Plus/Minus Over Time - Feb. 16, 2012

[timvp]

In a recent thread, MannyIsGod brought up the idea of plotting plus/minus numbers over time. I decided to give it a try ... and the results are interesting, to me at least.

In the following graph, the x-axis represents time (in games), while the y-axis is plus/minus per 48 minutes. I found that it takes about 200 minutes played for the noise to be limited enough to make plus/minus useful, so each player's plus/minus number only got plotted after they reached that 200-minute minimum.

http://oi41.tinypic.com/2entg69.jpg

Observations:

-Tim Duncan's number really started to sag at one point but he has recovered as of late. I think that correlates well to what we saw. Duncan went through a stretch where he wasn't too effective but has recently played much better.

-Richard Jefferson's hot shooting early in the season obviously helped his plus/minus number. But as he's cooled, so has his ability to help the team. It's not too surprising that he's slowly but surely headed south.

-DeJuan Blair was decent early this season, had a horrible stretch and lately has been treading water. His plus/minus numbers follow a similar trend.

-Kawhi Leonard :(. As good as he looks at times, he's still a 20-year-old rookie who doesn't yet know what it takes to win on the NBA level. It doesn't help him that he has played a lot against starters and that a big chunk of his minutes came next to Blair ... but this graph basically says that Leonard is consistently unhelpful.

-Tony Parker has been carrying the team and you can see exactly that in this graph. Comparing his line to the team line, it's apparent that he has been a steadying force who has a big influence on the team's overall play. Parker's consistency has been impressive, especially considering how the four starters he played next to for much of the season have been pulling his numbers lower.

-Tiago Splitter has really come on of late and you can see just that in the graph. He started off near the bottom but has rocketed upward. There was a time this season when it became painfully obvious that he's a much better player than Blair ... and you can almost pinpoint that exact moment using the graph.

-Matt Bonner, Plus/Minus King of the World. As we all know, the problem with Bonner isn't his regular season play -- it's the fact that he has never come close to replicating that level of play in the postseason. Another observation: though his line is impressive, it does have a number of jagged edges. That boom or bust type production is okay in the regular season but you don't want to rely on a role player in the postseason who might fall flat on his face and cost you a game.

-Danny Green's plus/minus numbers show how he's able to positively impact a game no matter how well he's shooting. Just by looking at his line, you can't tell that he was in an awful slump recently. This may also explain why Pop trusts him so much and why Pop has inserted him into the starting lineup. If you're Pop, you probably want to keep playing him until you figure out whether or not it's a fluke.

-Gary Neal wasn't very good coming back from injury. Specifically, his defense was a total mess. But he's played pretty darn well as of late. His plus/minus numbers reflect those improvements.

-James Anderson just recently qualified; you'll find his line up near the rarefied Bonner air. It seems like when Anderson has been given time, he either really helps or really hurts. But more often than not, he seems to cause the Spurs to go on a run. It's probably an anomaly because it has seemed more like he was just along for the ride ... but I guess we'll see as the season progresses.








Any further observations greatly appreciated.

Thanks.

[Borosai]

The alpha dog is simply leading the pack he's supposed to lead.

[Cant_Be_Faded]

I just can't wrap my mind around plus minus in the context of one game, much less over the course of a season. Leonard not once has a consistent stretch of positive plus minus? Its completely catheter intuitive, at least when he was supposedly shutting down Rudy Gay, and they little stretch he had around there.

[Blake]

How bout plus/minus lines during the playoffs in say, the last three years?

[Cant_Be_Faded]

Btw, this graph seems to imply that the are once again winning the wrong way. If Leonard is our best perimeter defender and he had worst plus minus, Bonner is our guaranteed choker, leading plus minus, we could be heading for another first round complete pwnage.


What the heck is up with Leonard paralleling Blair ?

[Spurminator]

Its completely catheter intuitive

That's one of the strangest auto-corrects I've ever seen. :lol

[WeNeedLength]

That's one of the strangest auto-corrects I've ever seen. :lol

:lmao I knew that looked strange. I thought it was some new term going around...

[Cant_Be_Faded]

That's one of the strangest auto-corrects I've ever seen. :lol

Crofl Crofl. Counter intuitive.

[Bruno]

Unlike a lot of teams, Spurs don't have, aside of Parker, a big drop off between their starting unit and their bench:
- At SG/SF, Neal, Green, Leonard and RJ are very different players but have a somewhat similar level.
- At PF, Blair and Bonner are both average.
- At C, Splitter isn't that far away of an aging Duncan.

At the end, Spurs starters suffer against quality units. Spurs bench, that isn't far from starter's level, plays against significantly worst units and do naturally better.

I'm not that worried about Spurs starting unit relative weakness. Ginobili will change a lot to the equation and give a big boost to the starters.

