The Good, the Bad, and the Wild From 21 Years of the NBA’s Annual GM Survey

The 22nd annual survey of NBA general managers went live on NBA.com Tuesday. As is the case every year, it offers a delightful preseason peek into the minds of the league’s top decision-makers.

This year, general managers mostly see a three-team title race between the defending champion Denver Nuggets, the Boston Celtics, and the Milwaukee Bucks. They believe Nikola Jokic will win his third MVP award in four seasons. And they think Victor Wembanyama is a generational talent, already placing him second—behind Jokic—as the player they’d most want if they were starting a new franchise.

But how often do general managers actually get it right? How much weight should we place on their predictions? We sought to investigate, using the 21 previous years of the GM survey.

The survey has expanded to contain 50 questions, and this report card won’t grade most of those. For instance, we didn’t test subjective areas like the league’s best players at each position, the standouts with the best skill sets, or the best coaches.

This year, for instance, general managers overwhelmingly believe that Jokic is the best center in the NBA, despite his loss to Joel Embiid in last season’s MVP race. They think new Celtic Jrue Holiday is the best perimeter defender, even though he struggled to contain Jimmy Butler in the 2023 postseason. And they gave Erik Spoelstra a near sweep of all the “best coach” questions. There’s value in what those categories reveal about internal league opinions, but we can’t evaluate them as simply right or wrong.

Instead, we chose to zoom in on a handful of objective questions that we can easily judge in retrospect. On the team side, were general managers right or wrong about their picks of Finals favorites and best regular-season squads? On the individual side, how well did they forecast races for MVP and Rookie of the Year? And, based on those outcomes, how much stock should we place in their predictions for the 2023-24 season and beyond?

NBA Champion

For a while, general managers enjoyed smooth sailing in the question that kicks off the annual survey: Which team will win that season’s championship?

From 2004-05 through 2017-18, general managers’ consensus preseason championship pick went on to win the title nine times in 14 seasons. There were some big misses along the way—only one general manager picked the Celtics in 2008, and nobody picked the Mavericks in 2011—but for the most part, they made the right choices when teams like the Kobe-and-Pau Lakers, LeBron-and-Wade Heat, and dynastic Warriors ruled the league.

Those correct predictions mostly reflected the overall consensus inside and outside the league, which was relatively predictable for more than a decade. The preseason Vegas favorite won the title in eight of those 14 seasons, per an analysis of Sports Odds History data.

But that era is over. Before the 2018-19 season, not a single general manager picked the Raptors to win the title—an understandable whiff, because the Warriors were going for a three-peat with Kevin Durant, but a whiff nonetheless. Only 11 percent picked the Lakers in 2019-20, which placed LeBron James and Anthony Davis’s team a distant third behind the favored Clippers and Bucks. And none of the past three champions have received a single preseason vote: not the Bucks in 2020-21, nor the Warriors in 2021-22, nor the Nuggets in 2022-23.

Those recent misses came even though the 2020-21 Bucks were the no. 2 preseason favorite according to Vegas odds, and the 2021-22 Warriors were tied for third. This chart shows all the title winners, along with their placement in the GM survey, since the 2002-03 season.

A closer examination of general managers’ picks suggests that when they’ve erred, they’ve mainly suffered from recency bias. They have picked a repeat champion 11 times in 21 years (not counting 2023-24), but they’ve gone just 3-8 in those 11 predictions. Their recent yo-yo with the Lakers is illustrative: Only 11 percent of general managers picked the Lakers to win in 2019-20, when they did indeed triumph, but a whopping 81 percent picked a repeat the following year, when L.A. lost early in the playoffs. As is true of many fans and public analysts, their predictions often come a year too late.

Regular-Season Success

General managers have fared better, and more consistently, when forecasting regular-season success, which makes some sense given that 82 games provide a larger sample than shorter playoff series. Over the past 21 seasons, 62 percent of teams that received the most votes to be one of the top regular-season teams in each conference ended up finishing that high. (Through the 2014-15 season, the survey asked for the six division winners; starting in 2015-16, it has asked for the top three seeds in each conference. We’ve graded it based on how the question was phrased during the season in question.)

Five Columns That Will Define the 2023-24 NBA Season

• Five Players
• Five Teams
• Five Rivalries
• Five Trade Candidates
• Five Questions

In that light, it’s notable—if not particularly surprising—that all 28 rival general managers picked the Celtics and Bucks to finish first and second in some order in the East this season. (For all these questions, GMs aren’t allowed to vote for their own teams or players.) Both inside and outside the league, people clearly see a wide gulf between the conference’s top two teams and other wannabe contenders.

However, an analysis of general managers’ regular-season picks suggests there’s typically one big surprise every year. Most recently, the 2023 Kings, 2022 Grizzlies, 2021 Suns, 2019 Trail Blazers, 2018 Trail Blazers, and 2018 76ers all landed in the top three seeds despite receiving no preseason votes to do so.

Perhaps because of better depth in the Western Conference, the West has hosted the vast majority of those surprise contenders. Only 12 Eastern teams in 21 years either won their division or landed a top-three seed after receiving less than 25 percent of votes. For comparison, 22 such teams did so in the West.

