As the Cleveland Cavaliers visit Boston on Sunday without former Celtics Isaiah Thomas and Jae Crowder, both dealt before Thursday's trade deadline, I think it's worth investigating how the Kyrie Irving trade went so wrong in Cleveland. Crowder and Thomas, the veterans acquired for Irving, along with rookie Ante Zizic and the Brooklyn Nets' 2018 first-round pick, lasted less than six unhappy months.
As someone who was on team "this is a pretty good return for Kyrie!" 6 months ago, I have another mailbag question: Has any NBA player fallen faster in perceived value than IT? ...From 5th in the MVP voting to a guy who was essentially treated as having zero trade value.
- Nate Silver (@NateSilver538) February 10, 2018
I, too, believed the package was a good return for Irving, giving the Cavaliers a much better grade on the deal (A) than the Celtics (C) initially. Those evened up to some extent when I regraded after we learned Cleveland had greater concern about the condition of Thomas' injured hip, but the Cavaliers still came out ahead (A-minus vs. B-minus).
So where did I go wrong? What happened to Crowder and Thomas?
Hip injury limited Thomas' play
A story by my former ESPN colleague Tom Haberstroh after the trade wisely raised the concern that other guards such as Michael Carter-Williams and Jonny Flynn had been less effective after hip labrum tears like Thomas' injury.
Even compared to those cautionary examples, Thomas has been disappointing. His box plus-minus (BPM) metric on Basketball-Reference.com (minus-5.9 rating per 100 possessions, putting him among the league's bottom-10 regulars so far) is worse than Carter-Williams was last season after his hip surgery (minus-2.4). Thomas has been better than Flynn was in 2010-11 (minus-7.2), but then, Flynn was coming off an up-and-down rookie campaign, not one in which he made the All-NBA second team.
In fact, if we switch to using player win percentage (the per-minute version of my wins above replacement player metric akin to BPM or PER), Thomas' decline is almost without precedent in modern NBA history. Last season, his .698 win percentage ranked 14th in the league, just behind Anthony Davis. That translated into more than 15 WARP. So far in 2017-18, Thomas' .395 win percentage puts him slightly below replacement level. Granting that Thomas' season isn't done yet, only one player since the ABA-NBA merger has seen his win percentage decline more from one year to the next.
None of the other players on this list had anywhere near Thomas' pedigree before their fall. World B. Free, a high-scoring and colorful one-time All-Star, might come the closest. But he was already 32 by the time the Cleveland Cavaliers let him go as a free agent as part of a youth movement in 1986-87. Free was unsigned until joining the Philadelphia 76ers in late December 1986 and played 20 miserable games that season, his second-to-last in the NBA.
Odom is probably the best comparison in terms of trade value. The Lakers got a first-round pick when they sent an unhappy Odom to the Dallas Mavericks days after a package sending him and other players for Chris Paul was rejected by the league. Joining the defending NBA champions the year after winning the Sixth Man Award, Odom shot 35.2 percent from the field. He, too, played only one more season in the NBA.
Rapid declines in trade value
Setting this list aside, a few other examples come to mind. The Detroit Pistons inexplicably gave up Chauncey Billups and Antonio McDyess (plus little-used Cheikh Samb) for 33-year-old Allen Iverson early in the 2008-09 season. While Billups helped the Denver Nuggets to the Western Conference finals, Iverson saw his true shooting percentage decline from .567 to .504 as the Pistons were swept in the opening round after six consecutive trips to the conference finals.
Unsigned until September, Iverson signed a one-year deal worth $3.1 million with the Memphis Grizzlies. He lasted three whole games in Memphis before the two sides agreed to part ways. Iverson was waived later in the season after returning to the 76ers, ending his Hall of Fame career.
Andrew Bynum is another interesting case. Philadelphia gave up an incredible haul to get Bynum as part of the four-team Dwight Howard trade: star Andre Iguodala, recent first-round picks Maurice Harkless and Nikola Vucevic, plus a future first-rounder. (The Sixers also took on the onerous contract of Jason Richardson in the process.) Yet Bynum never played a game in Philadelphia due to a knee injury, amid reports that the injury might have been exacerbated by bowling.
Nonetheless, as a free agent the following summer, Bynum was still able to command an incentive-laden, two-year deal worth up to $12 million annually from the Cavaliers. It wasn't until Bynum washed out in Cleveland that his market value went to nil. (Like Thomas, Bynum was traded as part of a cap-clearing move.)
