The most accurate (low-noise) adjusted plus-minus ratings ever to appear in the public domain.

I love this measure of a player’s worth. It takes a much better look at the impact a player has on defense, something that Hollinger always seems to ignore. Apparently 82 games tweaked the formula to make it more accurate although there are still some obvious anomalies. Some things I noticed:

1. Devin Harris is 8th overall. His defense counts for a lot more with this particular formula.

2. LaMarcus Aldridge is ranked 22nd defensively. This might have more to do with our lack of options when he is off the floor, but it’s nice to see none the less.

3. Gerald Wallace is ranked 23rd defensively and Tayshon Prince is 18th. I love those guys.

As with any rating system it has a million flaws, but it seems a little better than the normal plus-minus statistic.

over 3 years ago Winchester 11 comments 3 recs |

Story-email Email Printer Print

Comments

very nice

beat me to this by 3 minutes! this will keep the basketball stat heads happy for quite a while. there’s a lot of fun stuff to run through in this model. unlike hollinger’s PER stat, this takes into account defensive contributions.

Defense
it’s always reassuring when the numbers agree with what you see on on the court, e.g. garnett is far and away the best defensive presence in basketball, although i am surprised to see k-mart ranked so high. rounding out the top 10 are your usual defensive names: chandler, ming, artest, duncan, ben wallace, marion, camby, josh smith.

Offense
the top four are nash, lebron, cp3, and kobe. i find it interesting that nash is first, although this is most likely due to sampling data over the past five years. nash’s offense is offset by his horrendous defense, once again backing up what one sees on the court.

Other
anyway, there’s a lot of cool stuff to go through. as winchester said, LMA is ranked 22nd defensively and 18th overall (one spot ahead of sheed!). roy is ranked a surprising 62nd due to what this model says is overall poor defense. i think we all see that brandon has the capability to shut people down at times, but does not do this consistently yet. frye and pryz were the only blazers ranked in the top 25 of the “low minute” players.

JAH

by abetterbomb on Nov 4, 2008 2:26 PM PST reply actions

i should mention

bottom of the heap for the major minute guys: jarrett jack

third from the bottom for minor minute guys: sergio rodriguez

JAH

by abetterbomb on Nov 4, 2008 2:31 PM PST up reply actions

Travis Outlaw and Zack Randolph rank about the same on defense.

Not what I want to see.

If somebody hits you with an object you should beat the hell out of them.-Charles Barkley

by Winchester on Nov 4, 2008 2:30 PM PST reply actions

Of course, Zach's rating on defense has never been that bad in any plus-minus showings.

Amare, Boozer, Jefferson, Brand, and Jamison all rate worse than Zach.

by poster on Nov 4, 2008 2:43 PM PST up reply actions

That is due to defensive rebounding

Zack can’t guard anyone, but he cleans up the defensive glass a lot better than Trout.

by trk on Nov 4, 2008 3:47 PM PST up reply actions

I'm still unsure on this.

I agree that this is a very appealing metric. However, can it really be trusted? Most of the time I see this stat (or any plus-minus type stat), it’s only used to “prove” a point on a specific player, ignoring all other players.

For example, how many people would say that Outlaw (-1.13) is a better defender than Webster (-2.29)? Let’s include Danny Granger (-3.28), Josh Childress (-2.76), Matt Harpring (-4.13), Matt Barnes (-4.08), Mickael Pietrus (-1.27), Jeff Green (-3.03), Richard Jefferson (-3.18), and Marvin Williams among others. Some people with strong defensive reputations show up well, while others show very poorly.

The Roy example is a good one for why I can’t subscribe fully to this yet. We had a hard time beating anyone last season without Roy (1-7, with the only victory being a 2-pt one over the depleted Clippers). It also doesn’t agree with what our eyes see for net impact.

by poster on Nov 4, 2008 2:30 PM PST reply actions

Outlaw is defending guys from the other teams second unit

Granger is usually defending the other teams best guy. It makes the numbers look a little funny.

If somebody hits you with an object you should beat the hell out of them.-Charles Barkley

by Winchester on Nov 4, 2008 3:11 PM PST up reply actions

Adjusted plus-minus subtracts strength of opponent and strength of teammates, reportedly.

If the explanation you gave is correct, this stat isn’t doing its job correctly. If the explanation isn’t correct, then there are no conclusions to be drawn. I’m just saying that the numbers don’t always look quite right. Is Jamario Moon an elite player, for example? Maybe he is, but it’s hard to agree with that at face value.

by poster on Nov 4, 2008 3:29 PM PST up reply actions

Roy's impact

When Roy was in the game, that means that Jack and Sergio weren’t playing as much. Since Jack and Sergio were both pretty terrible in adjusted +/-, Roy only has to be a little above average to be a big upgrade.

by trk on Nov 4, 2008 3:54 PM PST up reply actions

Hmm, that brings up something else I've wondered about.

Last season said that our tall players were all above average (Jones, Outlaw, Frye, Aldridge, and Przybilla) while our short players were all below average except for Roy (Webster, Jack, Blake, and Sergio). Is this actually true or is the sample size even adding extra seasons still too small? I’m guessing this means we played well when we went unnaturally big and did poorly when we went unnaturally small for a few possessions.

Considering PGs rate poorly overall in the league in this, maybe going unnaturally big for spot minutes is simply good strategy. It’s just hard to understand how Jack could be so much worse than, say, Frye. I think Frye is better, but the degree is quite big here.

by poster on Nov 4, 2008 4:05 PM PST up reply actions

Tall Players

I think the trend may be a little exaggerated (due to sample size, and also because the Blazers were somewhat deficient in big men last year forced to play some players out of position) with the Blazers, but it is also a league-wide trend so there should be some truth to it. In general, tall players tend to do better than short players, especially on defense.

If you look at the defensive adjusted +/- ratings, you can see that there is a pretty clear correlation between size and defensive impact: bigger is better. Guards, even ones with good defensive reputations, will struggle just to break even on defense, especially if they are undersized. On the other hand, “big men” tend to have a positive impact on defense, as long as they aren’t completely abysmal. Part of this may be due to defensive rebounding, and though it also suggests that size may be more important on defense than quickness. Basically, if you want to have a good defensive team, it is a good idea to use players that are a little oversized for their positions.

by trk on Nov 4, 2008 4:23 PM PST up reply actions