clock menu more-arrow no yes

Filed under:

FiveThirtyEight projects the present and future of Zion, Lonzo, Ingram, and the 2021 Pelicans

New, comments

A guide to the newest RAPTOR projections from the data journalism giant

NBA: Memphis Grizzlies at New Orleans Pelicans
The present and future stars of the Pelicans franchise share a laugh
Derick E. Hingle-USA TODAY Sports

As the old saying goes, there’s nothing quite like an influx of regular season NBA data.

The Nets and Clippers kicked off the truncated 2021 NBA regular season with victories last night. Alas, only so much can be gleamed from a single night of games involving a fraction of the league’s teams. There are other, more effective ways than parsing extremely limited samples to try and predict what’s going to happen the rest of the way. NBA teams themselves likely do some version of this behind closed doors, but fortunately, not all of these efforts have been private.

FiveThirtyEight’s individual and team performance projections for NBA basketball were first published during the 2016 season and have evolved in the intervening years. I took a three-part, Pelicans-centric dive into the depths of their data last July to preview the coming year, but by mid-October they had revamped their entire system to revolve around a new statistic called RAPTOR. In preparation for the truncated season set to start this week, Neil Paine and the rest of the FiveThirtyEight team published their updates to RAPTOR.

With less than 24 hours until the Pelicans and Raptors battle it out in their season openers, my explanation of RAPTOR and subsequent analysis of the New Orleans projections should age better than last year’s attempt. If it doesn’t, I’ll just hope that it doesn’t take another 536 tries for me to find success.

What is RAPTOR?

RAPTOR was built to quantify and project all on-court contributions with a single statistic. FiveThirtyEight explains it in as much detail as you could ask for, but for the sake of this article, I’ll lay out some essentials and tips for exploration.

RAPTOR stands for Robust Algorithm (using) Player Tracking (and) On/Off Ratings, which is a fun way to commemorate the 2019 NBA champions that were predicted to defeat the Warriors by the site’s since-retired CARMELO projections while simultaneously identifying key components of the stat. Box score production and player tracking data for each player is paired with a variety of on-off calculations (with the former being weighted more than the latter) to determine how many points per 100 possessions a player is worth to their team on both offense and defense. For example, RAPTOR pegged Zion Williamson’s offensive production from last season at +1.2 and his defensive production at +0.2 for a net total of 1.4 points added per 100 possessions.

Like all plus-minus statistics, RAPTOR’s operatic assumption is that player performance is largely linear and additive. In other words, ratings from individuals can simply be added to approximate team performance. Factors like coaching, systems, and player synergies are not accounted for in these calculations. This is not to say that these things aren’t important or impactful, but rather that there are inherent limitations to such a model. Exclusions of this sort are often enough to dissuade some fans from the legitimacy of the stat, but the FiveThirtyEight team incorporated plenty of basketball logic to support the figures in their algorithms.

The net RAPTOR total is useful on its own, but it doesn’t account for playing time. That’s where WAR (Wins Above Replacement) comes in. A player’s RAPTOR is multiplied by their minutes played (plus some other factors) to determine how many more wins a player is worth than a replacement-level player. The RAPTOR for replacement-level players is -2.75 points per 100 possessions, which is derived from the historical performance of players on two-way contracts, who operate on the fringes of NBA-caliber play. By way of another example, though Williamson and Donovan Mitchell both had RAPTORs of +1.4, Mitchell’s WAR after 2,628 minutes of regular-season play finished at 5.7, whereas Williamson only totaled 1.4 in 668 minutes.

With the help of a dataset dating back to the 2014 season, when the NBA began making some player tracking data public, past performance can be extrapolated to project future performance. This data helps map the trajectories of past players with particular skill sets at certain points in their careers, which is then used to project the futures of similar players. The most similar players are listed on each player’s page. Let’s continue to use Williamson as an example and check out his player page:

Zion Williamson’s FiveThirtyEight RAPTOR projections

Williamson’s exceptional skill set is so much so that his most comparable player is 2020 Luka Doncic, whose similarity is only in the 16th percentile. By these calculations, those who have deemed Williamson a player without precedent appear to be correct (in what other context would). Speaking of percentiles, the charts on the left side of the screen describe his performance in each stat relative to the rest of the league (for those unfamiliar with percentiles, a player in the 25th percentile is bested by 75 percent of the league, where are a player in the 90th percentile is bested by only 10 percent.)