That also means, conversely, that big Western favorites were much more likely to come up short of expectations. Out of 41 Eastern teams that received at least 75 percent of general managers’ votes to land a top-three seed, 33 did so, for an 80 percent success rate. But out of 38 such Western teams, only 21 actually earned a top seed, for a mere 55 percent success rate.

MVP

The juiciest individual award is also where general managers are the least accurate. In 21 years of this survey, the GMs’ preseason MVP favorite has gone on to win the award only four times—and three of those four times, the victor was LeBron.

Meanwhile, not a single general manager picked Steve Nash to win MVP in 2005—or to repeat in 2006. Nor did any of them foresee Steph Curry’s first MVP win, in 2015, or Giannis Antetokounmpo’s first, in 2019. They also whiffed on Derrick Rose’s 2011 triumph and Jokic’s repeat in 2022. (Only a single manager had predicted Jokic’s first win the prior year.)

Put another way, in the past two decades, NBA GMs have missed entirely on six MVP seasons while correctly predicting only four as a group.

Splitting the GMs’ MVP prediction history into votes for LeBron and votes for other players makes their track record look even worse. At least 25 percent of managers picked LeBron to win the award in 12 separate seasons, and he won four trophies. But at least 25 percent of GMs have picked another player to win the award 21 times—and that group has hit only once, when Giannis won his second MVP in 2019-20. That’s a success rate below 5 percent.

That chart is bad news for Jokic, who received 43 percent of the votes in the 2023-24 survey, as general managers have finally started picking him with greater frequency after giving him scant attention prior to his actual MVP-winning seasons. The reigning Finals MVP might be the consensus best player now—but based on GMs’ track record with this question, MVP bettors would be wise to pick a longer shot this season.

Rookies

General managers do much better picking the Rookie of the Year winner than they do the MVP. In 15 of 21 years, their first or second choice for Rookie of the Year has gone on to win the award.

That rosy-looking chart hides some massive misses, to be fair. Eighty percent of GMs selected Victor Oladipo to win Rookie of the Year in his first season, but he didn’t. Neither did Anthony Davis (77 percent), Jabari Parker (75 percent), Zion Williamson (68 percent), John Wall (68 percent), or Lonzo Ball (62 percent).

What do all of those players have in common? They were all top-two picks in their respective drafts, so it’s hard to blame general managers for their confidence in those players—many of whom developed into stars.

It’s logical that GMs would be more accurate at forecasting Rookie of the Year because top picks tend to win. The four complete whiffs from that chart include Scottie Barnes (no. 4 pick in 2021), Michael Carter-Williams (no. 11 in 2013), Malcolm Brogdon (no. 36 in 2016), and, strangely, Derrick Rose, who didn’t receive a single vote to win the 2008-09 award even though he’d gone no. 1 in the draft.

(On the other hand, give plaudits to the 44 percent of general managers who looked further down the 2006 draft board to predict that no. 6 pick Brandon Roy would win Rookie of the Year. He was GMs’ top choice that year, and that faith paid off.)

General managers have a similarly mixed track record when given more subjective questions about the most recent draft class. Asked for the biggest steal of the draft, for instance, they’ve correctly tabbed Tyrese Haliburton (43 percent) and Shai Gilgeous-Alexander (27 percent) in recent seasons, which perhaps portends future stardom for the Rockets’ Cam Whitmore, the no. 20 pick and GMs’ consensus steal of the 2023 draft (43 percent).

Of course, a similar proportion of GMs also said the same about Norris Cole (44 percent), Dennis Smith Jr. (37 percent), Nickeil Alexander-Walker (33 percent), Justise Winslow (31 percent), Rodney Stuckey (30 percent), and Kelly Olynyk (28 percent) in their respective draft seasons—so take the “biggest steal” title with a hefty pinch of salt. It’s not as if any GMs correctly identified Jokic as his draft’s best sleeper, or else they would’ve picked the future two-time MVP themselves instead of letting him fall to Denver at no. 41.

The last question worth noting in this analysis has been a staple of the annual survey since its second year: “Which rookie will be the best player in five years?” The choice was clear this season: Wembanyama received 90 percent of the vote, the highest total for any rookie in the survey’s history. Players who received at least 50 percent in this vote have almost always matured into superstars.

Here, too, general managers naturally tend to default to the no. 1 pick. When they don’t, the answers can get weird. Evan Mobley edged out Cade Cunningham in 2021, which seems understandable two years out. James Wiseman beat Anthony Edwards (and LaMelo Ball) in 2020, which … doesn’t. Josh Jackson was the most common answer in 2017, narrowly ahead of Markelle Fultz and Jayson Tatum.

But in retrospect, the funniest answer to this question—indeed, the single funniest answer I found from poring over thousands of GM responses from more than 20 years of questions—comes from the second survey, before the 2003-04 season.

That year, 81 percent chose no. 1 pick LeBron James as the rookie who’d be the best five years later. Another 10 percent picked Dwyane Wade, and 5 percent picked Carmelo Anthony—both respectable choices for future Hall of Famers. But one other player received a vote for that question—an assertion that in five years, he’d be better than LeBron, better than Wade, better than Melo and Chris Bosh, too. That player was Mickael Pietrus.

So as you prepare for the 2023-24 season, and prepare to watch and read about the NBA for the next nine months, don’t blame broadcasters or analysts if you think they’re wrong in their evaluation of a player. Sometimes, GMs have wild takes too.