On the eve of the 2019-20 NBA season, let's take a final look at our projections using ESPN's real plus-minus (RPM). These differ from our initial projections, released in mid-August, by incorporating injuries and roster changes since then, as well as updating my guesses at playing time based on preseason rotations.
While it has certainly been an eventful month in the NBA, preseason seems to have had less impact on the projected standings than in most years. No. 1 overall pick Zion Williamson was the only starter to suffer what appears to be a potentially long-term injury, meaning changes in the projections are more attributable to rotations shaking out around the league than any other factor. After a quick explanation, let's get to our final projections.
How do the projections work?
Our RPM projections utilize the multiyear, predictive version of RPM as a starting point. They're adjusted for typical player aging and -- new for this season -- then regressed toward the player's projected offensive rating and defensive rating from my SCHOENE projection system, based solely on box-score stats. (For players without RPM projections, including rookies, the SCHOENE ratings are used instead.)
Games played are projected based on time missed over the previous three seasons. I then make a subjective guess at minutes distributions for each team. (You can see my full minutes projections here.) Multiplying those minutes by players' offensive and defensive ratings yields team ratings that translate into expected wins. I used those projections to simulate the season 10,000 times and record the average number of wins as well as how often each team made the playoffs.
Why are these projections so compact?
Amazingly, just three teams are projected for more than 50 wins this year. In part, this is a product of the conservative nature of projections. While we know that more than three teams will win 50-plus games, we don't know for sure beforehand which teams will do so.
That said, the parity the RPM-based model forecasts appears unique to this season. Using the exact same method with 2018-19 projections yielded six teams with projections of 50-plus wins and a seventh whose projection rounded up to 50.
What makes this season so wide open?
Besides the absence of a single dominant team, with the Golden State Warriors weakened by injury and attrition, this season is also unique because of the level of roster turnover this summer. I project just 62% of minutes leaguewide to be played by returning players, as compared to 76% in 2018-19. That's important because of the tweak to the projections I made last year to treat players who change teams differently from those who remain with the same team.
Regressing projections toward the player's SCHOENE projections rather than league average improved out-of-sample projections and no longer penalizes stars quite so harshly for changing teams. Still, it's clear that even stars do pay an RPM price for changing teams.
Within the sample I used for testing (back through 2012-13), 13 players who posted an RPM of 5.0 or better changed teams. On average, their RPM declined from 6.0 to 3.3. By comparison, players with an RPM of 5.0 or better who stay with the same team see a much smaller drop-off, from an average of 6.4 to 4.8. This effect is particularly relevant for the Brooklyn Nets, LA Clippers and Los Angeles Lakers, all of whom added players whose projections would rank among the NBA's top 15 if not for the adjustment.
Given that, it's no surprise that the top three teams in the projections all had a relatively high degree of continuity this offseason and return three of last season's top five players in RPM.
Western Conference
1. Houston Rockets
Average wins: 54.9
Playoffs: >99%
Counterintuitively, the Rockets' projection actually went up slightly with Gerald Green presumed out for the season with a possible Lisfranc fracture in his left foot. I assigned most of his minutes to Thabo Sefolosha, who projects slightly better by RPM.
Still, losing Green should test Houston's depth. The Rockets are already essentially down one spot because they can play Nene in only nine games before risking him earning a $2.4 million bonus that would take them into the luxury tax. Green's absence would cost them another spot, so minor injuries could severely compromise the Rockets' rotation.
2. Denver Nuggets
Average wins: 53.6
Playoffs: 99%
Given they return a league-high 88% of last season's minutes and have enviable depth with the additions of Jerami Grant and rookie Michael Porter Jr., the Nuggets might have the highest floor of any team this season. If you're picking the most likely team to win 50 games, it's probably Denver, coming off 54 wins in 2018-19.
3. LA Clippers
Average wins: 47.1
Playoffs: 87%
The Clippers' modest regular-season projection dipped further with the acknowledgment by coach Doc Rivers that All-Star wing Paul George will miss at least the first 10 games of the season. Though rumors were already swirling back in August that George would not be ready for opening night, I hadn't yet docked his projection any games in the original version.
4. Utah Jazz
Average wins: 46.8
Playoffs: 86%
In part because Mike Conley is conservatively projected for 64 games, the Jazz's offensive projection (11th in the league) might undersell their room for improvement. They hope to combine an offense that generated the league's highest-value shots (their quantified shot quality of 54.8% was the league's best, per Second Spectrum tracking) with shooters capable of actually taking advantage of those opportunities.