The chart on the top right contains Williamson’s WAR projections. The black dots above each season (‘20 represents the actual 2020 season total) are the official WAR projection for that year. The gray area both above and below these dots cover the middle 80 percent of likely outcomes, ranging from the tenth percentile at the bottom to the 90th percentile at the top. In the context of the coming season, Williamson’s most likely WAR total is 4.5, and totals that deviate from the dot in either direction are less and less likely. Ultimately, he is given equal ten percent chances of topping the peak of his range (roughly 10 WAR, which would have been good for 12th-best in the league last season) and falling below the bottom (roughly 2 WAR, which would have ranked 150th). Finally, the labels atop the chart categorize Williamson (“Future All-Star”) and place a relative dollar value ($118.3 million) on the 28 WAR he is projected to accumulate over the next five seasons.

What does RAPTOR think of the 2021 Pelicans roster?

Having established what RAPTOR is and how it works, let’s turn to its projection of the 2021 Pelicans. RAPTOR has a lot to say about present and future alike, but with the season set to start in Tampa Bay tonight against the Toronto Raptors (a correct sentence just weird enough to communicate the insanity of playing basketball in 2020), I’ll start with projected individual performances for 2021. The table below includes 14 players on the roster (as of December 23) accompanied by their projected and actual WAR totals from the 2020 season (when applicable) and their projected WAR total for the 2021 season, listed from highest to lowest in projected 2021 WAR:

Projected Single-Season WAR: Pre-Actual 2020 vs. Pre-2021

Player Proj. 2020 WAR Actual 2020 WAR Proj. 2021 WAR
Player Proj. 2020 WAR Actual 2020 WAR Proj. 2021 WAR
L. Ball 5.2 2.6 6.1
E. Bledsoe 7.5 5.3 5.2
B. Ingram 1.5 4 5
Z. Williamson 3.5 1.6 4.5
S. Adams 6.1 4.7 4.1
J. Hart 3.2 3 3.8
N. Alexander-Walker 0.6 -0.2 1.2
JJ Redick 2.6 1.9 1
T. Carr 0.5 N/A 0.9
N. Melli N/A -0.2 0.7
W. Gabriel 0 0.4 0.5
J. Hayes -0.3 -0.3 0.3
K. Lewis Jr. N/A N/A 0.1
W. Hernangomez 0.2 0.2 0
TOTAL 30.6 23 33.4
Data via FiveThirtyEight

Lonzo Ball and Eric Bledsoe top this list, but what is probably even more apparent is that both Williamson and Brandon Ingram do not. Ingram’s 5 WAR projection in the first year of his maximum contract extension may seem modest relative to his teammates, but it’s more than triple his 2020 projection (1 WAR) and would actually be a full win higher than his career-high output last year. Projection systems like RAPTOR have a long-term memory; the system was very low on Ingram’s track record just a season ago, and because performance from years past is still part of the picture, expectations are a bit tempered. Even so, if Ingram is able to prove this year that the massive strides from his breakout season can be sustained, or better yet, improved upon, he could easily rank first on this list heading into 2022.

Ball’s story is similar, but with a different trajectory. The leap that Ingram made between 2019 (0.7 WAR) and 2020 (4.0 WAR) is a tenth of a win from being exactly what RAPTOR expected for the young point guard (1.8 WAR in 2019, projected 5.2 WAR in 2020). He finds himself atop this list despite his much more modest year-over-year improvement (+0.8 WAR) because, like with Ingram, RAPTOR will not overreact to a single season from a young player. Even though his projection for 2021 (6.2 WAR) is even higher than the one he failed to meet last season, his ceiling and five-year value fell a bit.