The other names readers submitted, including Thomas' former Cavaliers teammate Derrick Rose, Gilbert Arenas and Penny Hardaway, all saw their value erode more gradually due to repeated injuries. What makes Thomas' change from fringe MVP candidate to expiring contract so shocking is just how quickly it happened. One day, you're getting traded for Irving. The next day, Jordan Clarkson.
Crowder disappointed in Cleveland, too
The magnitude of Thomas' decline helped obscure the fact that his fellow former Celtic also struggled with the Cavaliers. Crowder, who had provided valuable 3-and-D play in Boston, offered neither skill on a consistent basis in Cleveland.
I've gotten a few mailbag questions about whether Crowder's decline is evidence that ESPN's real plus-minus (RPM) doesn't predict well when players change teams, given he ranked 20th in the league in RPM last season. That's true, to a point -- RPM is less predictive when players change teams, or roles within the same team -- but in this case, no advanced stat would have seen Crowder's drop-off coming. Compare his RPM and his BPM over the course of his career.
While Crowder has generally rated better by RPM, which factors in lineup data in addition to the box score prior that is similar to BPM, he rated well in both metrics with the Celtics, and his decline with the Cavaliers has been nearly as pronounced. By any measure, Crowder has been a worse player in Cleveland, worse even than during his two-plus seasons with the Dallas Mavericks out of college.
Though not historic like Thomas' underperformance, Crowder's is notable in its own right. My SCHOENE projection system uses the development of the 50 most similar players at the same age to forecast a range of possible outcomes the following season in terms of player win percentage. Crowder's mean projection using this method was for a .533 win percentage, slightly better than league average (.500, naturally). His actual .388 win percentage so far would be a larger decline than any of those 50 comparable players experienced.
There isn't an injury to explain why Crowder has been so ineffective with the Cavaliers. One possible explanation is that Crowder wasn't in shape to start the season after spending the time between the trade and training camp dealing with his mother's death from cancer, which happened the same day he was traded.
After struggling with his shot early in the season, Crowder has made 35.0 percent of his 3-point attempts since the start of 2018, a little better than his career mark of 34.4 percent. Yet Crowder's defensive stats have not bounced back. His steal, block and defensive-rebound rates are all the lowest of his career. It's possible Crowder is no longer the same athlete he was in his early 20s. His steal rate first declined last year, and David Locke of Locked on Jazz pointed out that Crowded doesn't have a single dunk this season. (Crowder has never been a prolific dunker but had 13 in 2016-17, per Basketball-Reference.com.)
On Friday's post-deadline episode of the Lowe Post podcast, ESPN's Brian Windhorst offered an alternative theory that Crowder's defensive effort might have suffered because he was unhappy with his limited opportunities in the Cleveland offense. Indeed, in addition to using plays at the lowest rate (15.1 percent of the time) since 2013-14, Crowder also saw his frontcourt touches per minute and average seconds per touch decline according to Second Spectrum data on NBA Advanced Stats. As a result, Crowder went from handling the ball 4.4 percent of the time he spent on the court last season in Boston to 3.6 percent so far this year.
If Windhorst's theory is true, Crowder should be a good fit for the Utah Jazz, whose offense is predicated on sharing the ball. Every non-center on the Utah roster has handled the ball at least 4.4 percent of the time they've spent on the court this season.
Some of the Cavaliers' underperformance was predictable, but much was not
In reviewing my analysis of the trade from Cleveland's side, I see two big mistakes. First, I underestimated the possibility that Thomas would return from his hip injury a different player, focusing in my revised trade grade mostly on the time he would likely miss during the regular season -- not as important for a team like the Cavaliers, who are relatively unconcerned with seeding.
Second, in playing up Crowder's value in the trade, I screwed up in playing up that his RPM was better than that of Irving and Thomas, which made it sound like I thought he was the best player in the trade. As valuable as his 3-and-D play was for the Celtics, Crowder is not in their class as a player.
Taking those facts into account, I should have had more questions about whether Cleveland might drop off without Irving. Still, even the biggest skeptics of the deal never predicted a scenario in which both Crowder and Thomas offered the Cavaliers sub-replacement performance.
It's easy to fall victim to hindsight bias and believe the signs were there all along that Crowder and Thomas would disappoint. But those were balanced and outweighed by the indications that Cleveland had gotten two good players for its disgruntled superstar. The Cavaliers made a reasonable trade but got burned anyway.