Worst 10 starters in the league

[WigNosy]

One of the things I slipped into the parsing engine that runs the "lists.php" page and pulls in players' game logs is a little thing that tries to track the "primary defender" against a starter (the opposing starter at the same position). This doesn't work for bench players, and it's not perfect because obviously players shift around during the course of the game and the game doesn't give us a perfect look at who is guarding whom at all times... but I threw it in there in case I figured out something interesting to do with it later.

Well, I figured out something interesting to do with that information.

When I was still in high school, I ran across a book in my local library called "Basketball's Hidden Game: Points Created, Box Score Defense, and Other Revelations" by Bob Bellotti. He was one of the forefathers of what we now call "advanced metrics" and presented some interesting ideas in that book. It wasn't perfect - even in high school I found some problems with the methods he presented (and mistakes in his calculations!) but what that book presented has always been in the back of my mind.

Using the game logs I decided to fire up a system that follows some of his ideas about "Box score defense" - basically, the idea that you can use box scores to approximately measure how good a player is at defense by looking at the aggregate statistics that the opposing starter puts up. I decided it would be interesting to use a system I modified in a way I felt was a better representation than Bellotti uses to measure who is getting outplayed on a consistent basis by comparing each player's overall Productivity score with the net Productivity score put up by players he's starting against in games played against him.

Each player is listed with a "net differential" score - this measures roughly how many "points per minute" the opposing player can be expected to contribute relative to this player when they are matched up. A positive number is good, meaning on average, this player is more productive than his counterpart on the opposing team (playing this player puts you at an advantage). A negative number is bad, meaning on average, his opposing counterpart is more productive (playing this player puts you at a disadvantage). Without getting too deeply into the nitty-gritty, we have 219 players that qualified for comparison... about 65% of the them score between -0.2 and +0.2 and a full 85% score between -0.3 and +0.3... so these guys are well outside the "normal" range.

Note this list is restricted to players who have STARTED at least 10 games for their teams this season to make sure that we eliminate outliers that might have had only one start (maybe due to injury) and maybe happened to get matched up against AD or Kevin Love in that one start. With at least 10 starts, we can be confident all of these players have seen a significant cross-section of the league. Here they are, in order of "least worst" to "worst of the worst."

10. Arron Magness (PF, San Antonio Spurs) - Net differential -0.360

At first glance, Magness looks like he's having a good season - 16.5 ppg on 48% shooting. But when you take a look at his defensive ratings (10 PSD, 10 PRD), you see why he would be a net liability. He doesn't rebound well which means scoring is the only way in which he is productive and by not offering any defensive resistance, nor contributing anywhere else, he's the classic example of "empty stats on a bad team."

9. David Brunette (C, Los Angeles Clippers) - Net differential -0.378

Brunette is getting minutes mostly because the cupboard is bare in the frontcourt in Los Angeles. He's got some nice potentials, but he's still quite a ways off from the player he will become. He's the classic "not ready yet but there's nobody else to play and our season is going nowhere anyway so we may as well see what we've got" young player who figures to improve as he goes through a couple more training camps.

8. Al-Farouq Aminu (PF, Cleveland Cavaliers) - Net differential -0.397

Aminu may be getting the starting nod in Cleveland based on reputation (color and past defensive greatness) but he's no longer a net positive (if he ever was). His free throw and three point shooting have slipped this season and the Bellotti system - unlike PER - dings players for taking and missing shots. His fouls have jumped up this season and his scoring has dipped. Again, this is the classic example of a player who's getting minutes based on name, not production.

7. Barrett Reyes (SF, Charlotte Bobcats) - Net differential -0.436

This one is pretty easy to explain. The Bobcats are imploding under the weight of taxation and so it's time to start tanking. Reyes' ratings suggest he has a nice little jump shot but doesn't do much else well and has F's in PSD and PRD, meaning he's getting destroyed on the defensive end. He reminds me a lot of classic real life Clipper Loy Vaught for more reasons than just his photo. He's a rookie but barring a TC jump, this is probably the only time you'll ever hear of him.

6. Rodney Hood (SF, Los Angeles Clippers) - Net differential -0.441

A lot of the comments I made about Aminu apply to Rodney. He may be getting minutes because of his name and his contract, but he's a bad defender, he doesn't rebound well, and he doesn't shoot particularly well except from three... though he *is* really good at fouling. I think his time with Anthony Davis in Detroit made teams assume he was better than he was... "borrowed greatness by association."

5. Myron Logsdon (PG, Orlando Magic) - Net differential -0.442

Myron has been pressed into starting duty as the Magic trade away everything remotely resembling a current player asset to get into the tank game... he only has 14 starts so this could be a function of small sample size. But then again, he's got no potential rating about a C, very low strength, and pretty low jump... all of which combine to see him shooting 33% from the field. He will have to take a TC jump before his rookie contract is up to see a starting job again after this season.

4. David McClanahan (PG, Charlotte Bobcats) - Net differential -0.444

Another player being thrown to the proverbial wolves by the tanking Bobcats. McClanahan has some potential and nice athleticism but is super raw now and is getting his lunch handed to him. But the Bobcats don't care because this is a tank season so they may as well get him some burn and if it costs them games, that's a feature, not a bug.

3. Julio de la Rosa (SG, Boston Celtics) - Net differential -0.466

De la Rosa has had some spot starts this season (15 total) but has gotten consistent playing time throughout the year. Everyone had high hopes for him when the Wizards drafted him a couple of seasons back, but so far, he's looking like a great big bust. He has regressed in shooting from his rookie season and isn't showing much ability to contribute off the ball either. His potential ratings suggest he'll eventually grow out of it, but that may STILL be two or three seasons away when he's halfway through a max contract... and unrealized potential gets GM's in hot water.

2. Tua Tagovailoa (SF, San Antonio Spurs) - Net differential -0.549

Another spur getting thrown into the deep end and asked to swim... his defense is shoddy, his shooting is suspect, and the Spurs don't care if they lose. That means Tua will get as much playing time as he wants right now even if it hurts the team. I feel less worried about him than De La Rosa since this is just his first year (compared to Julio's third). He could probably use a little points training on the defensive side, but even if he doesn't get that, once he grows into his offensive ratings, given his athleticism he should be a net positive.

1. Rusty Keenan (PG, Oklahoma City Thunder) - Net differential -0.573

The Thunder have flat-out sucked this season and unfortuantely for Rusty, he's a big reason why. A first-round pick three seasons ago who has never really done much with his career, you can see why the Thunder are floundering with him at point guard. Do you really want a floor general who only averages 2.7 assists versus 2.5 turnovers and shoots 35% from the field and 54% from the line? Not unless you're tanking. Once the Thunder get some pieces next to Jabari Parker, Rusty will be out of the league faster than you can say, "ded." At 30 minutes per game, that -0.573 differential suggests he personally is making a massive contribution to that -11.8 +/- he's sporting rather than it being a result of bad teammates.

[JNR]

I really liked this idea. Great article!

[IamQuailman]

Poor Julio....

[Conroy]

[ballsohard]

Julio and Kat have been huge disappointments so far, I think they’ll turn the corner

[ballsohard]

5 Points sir