Before getting to long-term value, the dollar version of the above WAR projections will help to further illuminate what may be in store for the coming year (listed from highest to lowest in projected 2021 value):

Projected Single-Season Value: Pre-/Actual 2020 vs. Pre-2021

Player Proj. 2020 $ Actual 2020 $ Proj. 2021 $ 2021 Contract $
Player Proj. 2020 $ Actual 2020 $ Proj. 2021 $ 2021 Contract $
L. Ball $21.7M $9.6M $24.5M $11.0M
E. Bledsoe $32.6M $20.9M $20.2M $16.9M
B. Ingram $6.8M $15.2M $19.9M $27.3M
Z. Williamson $14.6M $6.3M $17.3M $10.2M
S. Adams $25.7M $18.1M $15.8M $27.5M
J. Hart $13.3M $11.3M $14.9M $3.5M
N. Alexander-Walker $3.7M $1.3M $5.4M $3.1M
JJ Redick $10.9M $7.3M $4.8M $13.0M
T. Carr $3.2M N/A $4.5M $0.9M
N. Melli N/A $1.3M $3.9M $3.9M
W. Gabriel $1.6M $3.1M $3.1M $1.6M
J. Hayes $0.8M $0.9M $2.5M $5.1M
K. Lewis Jr. N/A N/A $2.1M $3.6M
W. Hernangomez $2.4M $2.1M $1.7M $1.7M
TOTAL $137.3M 97.4M $140.6M $140.6M
Data via FiveThirtyEight

Projections are less kind to veteran players who underperform expectations, and Bledsoe and Steven Adams are no exception. Both players missed their lofty projections by significant margins (approximately 29 percent in Bledsoe’s case, 23 percent for Adams), and while their 2021 projections are relatively close to what they did last year, it is less likely (though not impossible) that they, at 31 and 27 years old respectively, will be able to reorient their career trajectories towards improvement without drastic changes to their games.

Their combined expectation of $36 million isn’t quite as valuable as the $34.1 million that Jrue Holiday could provide the Milwaukee Bucks with a single roster spot, but in the context of a team looking for average-or-better complementary pieces, they are in line to be real difference makers for Stan Van Gundy’s first Pelicans team.

JJ Redick is the lone player set to earn double-digit millions without reaching that mark in value added. It’s no coincidence that he is also the oldest player on the roster. The quality of his production hardly wavered, and even improved in some respects (his 3-point percentage leaped back up to 45.3 percent from a three-year low of 39.7 percent), but history has not been kind to players on the back end of their thirties, which is the primary reason for the low projection.

The shortened regular season artificially reduced his overall output in 2020, and the reputation that Redick has established for himself throughout his career is reason to believe he won’t give way to Father Time without a fight. His ability to fulfill the spacing and shot-making role that he was signed for hasn’t waned, but his minutes may continue to. In short, RAPTOR thinks the Pelicans may be slightly overpaying Redick for what he’ll provide, but that doesn’t mean New Orleans is at fault for employing him.

In terms of outperforming their contract, Josh Hart bested all of the Pelicans last year and is projected to do the same this time around. He made $1.9 million last year and contributed $11.3 million (or 3 WAR) in on-court production, which equates to a return on investment of 595 percent. That’s very good! That’s basically the equivalent of paying $2 at McDonalds and getting 50 chicken nuggets, but you get 13.5 points and 8.7 rebounds per 36 minutes and an occasional GIF.

Many Pelicans supporters agree with RAPTOR’s assessment that Hart is due for a pay day, but the franchise hasn’t given him one just yet. His future in New Orleans after 2021 is clouded, but for a player who has been as consistent as the 24-year-old Hart, the excess value he can and likely will provide is remarkably clear.

Hart was a strong player from the moment he entered the league, and while he’s young enough that a level-up is far from out of the question, the steadiness of his statistical production has capped his long-term value. Speaking of which, here are the projected five-year values of the players listed above, ranked from the highest 2021 projection to the lowest:

Projected 5-Year Value: Pre-2020 vs. Pre-2021

Player 2020 Proj. 5-Yr. $ 2021 Proj. 5-Yr $ 5-Yr. Proj. $ Change
Player 2020 Proj. 5-Yr. $ 2021 Proj. 5-Yr $ 5-Yr. Proj. $ Change
L. Ball $155.2M $130.4M $-24.8M
Z. Williamson $144.2M $118.3M $-25.9M
B. Ingram $48.0M $102.0M $54.0M
J. Hart $70.4M $67.0M $-3.4M
S. Adams $103.4M $63.2M $-40.2M
E. Bledsoe $114.5M $59.4M $-55.1M
K. Lewis Jr. N/A $47.9M N/A
J. Hayes $23.0M $42.1M $19.1M
N. Alexander-Walker $48.8M $38.2M $-10.6M
T. Carr $13.5M $38.1M $24.6M
W. Gabriel $9.2M $18.7M $9.5M
N. Melli N/A $13.2M N/A
JJ Redick $24.3M $10.7M $-13.6M
W. Hernangomez $11.9M $8.1M $-3.8M
TOTAL $766.4M $757.3M $-70.2M
Data via FiveThirtyEight

