Why NBArank is tough on Melo, Kobe and DeRozan

Why do players like Carmelo Anthony finish lower than some fans expect in NBArank? Isaiah J. Downing/USA TODAY Sports

NBArank is a forecast, and one year ago, ESPN's NBArank panel predicted which players would perform the best in 2017-18. LeBron James won, but the player who drew the most attention was ranked No. 64: Carmelo Anthony.

Melo responded immediately to the apparent insult, and he found a chorus of players and media voices in support, including some at ESPN. The loud consensus was that the ranking was ridiculous and indicated an embarrassing lack of basketball knowledge among the panel.

Within months, however, the thunderstorm of criticism was directly over Melo himself -- Stephen A. Smith said during the playoffs, "I cannot believe what I am seeing. He looks bad. He looks bad. He looks incapable of doing much. ... Carmelo Anthony looks done."

How this happened is what makes NBArank unique. And the results have created controversy when popular players like Anthony, Kobe Bryant and DeMar DeRozan are ranked lower than fans expect.

Melo: When fan favorites fall

One primary accusation against the NBArank panel every year is that it lacks "respect." Melo himself said it best in his response:

But respect is not a factor in NBArank. In truth, NBArank is designed to take respect out of the equation as much as possible.

NBArank isn't a lifetime achievement award or a celebration. We have plenty of hugs and high-fives in the NBA. Rather, it is supposed to be an objective assessment -- a way to transmit information about what fans should expect in the coming season.

So we made NBArank a prediction. That decision was inspired by social science, and particularly the finding by the The Brookings Institution that one of the most accurate ways to forecast a presidential election is to ask a diverse bunch of people to predict who they think will win. It's the wisdom of the crowd in action. It sounds simple, and it is.

We have a large and diverse group of voters, from across ESPN, The Undefeated and FiveThirtyEight, including analysts, reporters, editors, broadcasters, producers, researchers and other close NBA observers. This year we had nearly 12,000 votes -- with two players listed each time, that's an average of more than 150 predictions on each player, with lots of matchups of superstar vs. superstar.

On top of that, three years ago we changed the ballot to make NBArank even more accurate. In consultation with polling experts and analysts at Microsoft Research and the Wharton School, we now use a pairwise ballot, which asks voters to choose between two players. Last year, we asked, "Which player will be better in 2017-18?"

This style of ballot asks voters to make a real prediction between two players and provides a more fair and even-handed result.

And last year, we made another change to boost accuracy -- we started giving voters the option of seeing each player's projected impact on winning, according to real plus-minus (RPM). Voters can choose to ignore this information, or they can choose not to see it at all. But the idea is to help the ESPN panel move on from the past and look more objectively at the coming season. It increases the focus on what matters in our predictions: which players will do the most -- in terms of both quality and quantity -- to help their teams win.

Last year, some voters had Melo winning a lot of his matchups. But other voters looked at the decline in his effectiveness -- which was captured by his projected RPM -- and decided that the 2017-18 version of Melo might not be the force he had been in his previous 14 seasons. Not only was Anthony a below-average defender, but his offensive efficiency had dropped for four straight seasons.

Carmelo Anthony is one of the most popular players in NBA history, and he is going into the Hall of Fame, probably the first year he's eligible. He has been a prolific scorer, he led Syracuse to an NCAA championship in 2003, and he has been acclaimed as perhaps the most accomplished player in Team USA's glorious international history.

But that's history. And it said nothing about what he would do in 2017-18, and that's what we asked the panel to predict.

And this is how Melo landed at No. 64.

Kobe: When great players get injured

The case of Kobe Bryant is more complex. And it's a reminder about how destructive injuries can be, even to the greatest players.

Kobe and NBArank added up to a contentious topic even before his devastating Achilles tendon injury of April 2013. Although he had always landed in the top 10, Bryant had scoffed openly at his ranking, particularly after being ranked No. 7 in 2011. And his supporters -- fans and many media members -- generally had his back.

But it was the 2013 rankings that put the Kobe cult on tilt.

Before the 2013 poll, we made an important change: NBArank became a prediction, rather than just a rating. As mentioned, this was to eliminate "respect" from the equation and to focus on what the player could provide in the coming season.