Aside of that, it's quite cool that +/- numbers shows the improvement in Duncan's and Neal's play.

[Kewni Leonard]

Hopefully this illustrates for people who don't understand basketball the value of Matt Bonner and what a great roleplayer he is to have. Fucking bargain of a contract, too.

[ElNono]

If you take Anderson as an outlier, the most fairly recent productive lineup is:

Parker-Neal-Green-Splitter-Bonner

As well as Duncan has played lately, you would think he would be part of that group.

Neal would likely be replaced by Manu over time.

And I agree about the sharp swings... You want production to be smooth as possible. That indicates the player is productive regardless who he is on the court with.

[DMC]

Much of that says the opponents have shit for benches, and the fact that our starters have been bailed out by the bench in a few games backs that notion. Sure we have a good bench, but they have poor ones, and our starters are not as good as theirs in many cases.

[timvp]

Its completely catheter intuitive, at least when he was supposedly shutting down Rudy Gay, and they little stretch he had around there.

lol catheter

The Gay game was #22 but unfortunately he had bad plus/minus games around it.


What the heck is up with Leonard paralleling Blair ?

They each play about half their minutes with each other on the court, so their line is going to be pretty similar.

Leonard and Blair together are at about -5.9 points per 48 minutes.

Leonard without Blair is at about +0.3 points per 48 minutes.

Blair without Leonard is at about +5.2 points per 48 minutes.




Tbh, those numbers are drastic enough that perhaps Blair and Leonard together should be avoided.

[SA210]

How bout plus/minus lines during the playoffs in say, the last three years?

[chazley]

You mean I've been right about Bonner the last 2-3 years? What a shocker.

[The Truth #6]

Unlike a lot of teams, Spurs don't have, aside of Parker, a big drop off between their starting unit and their bench:
- At SG/SF, Neal, Green, Leonard and RJ are very different players but have a somewhat similar level.
- At PF, Blair and Bonner are both average.
- At C, Splitter isn't that far away of an aging Duncan.

At the end, Spurs starters suffer against quality units. Spurs bench, that isn't far from starter's level, plays against significantly worst units and do naturally better.

I'm not that worried about Spurs starting unit relative weakness. Ginobili will change a lot to the equation and give a big boost to the starters.

Aside of that, it's quite cool that +/- numbers shows the improvement in Duncan's and Neal's play.


It will be interesting to see how our deep bench will be utilized in the playoffs.

1. I'm curious if Pop will maintain a semblance of his current rotation in the playoffs.

2. Assuming a playoff opponent doesn't play their bench, will our bench be able to maintain its current contributions against the other teams' best players?

[E-RockWill]

I find it funny that Bonner & Blair's lines are almost a mirror of each other.

[Blake]

could be wrong, but it seems to me the biggest +/-flop during the playoffs the last couple of years was George Hill.

Could partly explain his exit from San Antonio..

[z0sa]

WTF Leonard?

[therealtruth]

could be wrong, but it seems to me the biggest +/-flop during the playoffs the last couple of years was George Hill.

Could partly explain his exit from San Antonio..

George Hill was an enigma. The Spurs don't beat the Mavs without him in '10 because of his defense and offense. But then he makes Nash look unstoppable in the next series.

[rmt]

It will be interesting to see how our deep bench will be utilized in the playoffs.

1. I'm curious if Pop will maintain a semblance of his current rotation in the playoffs.

2. Assuming a playoff opponent doesn't play their bench, will our bench be able to maintain its current contributions against the other teams' best players?

I'm not holding my breath that Pop will maintain a semblance of his current rotation in the playoffs. Too many memories of me screaming at the TV for Pop to insert Splitter (last year) and Hill (the year before).

[Blake]

George Hill was an enigma. The Spurs don't beat the Mavs without him in '10 because of his defense and offense. But then he makes Nash look unstoppable in the next series.

He stunk pretty bad last year against Memphis as well

[timvp]

Spurs bench, that isn't far from starter's level, plays against significantly worst units and do naturally better.I'm sure that's true. The bench this year has been a positive in just about every game.


I'm not that worried about Spurs starting unit relative weakness. Ginobili will change a lot to the equation and give a big boost to the starters.In the past, Ginobili was useful to give the Spurs a powerful bench. Since their bench is already damn good, this might be the first year where Ginobili absolutely must start.


Aside of that, it's quite cool that +/- numbers shows the improvement in Duncan's and Neal's play.Yeah, Neal's defensive improvements in recent weeks was hard to quantify ... but his plus/minus line illustrates it pretty well.

[RodNIc91]

Is there a stat similar to plus/minus but that reflects deefensive performance?

[TJastal]

Btw, this graph seems to imply that the are once again winning the wrong way. If Leonard is our best perimeter defender and he had worst plus minus, Bonner is our guaranteed choker, leading plus minus, we could be heading for another first round complete pwnage.