A lot of what has been discussed above applies to this list as well. The underperformances of Bledsoe and Adams hurt but did not entirely flatten their long-term expectations, Ball and Ingram, despite wildly different 2020 seasons, both find themselves ahead of the $20 million of value per year threshold, and Hart could very well just continue to be Hart until the back half of the decade. Redick finds himself among end-of-the-bench players like Wenyen Gabriel, Nicolo Melli, and Willy Hernangomez, but it wouldn’t be too much of a surprise if he winds up being the most productive of this group in 2025, even as a 41-year-old. Rookie Kira Lewis Jr. barely edges out the second-year pair of Jaxson Hayes and Nickeil Alexander-Walker as well as surprise guest Tony Carr, whose stats from his college days at Penn State apparently grade quite well in RAPTOR’s book.

But what about Williamson?

No individual player’s future matters more to the future of the franchise, and given what the world saw from the number one pick in the 2019 NBA Draft, that bodes will for the Pelicans. Heading into his rookie season, Williamson’s five-year value was projected to be $144.2M, or $28.8M per season, figures that any player in the league would be thrilled to reach, much less one who presumably wouldn’t even have reached their peak by the time that period ended in 2024.

Preseason surgery delayed his regular season debut and limited him for the entire season, which itself was limited by the COVID-19 pandemic. That Williamson was able to impress as much as he did despite this is why his reduced five-year value of $118.3M should be received with optimism. Struggles on defense should cap anyone’s potential, and RAPTOR acted accordingly for both the present and the future after what was an uneven rookie showing on that end. Williamson’s 90th-percentile WAR projections through 2024 capped out at about 16 WAR before last year, but sit at just over 10 as of December 23.

Even with those struggles on the books, though, he’s still projected to be worth a respectable $17.3M in his sophomore campaign. However, given that much of Williamson’s domination at Duke was a product of his defensive prowess, there is reason to believe that a healthy and unrestricted second season can put him back on the fast track to being considered among the best players in the league.

How far can the 2021 Pelicans go?

A breakout season from a fully-operational Williamson would accelerate the Pelicans contention timeline in a big way, but even a tiny boost could be enough to vault the team back into the postseason. As of December 23, New Orleans finds themselves on the outside of RAPTOR’s top eight with eyes on the play-in tournament:

FiveThirtyEight’s projected standings for the 2021 NBA season

All of the RAPTOR data for individual players described throughout this article is put together for each team and, with the help of minutes projections and ELO ratings, regular season records and Finals chances are projected. (Each team has a page where all of this data is visible, and the Pelicans page can be found here.) These exact figures I won’t explain the nuts and bolts of the models at work here, but FiveThirtyEight did so to great effect if you’re interested.

Worthy of note is that the ninth-seeded Pelicans are projected to be just five games worse than the fifth-seeded Suns. The bottom half of the West is shaping up to once again be a free-for-all that will be tracked until the season’s final day. Fortunately, as games get played and injuries occur each day, these projections adjust and offer updated data, which makes them a fun tool to use throughout the course of the season.

There is no one way to enjoy basketball. The increased prevalence of Ivy Leaguers and advanced statistics in front offices and fandom alike has often been accompanied by a stench of superiority that justifies the handwaving it often receives. My goal here is not to shame anyone for the way they choose to experience the game, but to try and provide tools and tips that could help craft a new lens that adds to their fan experience. RAPTOR isn’t a perfect stat, as evidenced by the constant changes it undergoes. It’s merely trying to figure out how to understand this strange game.

Just like us fans.


*For those curious, the complete WAR calculation is the following:

WAR = (RAPTOR+2.75) × Minutes Played × ((LeaguePace+Individual Pace Impact)/League Pace) × WAR multiplier

The WAR multiplier is derived from Bill JamesPythagorean expectation

For more Pelicans talk, subscribe to The Bird Calls podcast feed on iTunes, Spotify, Stitcher or Google Podcasts. You can follow this author on Twitter at @NolaYanks2740.