5. Los Angeles Lakers
Average wins: 46.4
Playoffs: 84%
Technically, DeMarcus Cousins' ACL tear occurred after the original projections, which we updated later that week once the injury was confirmed. The addition of Dwight Howard, projected slightly better than replacement level for a center, didn't do much to affect the Lakers' outlook either way.
6. Golden State Warriors
Average wins: 45.6
Playoffs: 80%
Golden State's projection declined a touch with Alfonzo McKinnie's minutes going to rookies Eric Paschall and Jordan Poole after he was waived to make room on the roster for training-camp invitee Marquese Chriss.
7. Dallas Mavericks
Average wins: 43.9
Playoffs: 68%
The Mavericks' position as a solid playoff team ahead of two teams that reached the playoffs last season (Portland and San Antonio) remains one of the biggest surprises from RPM's projections, but it's an outlook shared by many stat-based projections. FiveThirtyEight's similar RAPTOR model has Dallas averaging 45 wins, the West's eighth-best total.
8. Portland Trail Blazers
Average wins: 40.6
Playoffs: 42%
At Blazers media day, All-NBA guard Damian Lillard told reporters, "What the experts' percentages of us making the playoffs are, that's I would say the least of our concerns." There's certainly a disconnect between Portland's internal expectations of competing for a championship and its statistical projections (FiveThirtyEight also has the Blazers at 41 wins on average and less than 50/50 to make the playoffs).
9. Sacramento Kings
Average wins: 39.8
Playoffs: 35%
While the Kings' projection is close to the 39 games they won last season, it still represents a huge step forward from this time a year ago. Sacramento's 26.1-win projection entering 2018-19 was the league's second lowest.
10. Minnesota Timberwolves
Average wins: 39.5
Playoffs: 32%
Minnesota star Karl-Anthony Towns told Marc J. Spears of The Undefeated that observers should "keep sleeping on us," but statistical projections aren't. FiveThirtyEight's model is even more bullish on the Timberwolves, who average 43 wins in those projections.
11. San Antonio Spurs
Average wins: 38.6
Playoffs: 26%
If the Spurs were any other team, we'd look at their weak projections and rough preseason (San Antonio's minus-6.4 net rating ranked 28th in the league, per NBA Advanced Stats) and say their 22-year playoff streak is in grave jeopardy. However, betting against Gregg Popovich's teams has typically been a losing proposition.
12. Phoenix Suns
Average wins: 38.2
Playoffs: 23%
With Zion's injury, the Suns jumped New Orleans. The 8.2-win difference between Phoenix's RPM projection and the team's over/under win total (30) at Caesars Sportsbook is the single largest in the NBA.
13. New Orleans Pelicans
Average wins: 37.5
Playoffs: 19%
As noted in Friday's analysis of Williamson's injury, projecting him out for the team's first 20 games costs New Orleans 0.7 wins on average and drops the team's chances of making the playoffs from 22% beforehand.
14. Oklahoma City Thunder
Average wins: 36.7
Playoffs: 15%
To illustrate the depth of this year's West, last year's No. 14 team in the projections was Sacramento at the aforementioned 26.1 wins. That depth could end up altering Oklahoma City's outlook. If the Thunder are actually 14th in the West, they'll likely trade impending free agent Danilo Gallinari by the deadline.
15. Memphis Grizzlies
Average wins: 32.4
Playoffs: 3%
For a likely last-place team in their conference, the Grizzlies figure to be competitive behind veterans Jonas Valanciunas and Jae Crowder and their young core of Ja Morant and Jaren Jackson Jr. Remarkably, they win more games on average than the bottom four teams in the East.
Eastern Conference
1. Milwaukee Bucks
Average wins: 50.7
Playoffs: 99%
Though the Bucks' win projection is rather modest, they're the only team in the league forecast to finish in the top five in both offensive (fourth) and defensive (first) rating. Houston (second in offensive rating, sixth on defense) just misses out.
2. Boston Celtics
Average wins: 47.4
Playoffs: 95%
Because of their strong perimeter defenders, the Celtics are projected to finish seventh in defensive rating. That might overstate Boston's potential with weakened rim protection after losing both Al Horford and Aron Baynes.