What the heck is up with Leonard paralleling Blair ?

Maybe because they play the same minutes together of most games? Maybe that's why Leonard is sucking, Blair fatass is dragging him down.

[TJastal]

Hopefully this illustrates for people who don't understand basketball the value of Matt Bonner and what a great regular season roleplayer he is to have. Fucking bargain of a contract, too.

fify

[Borosai]

This is Bonner's year. Guaranteed.

[Interrohater]

Does anybody have a plus minus of Bruce? Maybe a championship season, earlier, whatever.

[RodNIc91]

Other thing I would be interested in looking would be the plus/minus per 5man unit if there is such thing. For example, from the graph I can see how great of a combination is Tony + Tiago, but we are not able to see who fits there as well.

[Cry Havoc]

Just musing here, is it possible for a 3 point specialist to have a smooth line of +/-?

Seems like guys who shoot almost exclusively 3s are going to be feast or famine by nature.

[Cry Havoc]

Mmm, thinking about it, it actually makes a lot of sense about Bonner.

If he's hitting, he puts a big on the floor who also stretches it. It makes us virtually unguardable as a team.

If his shooting is off, he basically makes us 4 on 5 on offense and he isn't very productive as a defender.

Yes, Bonner can have games where he plays well outside of shooting, but most of the time, that's not the case.

[maverick1948]

+/- is a useless stat. You read into it what you want. But it doesnt show a single thing that makes sense.

Take the Miami 3rd quarter blowout. Who had the best game, Jefferson or Green?

Green had 20 pts in 23 mins. Jefferson had 0 pts in 23 mins. Green had a +/- of -21. But Jefferson had a +/-of 0. Does this mean that Jefferson played better D than Green?

How about Philly? Parker had 37 pts in 38 mins. Green had 0 pts in 30 mins. Parker +/- of +5, Green +13.

Look at the Toronto game. Bonner +17, Jefferson -8. Bonner 20 mins, Jefferson 26 mins. Both scored 10 pts. Does this mean that Bonner played DEFENSE and Jefferson didnt?

How does +/- show anything worth knowing?

[dylankerouac]

How does +/- show anything worth knowing?

A person can bring on-the-court chemistry or a special skill set which enables the rest of their team to score more points than the other team, on average.

Just because players don't score doesn't make them useless to a team; if they were useless, generally the opposing team should benefit in some way. Of course the worst way the opposing team can benefit is by scoring more points.

This stat can be pretty helpful in arranging line-ups.

[Obstructed_View]

The more I look at some of these advanced stats, the more sure I am that either they're not being displayed properly or they are completely useless in determining anything outside of themselves.

[GSH]

Matt Bonner's +/- per 48 tonight was... wait for it... -155.45! That's gotta be some kind of record.

[sabar]

Problem with plus minus is that it is useless unless rotations are purely randomized. Bonner has a great plus/minus because he generally is in a group that is playing very good team ball.

It is really easy to see this if you view plus/minus live during a game, seeing it change with each bucket. You can see Bonner's plus/minus go up even on nights where he hits nothing and plays only average D. The reason? It is clear from watching that players like Splitter and Neal were large reasons for the high rating.

Of course, there are nights where Bonner is the reason for a high plus minus as well. But you really do need to watch the game to see the true value of this particular stat. It is very deceiving when a player is constantly playing with the same group of guys.

[mathbzh]

Problem with plus minus is that it is useless unless rotations are purely randomized. Bonner has a great plus/minus because he generally is in a group that is playing very good team ball.

It is really easy to see this if you view plus/minus live during a game, seeing it change with each bucket. You can see Bonner's plus/minus go up even on nights where he hits nothing and plays only average D. The reason? It is clear from watching that players like Splitter and Neal were large reasons for the high rating.

Of course, there are nights where Bonner is the reason for a high plus minus as well. But you really do need to watch the game to see the true value of this particular stat. It is very deceiving when a player is constantly playing with the same group of guys.

I do not disagree on the fact that you have to watch game, take a look at the rotation... But Bonner has been a constant regular season (I insist on that) +/- king since he is here.

You can't just assume he is not a reason for that +/-.
Splitter Benefit from Bonner presence just as much as Spitter from Bonner.

Actually, Bonner does not need to hit shot every nights. All it takes is teams having to respect his shot. You can't let Bonner wide open.

Personally, I use +/-
- As a hint about a player impact on the rotation.
Clearly Blair/Bonner is terrible while Blair/Splitter kills any second unit.
- As a grain of salt (If a player has great stats with terrible +/- or the other may) indicating that maybe I am missing something.

I also try not to look at +/- for individual games too much.

Now what is clear is that you can't use +/- to rank players.