You may have seen that David Berri’s wins produced formula ranks Glen Davis among the 45 least productive players in the NBA this season. (For an explanation of the WP formula, go here). By net wins produced, only eight players in the league were worse than Davis; when broken down by minutes played (the WP/48 column in Berri’s post), 32 players on the list of 45 rank worse than Davis–he’s penalized in the net WP category because he played more minutes than most of the guys on this list.
No one reading this site thinks Glen Davis is a star. But to be told he’s one of the least valuable players in the league is surprising, if only because fans naturally develop an affection for players they watch a lot.
But is Baby really this bad? My quick reaction is that he scores badly here (and in PER) for two reasons: 1) He’s a bad defensive rebounder, and both metrics value rebounding highly; and 2) He was awful on offense until around mid-January. Let’s take these one at a time.
I want to start here, because I think net statistics miss the broader narrative we’ve seen unfold during Davis’ second season in the league.
To say Glen Davis has become a different offensive player this season is an understatement. Few NBA players change their offensive games from one season to the next as dramatically as Glen Davis did. In his rookie year, 29 percent of Baby’s shots were jumpers; 67 percent came from in close. This year, he nearly reversed those percentages–60 percent of his attempts were jump shots, and just 36 percent came from in close, according to 82games. I wonder how many players adjust their offensive games to this extent through their entire careers.
Let’s look at this in graph form, since graphs are fun and have colors. First, here’s Baby’s NBA hot spot chart from last season:
Glen Davis took 37 jump shots last year. Thirty-seven jumpers. All season.
Now, here’s the same chart for this season:
Glen Davis took more jumpers from the left baseline this season than he took from all over the court combined in 07-08. (Yes, Baby played 8 more minutes per game this season, but that doesn’t disprove the larger point). It’s a jarring change when you look at it this way.
And it started horribly. On January 12, Kelly Dwyer wrote the following on BDL: “Davis is killing the Celtics. Destroying this team. His PER on the season is around 8, which puts him amongst the worst rotation players in the NBA, and PER doesn’t account for his defensive attributes. Which are awful. Ruddy awful.”
(I love that Kelly Dwyer uses the word “ruddy” on a regular basis. He is awesome).
And Kelly was right. Davis’ PER was under 8, which is really, really hard to do if you’re playing real minutes in the league, if only because you’re not good enough to play real minutes. On Jan. 19 on my old blog, I pointed out that, on that date, Glen Davis was shooting 28 percent on jump shots for the season. I criticized the shot selection (and the connected decline in FTAs) and implored Baby to go to the hoop more.
The guys at Red’s Army, on the other hand, took a longer view. In posts I can’t find right now (Red’s is re-designing, it appears), Red’s called for Davis to keep shooting, arguing that he could help the team most by developing a jump shot.
Red’s was right, and I was wrong, and Glen Davis developed a jump shot. So what happened? Davis’ PER for the season is now around 11, and his shooting percentage on jump shots is up to 37 percent. Neither of those numbers are good (an average PER is around 15), but considering Davis had a 7.5 PER and a 28 percent mark on jumpers after 50 games, they show considerable development. Since late January, Glen Davis has compiled about a league average PER.
Rebounding and Defense, after the jump.
Glen Davis is not a good defensive rebounder. And if there’s one thing that makes people uncomfortable with Wins Produced and PER (note: I said “people,” not “me”), it’s that both systems give huge weight to rebounding. (Please note: Berri’s system makes adjustments for the fact that big guys get more rebounds than guards, and the numbers he ends up with rank players relative to those who play the same position/s).
Davis’s defensive rebounding rate is 12.8, meaning he grabs 12.8 percent of available defensive rebounds. This is terrible for a power forward. Of 81 forwards (SFs and PFs) who played enough to be eligible for the scoring title, only eight had lower DRB rates than Davis. In order from “best” to “worst”: Hilton Armstrong, Brandon Rush, Darius Songalia, Nicolas Batum, Thaddeus Young, Al Thornton, Steve Novak, Bruce Bowen and Jason Kapono.
Not good company for a burly power forward. (Celtics trivia: Pierce is 53rd; Powe 29th).
To his credit, Davis’ 9.4 percent offensive rebounding rate ranks him 22nd among these 81 players. That type of ORB-DRB rate gap seems unusual; the same players (other than Armstrong) populate the basement of both lists. (Celtic trivia: Leon Powe is #2 in the entire NBA in this category, trailing only Kevin Love. Get well, Leon).
I’m not sure why Davis is so bad at rebounding on one end and decent at the other, and we’re nearing the 900-word mark, so I’m not going to speculate now (though I have e-mailed a couple of NBA experts to weigh in).
As for defense, opposing power forwards put up a solid 18 PER against Baby, and the C’s defense gives up five more points per 100 possessions with Baby on the floor. The new Basketball Prospectus stats show he’s cut his opponents production by about seven percent, but that he’s typically matched up against bad players.
I’ll leave the full defensive analysis of Baby for another post. My goal here was to provide some explanation for why he fares so poorly in some of the most prominent player value metrics we have (including Hollinger’s new Wins Added, where Baby ranks 288th in the league, having “added” -0.6 wins to the Celtics total this season).
I’d be very curious to see how the New Baby–the post-Jan. 15 Baby–would finish in these metrics over the course of a full season. I wonder how much the Celtics are willing to pay to find out?