C.J. McCollum: Opportunity or improvement?

The Most Improved Player award has a noble intention but every year we run into the same issue: people focus on a big jump in points per game, or another basic stat, and the award becomes Biggest Improvement in Minutes Per Game. That’s not honoring the spirit of the award, and real improvement is subtle and complicated.

C.J. McCollum of the Portland Trail Blazers is the most recent winner. But is he simply another Biggest Improvement in Minutes Per Game winner or did he legitimately get better?

This season, McCollum had one of the largest increases in points per game in NBA history. In 1987, Dale Ellis had the largest increase ever, and naturally he won this award. In fact, since this award was created in 1986, most of the players on the list below of biggest year-to-year scoring increases of have the award.. This is no surprise to anyone but given where C.J. ranked historically you can see why he was a lock to win. We rarely see increases this large and they often happen to young players becoming stars

Largest Year-To-Year PPG Increases in NBA History, Min. 60 Games

(Source: Basketball-Reference)

There are different ways of measuring improvement, if you want to look beyond an increase in scoring average. One could, for example, use Basketball-Reference’s Box Plus-Minus, a box-score derived estimate of a player’s net impact per 100 possessions. McCollum doesn’t fare well there, however, but there’s an issue — BPM is tied directly into a team’s defensive rating, and Portland’s fell from 2015 to 2016. Consequently. McCollum’s defensive BPM did too, even though his defense has gotten any worse. NBA metrics are estimates of reality, not reality itself. Even ESPN’s Real Plus-Minus has his defense slipping by a large amount. Some of the most important improvements a player can make are hard to quantify and nearly impossible with public data, like defensive fundamentals. Looking just at the offensive component of Real Plus-Minus, McCollum had the third greatest improvement behind Kawhi Leonard and Ian Mahinmi. It wasn’t just his increase in minutes; he really did improve there.

There’s another tricky statistics issue to deal with that pertains to McCollum — in 2015, he had a bench role, compiling his stats in a different environment, whereas in his most recent season he started the majority of his games. Most people would assume by consequence that his stats on the bench were inflated because he played against inferior opponents compared to starters, but there a few issues with this. Lineups are more mixed than people think. If you’re a bench player, there’s no guarantee you play against other bench players because teams all have different substitution patterns and there are numerous random effects, like injuries, or foul problems that disrupt when starters play. McCollum played had nearly 1500 minutes before his breakout season to show us what he could do against NBA competition.

There’s also this concept known as the Millsap Doctrine that I love to reference because every season there are a couple of strong candidates who fit the previous findings from the research. Kevin Pelton and Tom Ziller discovered that players who got large increases in playing time due to an independent event like a starter getting injured — this ensures the guy’s playing time didn’t increase just because he got better — saw their per minute production stay the same or even increase in some cases. This goes against conventional thinking; players with elite per minute production but small roles are usually dismissed. However, Paul Millsap and others have flourished with full-time starting gigs.

There’s a tangible reason for that effect too. Per research by another analyst who now writes for ESPN, players perform better when they have longer stints on the court. In other words, it’s tougher to produce and help the team when you only play two minutes in a row versus six. This isn’t just about advanced stats either. If you’ve played basketball, you understand how you need a little time to get into the flow of the game. In fact, the research shows that the effect is greater on offense than defense. This is relevant in comparing players who only get short little bursts on the court here and there. We don’t see them fully “warmed up” and comfortable, and it negatively affects their game.

Going back to McCollum, he improved in one area that you can’t fake, and his promotion probably helped — his usage rate skyrocketed. He’s taking more shots per minute than he was the previous season while his efficiency has actually improved. His assist rate has also doubled. Usage rate is a pretty stable statistic, not prone to wild fluctuations. There was something real and substantial about his leap. Some people cite his playoff performance last year as a harbinger of his growth but his usage rate didn’t surge and he had few assists.

Using the NBA’s player tracking data, McCollum’s time of possession per minute nearly doubled. This means that he held the ball twice as long this season than the previous one, even after adjusting for minutes played. One could argue he’s had this talent all along, and only now with LaMarcus Aldridge, Wes Matthews, and Nic Batum gone has he been able to utilize his full package of skills. But when that happens, players don’t usually improve their efficiency too.

Strangely, McCollum’s greatest improvement is not one people would expect. After two years of hitting roughly 69 percent of his free throws, he finished the season at a respectable 82.7 percent. For a ball handler like him, that’s an important skill because his next area of improvement is his foul-drawing ability. . Like usage rate, free-throw percentages are usually pretty stable. McCollum shot 82.5 percent in college, so this is probably just a case of him regressing to his mean after two down seasons in the NBA.

With a new role and set of responsibilities after the exodus of most of Portland’s starters, McCollum improved and adapted into his environment. There are a few natural reasons for his development, from age to opportunity, but he had appreciable advances in some important areas of his game. It was in fact one of the most predictable Most Improved Player awards ever; an award like that is by its nature usually unpredictable. While he was a lock for the award because of how his points per game average changed, he really has gotten better.

It’s just tough to compare his improvement to others — defense is subtle and largely ignored, while there’s a tangled web from the interaction of new schemes, new teammates helping each other, opportunities to show unused skills (McCollum’s teammate, Mason Plumlee, is a good candidate for Most Inexplicable Improvement with his passing), and other factors. There’s a legitimate case for someone like Kawhi Leonard. But at least McCollum has a case too that extends beyond the usual shallow marker of points per game.