3. Philadelphia 76ers
Average wins: 47.2
Playoffs: 95%
As a reminder, Philadelphia's modest win projection owes in large part to both Joel Embiid and Ben Simmons being projected for 66 games based on time missed over the past three seasons. While that seems reasonable for Embiid, who has averaged 63.5 games the past two years, it's harsh for Simmons. After sitting out his entire first season, Simmons played all but four games over 2017-18 and 2018-19.
4. Orlando Magic
Average wins: 46.4
Playoffs: 93%
Orlando still finishes fourth in the East on average, even with a shift to give Mo Bamba primary backup center minutes instead of the higher-rated Khem Birch. Bamba averaged 17.6 minutes to Birch's 12.8 in the preseason and is looking more productive than he was as a rookie.
5. Toronto Raptors
Average wins: 45.8
Playoffs: 91%
The Kyle Lowry extension makes it all the more likely that the Raptors will keep what's left of the championship team intact and contend for home-court advantage in the first round instead of tearing it down at midseason. Of course, big changes will likewise shift their projections.
6. Miami Heat
Average wins: 43.0
Playoffs: 78%
Despite the excitement over his preseason play, Tyler Herro projects worse than replacement level as a rookie based on his translated performance at Kentucky. In particular, Herro is unlikely to keep shooting so well on jumpers. He made 20 of 38 attempts outside the paint during the preseason (53%), as compared to 41% in college, according to Synergy Sports tracking.
7. Indiana Pacers
Average wins: 42.2
Playoffs: 73%
A key question for the Pacers is when Victor Oladipo will return from a quadriceps tendon rupture and how much he'll play thereafter. Oladipo is projected here for 49 games and a little less than 1,400 minutes under the assumption he'll be back in mid-December.
8. Brooklyn Nets
Average wins: 41.3
Playoffs: 65%
It's worth watching how well the Nets play defensively. Their 109.0 defensive rating last season was almost exactly league average, but the team's top six regulars in on-court defensive rating, according to NBA Advanced Stats (Ed Davis, Shabazz Napier, Rondae Hollis-Jefferson, DeMarre Carroll, Allen Crabbe and Treveon Graham) all left over the offseason. That helps explain why Brooklyn is projected to decline to 23rd defensively.
9. Detroit Pistons
Average wins: 39.6
Playoffs: 52%
Detroit's projection benefits slightly from the news that preseason standout Christian Wood will make the team. Giving him some of the minutes previously assigned to Thon Maker strengthens the Pistons' rotation in terms of projected RPM.
10. Chicago Bulls
Average wins: 38.9
Playoffs: 46%
Similarly, Chicago could benefit if opening-night starter Tomas Satoransky establishes himself as the clear choice at point guard ahead of Kris Dunn and rookie Coby White, who projects as an ineffective contributor during his first season.
11. Washington Wizards
Average wins: 33.0
Playoffs: 9%
The Wizards project as a deep fringe playoff contender in the East exclusively because of All-Star shooting guard Bradley Beal. He alone projects to 9.2 wins above replacement based on RPM, as compared to 1.3 wins below replacement for the rest of the Washington roster.
12. Atlanta Hawks
Average wins: 30.7
Playoffs: 3%
After a strong second half, the Hawks are a trendy pick to take a step forward this season. That hype is likely a year ahead of schedule. There's scant evidence that in-season improvement tends to carry over to the following campaign, and Atlanta is projected to finish 28th in defensive rating.
13. Charlotte Hornets
Average wins: 30.1
Playoffs: 3%
Despite significant rotation adjustments as the Hornets committed to their young players during the preseason, starting recent first-round picks Miles Bridges and PJ Washington together at forward, the Hornets' win projection barely budged at all -- probably an indication Charlotte is right to sit the veterans.
14. Cleveland Cavaliers
Average wins: 26.2
Playoffs: <1%
The Cavaliers' final record may depend largely on how much Kevin Love plays this season after being limited to 22 games in 2018-19. He's projected here for 60 games and about 1,700 minutes, similar to his 2017-18 total. Cleveland was 6.9 points per 100 possessions better with Love on the court, per NBA Advanced Stats.
15. New York Knicks
Average wins: 26.1
Playoffs: <1%
Out of the 10,000 simulations of the season, the Knicks made the playoffs a league-low 31 times -- 0.3% of the time.