Skip to main content
ABC News
Superteams Like The Warriors Usually Underperform — Will The Kevin Durant Era Be Different?

As we approach the sixth anniversary of LeBron James’s infamous “Decision,” another free-agent megastar has taken his talents to a superteam. Kevin Durant announced Monday that he was leaving the Oklahoma City Thunder for the Golden State Warriors, joining a team that had already set the league’s all-time wins record without him. The resulting roster could easily end up being the most stacked in NBA history — though there’s also good reason to doubt that they’ll be quite as dominant as they seem now.

The best ever — on paper

First things first: These new-look Warriors probably won’t match their own record for victories, no matter how much of a boost they get from Durant. Sure, it could happen again, but too many things would have to go right at once to make another 73-win season likely. Even last season’s Warriors barely held on to break the 1995-96 Chicago Bulls’ record, clearing it by a game after losing twice in the regular season’s penultimate week. (To say nothing of the team’s relative slog through the playoffs, in which the Warriors lost another nine times in only 24 games.) Winning all but eight or nine games in an 82-game schedule is ridiculously hard, no matter how gifted your roster is.

But make no mistake, by the numbers, the 2016-17 Warriors could easily prove to be the most talented team in modern NBA history. If we use’s Simple Projection System to predict each player’s Box Plus/Minus for the following season and sum up those projections for each team,1 next year’s Warriors would be the most-talented team since 1979-802 if they merely surrounded their existing players — Stephen Curry, Klay Thompson, Kevin Durant, Draymond Green, Andre Iguodala, Shaun Livingston, Zaza Pachulia and David West, who is reportedly signing with Golden State on the veteran’s minimum — with replacement-level talent (think Luc Mbah a Moute or Norris Cole). In other words, even if they don’t add another meaningful player, Durant’s Warriors will probably surpass James’s 2010-11 Heat as the best collection of talent in the league’s past four decades or so.

2017 Warriors* 8.2 4.1 12.3
2011 Heat 7.6 4.2 11.7 4.3 3.2 7.5
2015 Cavaliers 9.5 0.6 10.2 5.5 -1.1 4.4
1993 Bulls 5.9 3.4 9.3 4.9 1.6 6.5
2012 Heat 4.6 4.3 8.9 2.6 3.7 6.3
1997 Bulls 3.6 5.0 8.6 7.6 4.3 11.8
1998 Bulls 3.7 4.6 8.3 3.1 4.9 8.0
2016 Thunder 6.1 1.9 8.0 6.5 0.9 7.4
1991 Lakers 6.3 1.6 7.9 4.1 2.8 7.0
1996 Bulls 1.7 6.1 7.8 7.3 5.6 13.0
1995 Magic 4.9 2.8 7.7 7.0 -0.2 6.8
The most talented “on paper” NBA teams since 1980

* 100 percent of leftover minutes given to replacement-level players.


And as my colleagues Nate Silver and Kyle Wagner wrote Monday, these superteams often manage to squeeze better-than-replacement talent onto their rosters despite a lack of cap space, since their high odds of winning a championship are a strong lure for veterans. So it’s possible that Golden State will cobble together even more pure talent before the offseason is over.

Talent vs. diminishing returns

But that Miami team should also serve as a powerful restraint on enthusiasm about the 2017 Warriors’ potential. Before 2010, we’d never really seen a group of team-carrying stars assembled in their primes like James, Dwyane Wade and Chris Bosh; it was uncharted territory, and the possibilities seemed endless. But despite such unprecedented ability, Miami struggled to live up to its perceived potential, becoming the first in a line of superteams that would not be as good as the sum of their parts.

Now that we’ve seen a few examples of this type of team, we can see that as a team’s talent level rises (as measured by BPM), so too does its likelihood of underperforming relative to that talent:


The root cause is probably age-old: diminishing returns. And it’s a problem amplified in basketball. With only five players on the court at any given time — and only one who can possess the ball — adding talent can help a team only so much. Eventually it becomes more important to surround a star with less-talented players whose skills complement his own, rather than another star whose overlapping abilities are largely wasted. Although some superteams have been constructed in a way that harnessed their stars’ full potential — the 1996-97 Bulls surpassed expectations, for instance, becoming greater than the sum of their hyper-talented parts — most stockpile too much of one skill set or another and end up underperforming as a result.

The superteam’s secret formula

So what’s a superteam like Golden State to do? Is there a formula for battling the scourge of diminishing returns?

Perhaps — though only at the margins.

Among the top 10 percent of teams in our sample according to projected talent,3 a few trends emerged that had some small predictive power over a roster’s ability to meet its expectations. Unsurprisingly, even after controlling for their overall projected talent levels, teams that collected more high-usage players4 tended to underperform, which speaks to the dangers of assembling a lineup of scorers and having them fight over one ball. Also, teams stocked with rebounders tended to fall short of expectations more often than the average squad.

At the other end of the spectrum, teams rife with shooters, passers and defenders5 tended to resist the pull of diminishing returns more effectively than their peers. This makes perfect basketball sense: Unlike on offense, adding defensive talent seems to only make a team stronger, all else being equal, and ball movement and floor-spacing have long been known to help an offense become more than the sum of its parts.

But again, these factors make a small difference at best. And how they pertain to the 2016-17 Warriors is still a little complicated. Regardless of roster construction, Golden State will probably still fall well short of its projected team BPM when the team members take the court together, and Durant gives them a third scorer (with Curry and Thompson) who ranked among the NBA’s top 30 players by usage rate (two were in the top 10) last season. Think of the diminishing returns!

On the other hand, staggering Durant’s minutes with Russell Westbrook’s worked well for Oklahoma City last season, and with four stars to spread across bench lineups instead of two, Golden State may find ways to minimize wasted potential possessions for its stars throughout the game. Besides, the Warriors also figure to have a mix of shooting, passing and defense6 that will be unrivaled in the league. That combination alone will probably help Golden State gel better than other superteams have in recent seasons, even if it can’t fully escape the force of diminishing returns.


  1. Weighted by each player’s actual minutes played during the season in question — or, in the case of the 2017 Warriors, predicting their minutes using our CARMELO system in concert with updated depth charts.

  2. The earliest season for which BPM projections can be calculated.

  3. Ninety-seven teams.

  4. As defined by their percentile ranking in projected usage rate, according to the same Simple Projection System methodology we used to project BPM.

  5. Measured by the percentile rank of their projected true shooting percentage, assist rate and defensive BPM, respectively.

  6. Don’t forget, the Warriors were the league’s sixth-best defensive team last season, and Durant has worked hard to get pretty good at that end of the floor, too.

Neil Paine was the acting sports editor at FiveThirtyEight.