I read the KP CAREMLO article some time ago. It's pretty interesting stuff. KP is so unusual that you can't accurately compare or predict what he will be.
If those assumptions hold, the Knicks have a hell of a prospect on their hands. The NBA is a tough league for rookies, so merely staying in a team’s rotation at age 20 is often a sign of a bright future. So far, however, Porzingis has not just been a rotation player but an above-average one4 — usually a sign of superstar potential. In fact, Porzingis’s long-term upside score, based on his wins above replacement projection from 2016-17 through 2021-22, is 36.3. That’s very good; at the start of this season, it would have made Porzingis the 17th-most-valuable franchise player in the league, in the same vicinity as Giannis Antetokounmpo, John Wall and Jimmy Butler.But Porzingis’s projection also involves tremendous uncertainty. In 2018-19, for example, his 90th percentile projection5 is 13.3 WAR, good enough to put him on the fringes of the MVP discussion. Meanwhile, his 10th percentile projection is just 0.0 WAR (exactly replacement level), or roughly the same range as Quincy Acy. The error bars around CARMELO’s forecasts for young players are often wide, but these are especially so.
OK, but is he Dirk?
What gives? Part of it is the problem we alluded to before. CARMELO works by identifying comparable players, and Porzingis is a hard guy for which to find historical precedents. Only one player, Brook Lopez, achieves a similarity score of 50 or higher with Porzingis6 — and if we’re being frank, the Lopez-Porzingis connection doesn’t have the “eye test” appeal that CARMELO comparisons often do.
But faced with unusual players like Porzingis, CARMELO has to make some sacrifices. In the case of Lopez, it ignores the fact that Lopez almost never shoots from behind the arc (although he does have a decent midrange game). In the case of Kevin Love, Porzingis’s No. 4 comparable, it finds another big man with a good outside shot and strong rebounding numbers — but ignores that Love is shaped much differently than Porzingis, 5 inches shorter but quite a bit bulkier, especially in his youth. And since CARMELO relies on metrics that utilize box score stats, it can be difficult to translate calling-card traits like “quick feet in defensive transition” to on-the-stat-sheet comps, which in turn means the model may promote the formal similarity of Porzingis’s defensive numbers to those of Shawn Kemp, when he’s closer to Andrei Kirilenko stylistically.
You can get a fuller sense for the range of possibilities when sorting through Porzingis’s top 50 CARMELO comparables, a list that includes everyone from Shaquille O’Neal to Darko Milicic. There’s also a cameo appearance from Porzingis’s idol, Dirk Nowitzki, who checks in at No. 17. Why doesn’t he rank higher? Because, as Nowitzki correctly points out, Porzingis has been considerably better so far at age 20 than Nowitzki was at the same age. CARMELO “thinks” the Porzingis-Nowitzki comparison is unflattering — to Porzingis.