Of course, fans (and many of our panelists) weren't ready to hear that Bryant's Achilles injury had likely curtailed his career, or at least his ability to contribute. Kobe had a well-earned reputation for grit and hard work, and to many, his injury seemed like a pit stop on his way back to greatness. This was a guy who tore his Achilles and then made two free throws before he walked off the floor.

So our panelists, despite knowing how difficult it might be for Kobe to return to health, and despite our voting system minimizing the impact of "respect," did extend Bryant a large measure of respect and put him in the top 25 -- a sort of compromise between crediting his previous all-world talents and assessing his new, more limited powers. Of course, that's not how Kobe and his supporters saw it -- at all. The public outcry was deafening, and Kobe's response was creative: He changed his Twitter avatar to read "1225," after ESPN had forecasted the Lakers to finish 12th in the Western Conference in 2013-14 and Bryant to be the 25th-best player.

He went on to play six games that season and subsequently ranked 40th and 93rd before his final two seasons, reminding folks in 2014 about his disdain for the ESPN panel: "I've known for a long time that they're a bunch of idiots."

As the architect of NBArank, I have always found these injury-related outcomes both fascinating and frustrating. While the conversation can be stimulating, it's no fun announcing that we think a player will be a shadow of his former self, or that fans shouldn't expect a miracle.

Assessing players who are dealing with injuries is the hardest part of NBArank, and the most unfair in some ways -- to both the players and the panelists. There is a lot of uncertainty on all sides. We've been through it with Kobe, Derrick Rose, Deron Williams, Rajon Rondo, Dwight Howard and many others. We want to be wrong, and we want to give them the benefit of the doubt, even as we try hard to get it right.

This year, we have two former top players trying to return from serious injuries: Kristaps Porzingis and DeMarcus Cousins. Like Kobe, Cousins is dealing with an Achilles tear. For that reason, he has dropped 52 spots this season. Porzingis has slid 36 spots because he's expected to miss part of the season with a torn ACL. Let's hope next year we're talking once again about whether they're top-20 players.

DeMar: When big scorers don't play D

DeMar DeRozan is a skilled scorer who has put up as many as 27.3 points per game. And because scorers and shot creators are valued so highly in the NBA, it's no surprise that DeRozan is a four-time All-Star.

DeRozan's ranking has reflected his positive contributions, as he has finished with an average ranking of 39 from 2014 to 2018. And that's exactly where he sits this year, just like last year: No. 39.

But that ranking doesn't square perfectly with his stardom, so why is that?

To put it bluntly, DeRozan's advanced stats don't match up to his reputation, for two reasons. The minor reason is that his offensive production -- his midrange game -- has not been especially efficient. The major reason is that he has been a disaster on defense. According to RPM, he is consistently one of the worst defenders at shooting guard and plays D at a level that costs his team dearly. The Raptors often played better with DeRozan sitting than with him on the floor.

Our voters predicted -- by a slender margin -- that several two-way players will have a bigger impact this season on winning. That list includes Celtics wing Jaylen Brown, Wizards forward Otto Porter Jr. and Bucks wing Khris Middleton, all of whom are known for providing relatively efficient offense and sturdy defense. Each player easily outpaced DeRozan in projected RPM as well.

Of course, DeRozan is not the only one-way player near the top of the food chain. At various times, that could describe James Harden, Kyrie Irving and many other stars. But when we pair DeRozan's lack of defense with his so-so offensive efficiency, and mix in his well-documented struggles in the postseason, we end up with a player who contributes less to winning than it might appear. That's something the voters take seriously.

Could our prediction on DeRozan be wrong? Yes, indeed. He's still just 29, and his move to the Spurs offers him an opportunity to prove himself in a new system.

For inspiration, he can look right down the road. In 2012, before NBArank was a prediction, we ranked Harden at No. 26, as he was a sixth man coming off a miserable NBA Finals performance for the Oklahoma City Thunder. But shortly after our vote, Harden was traded to Houston. When the Rockets gave him the keys to the offense, he soared into the MVP conversation and has resided there ever since. DeRozan will have his chance to do likewise in San Antonio.

NBArank is intended to inform, not confirm, the public consensus. It's not about popularity or parroting conventional wisdom. It's about each player's contributions to winning, whether he has great teammates or not. It's about looking ahead with the question: Which player will be better in 2018-19?

NBArank is conducted in collaboration with Etan Green of The Wharton School and David Rothschild of Microsoft Research.