke the wins record, they had the point differential of a mere 65-win team. Their defense also had a lucky break: their opponents hit a lower 3PT% than expected, and a basic adjustment would push their point differential down almost an entire point. They weren’t perfect, but their Death lineup was a great weapon and they deployed it at the end of games, helping them secure more close wins than expected.

Rotation players in: Kevin Durant, Zaza Pachulia, David West.

Rotation players out: Harrison Barnes, Andrew Bogut, Festus Ezeli, Brandon Rush, Mo Speights.

In bringing in Kevin Durant, the Warriors had to lose a few of their role players. Andrew Bogut never played huge minutes and wasn’t a part of the Death lineup, but his defense and rebounding will be missed — same with Ezeli. Harrison Barnes’ ability to defend power forwards could be missed, but he really provided very little to the team. Brandon Rush was a useful wing, and the Warriors actually performed very well with him in their typical lineups. Mo Speights was a shoot-first big man who was a little out of place on the team. David West is a good replacement, as he has a good midrange jump shot too but is a better defender. Finally, Zaza has rated surprisingly well in advanced stats, thanks to his passing and no-frills defense, and the fact that the Warriors got him so cheaply is a crime.

2016-17 projected

The Warriors replaced Harrison Barnes with Kevin Durant, and several betting sites are setting their over/under win projection at around 66. There are several factors entangled here, and it’s tough to suss out exactly how things will unfold. First, as already discussed, the team had a point differential of a 65-win team, and the ability to win close games is not something that usually holds over from one season to the next. The Warriors could be an exception, because their now Mega-Death lineup is so powerful. But that is the baseline for analytics-based projections. Also, again, their defense was lucky, as I found that opponent three-point percentage was not a controllable attribute.

The disparity between their win total and their team rating last year will be eaten up by the difference between Kevin Durant and Harrison Barnes, however. I’ve explained this before. There is little evidence Barnes was an above average player, and he received little defensive coverage when Stephen Curry and Klay Thompson were on the court. Even so, he did not capitalize very much on the decreased defensive pressure. Kevin Durant, meanwhile, will receive a lot of attention on defense, and he has the ability to destroy single coverage plans. Teams figured out they could put their big man on Barnes in the post-season, but that won’t work anymore; Durant can beat slower players off the dribble with ease.

One could cite some fit concerns, but there’s really only one legitimate issue, since Durant appears to be fitting in seamlessly already: there’s still only one ball to go around. When high-usage players come together, the sum of the parts is less than expected because their value was predicated on having a high usage rate. Since Kevin Durant is coming from Oklahoma City where he shared the ball with Russell Westbrook, it’s not a huge adjustment for him, but overall he’s going to be taking away a few shots from everyone — Harrison Barnes shot the ball a lot less frequently. Overall, the usage adjustment hurts them a little, but they’re so far into the stratosphere it’s not damaging.

The Warriors on paper have a team that projects past 70 wins, and for win projection systems, which are built to be conservative, that’s absurd. Kevin Durant is usually worth a few more wins, but there are diminishing returns with respect to wins at the extreme ends. So the limiting factor, then, is how much they’ll rest their players and how careful they’ll be with injuries. Of course, if you blow out most of your opponents, you don’t have to play your key guys in the fourth quarter. Their top guys last season were only playing 33 or 34 minutes per game. But one would have to craft some low minutes totals for a projection to suggest anything less than 66.

Even though Golden State’s dominance is self-evident, the record books have not already been written through the end of the year. Chaos is still a huge force in the league, and even an overwhelming favorite only wins the title about 50 to 60 percent of the time. There are still many scenarios in which an underdog wins, and that’s gotta be a compelling storyline to follow for any fan in the playoffs. They are historically great, but no team is invincible.

Quick statistic

Stephen Curry’s three-point shooting prowess is Ruthian and it’s hard to understand his value there in comparison with other players because he’s so far away from the pack. For example, he’s already 19th all-time in three-pointers, but at this current pace the record of 2973 from Ray Allen will be obliterated. As you can see from the graph below, Curry’s pace is incredible. He’s made nearly 1000 treys in three years, and with another season like the last one he’d be fifth all-time. And at that rate he’s shown over the past three seasons, he could hit Reggie Miller’s total by age 30.



I want to repeat this again: the Golden State Warriors replaced Harrison Barnes with Kevin Durant. Now defenses have to worry about guarding the trio of Durant, Stephen Curry, and Klay Thompson outside the three-point line. They were already the third best offense ever, in terms of offensive rating relative to the league average. Dallas in 2004 is first, and Phoenix in 2005 was second. And yes, Steve Nash was on both of those teams, and it’s mind-boggling to imagine what’s going to happen now. They still have to contend with usage/fit issues, how they had great luck in close games and opponent three-point percentage last season, and they’ll probably rest their players a lot more frequently. But they will be ludicrously great.

Win predictions:

Mine: 69.9. A blend of several metrics, including Dredge, with a few other factors considered, like coaching and usage.

Andrew Johnson’s: 72. A combination of PT-PM (a SportVU player tracking metric) blended with RAPM. Two-time reigning champion of the APBRmetrics board predictions contest.

Nick Restifo’s: 71. A simulation using BPM and RPM for player value, which includes game effects like altitude and rest.

Kevin Ferrigan’s: 70. A player projection system with inputs from RAPM, BPM, height, and age.

This article originally appeared on