Statistically Speaking: The Effect of Subtracting Richard Jefferson

[timvp]

Following the trade of Richard Jefferson, many of us were saying that at the very least the transaction will go down as addition by subtraction. In other words, Jefferson had become such a liability that the Spurs will be a better team simply by removing him from the equation.

Do the statistics support that hypothesis? The obvious first step is to compare how the Spurs performed when Jefferson was on the court compared to when Jefferson was on the bench.

Defensively, the hypothesis appears to be wrong. When Jefferson was on the court, the Spurs allowed 103.22 points per 100 possessions. When Jefferson was on the bench, the Spurs allowed 104.67 points per 100 possessions. Thus, there is an expected increase of 1.45 points allowed per 100 possessions -- not exactly great news for an already shaky defensive squad.

However, the offense is an entirely different story. With Jefferson on the court, the Spurs scored 105.74 points per 100 possessions. With Jefferson riding the pine, the Spurs scored 112.64 points per 100 possessions -- or an improvement of 6.9 points scored per 100 possessions.

Why have the Spurs been so much better on offense without Jefferson? A lot of factors go into it. The most glaring difference is the fact that San Antonio has gone to the free throw line much more without Jefferson on the court. The overall percentage is basically the same but, per 48 minutes, free throws made (14.4 to 17.6) and attempted (19.8 to 24.3) go way up. Overall shooting percentage also rises without Jefferson, from 45.3% to 48.3%. And while three-point percentage drops a smidgen (from 40.3% to 39.1%), the decrease is somewhat negated by an increase in made three-point field goals (8.0 to 8.5).

Offensive rebounding percentage is also up without Jefferson, from 22.6% to 25.9%. However, not all the offensive news is good: Turnovers go up without Jefferson (12.8 to 13.8) and assists go down (22.7 to 21.5). That said, the aforementioned increase in overall offensive efficiency (+6.9 points per 100 possessions) should quell most fears. The stats paint a picture that shows, going forward, the Spurs will lose some precision and ball-movement but an increase in aggression will result in better quality of shots and more trips to the charity strip -- with the overall result being a much improved offensive attack.

Team-wide, we see the Spurs will get slightly worse on defense but a lot better on offense without Jefferson. To break it down even further, let's take a look at how his departure will effect each particular player. Of the three following charts, the first one shows how the Spurs performed offensively when the player was paired with Jefferson and and how he fared without Jefferson. The second chart concerns defense and the final chart is the sum of offense and defense.


Points Scored Per 100 Possessions - Team
http://oi41.tinypic.com/2vnnvop.jpg



Points Allowed Per 100 Possessions - Team
http://oi40.tinypic.com/hrn1ag.jpg



Point Differential Per 100 Possessions - Team
http://oi39.tinypic.com/m7pc4.jpg


Offensively, the big winner of the Jefferson trade is Kawhi Leonard. He has massively struggled when paired with Jefferson this season. Without Jefferson, the rookie has actually been quite good. Two other players who shouldn't be upset with Jefferson leaving are Tiago Splitter and Matt Bonner. The two bench bigs are fantastic without Jefferson. All in all, only DeJuan Blair doesn't see an improvement.

Defensively, the numbers aren't nearly as pretty on first glance. However, a few things worth noting: First of all, Tim Duncan has been really good on defense as long as Jefferson isn't on the court. Secondly, even the negative numbers in the Difference column are all rather negligible. Well, other than Blair, who sees a gargantuan drop without Jefferson around. To put it another way, the Spurs allow 115.32 points per 48 minutes when Blair plays without Jefferson also on the court.

(What is the logic behind Blair's struggles without Jefferson? That number is shockingly bad but I guess it just goes to show that Blair is useless on defense unless he's playing with the starting lineup. I assume the rest of the starting lineup has figured out how to play with Blair and that without that knowledge, Blair's weaknesses on defense are exposed.)

In the final chart, we see that a number of players can expect to see addition to their effectiveness by the subtraction of Jefferson. Youngsters Leonard, Danny Green and Gary Neal go from liability to asset. Duncan also appears headed for a large improvement.

To zoom in even closer, let's look at the scoring ability for each player with and without Jefferson.

The first chart shows how many points the player scores per 40 minutes with Jefferson on the court and how many points the player scores with Jefferson on the bench. The second chart shows the points per shot for each player with and without Jefferson (points per shot is simply points scored divided by field goal attempts).


Points Scored Per 40 Minutes - Individual
http://oi40.tinypic.com/m93w5f.jpg



Points Per Shot - Individual
http://oi41.tinypic.com/f02nwl.jpg


The first chart shows Parker's scoring goes from good to great. Green goes from below average to rather potent. Splitter becomes an elite scoring bigman without Jefferson.

In the second chart, the improvement that really stands out to me is Duncan's increase. Duncan scoring at a rate of 1.12 points per shot is honestly rather disappointing; the Spurs can't win in the playoffs with that number so low. However, without Jefferson, Duncan's number improves to 1.33 -- which is slightly higher than Duncan's career rate of 1.31. If the Jefferson trade allows Duncan to become a force on the offensive end, that alone will make the trade a beneficial one.

And while Ginobili is at the bottom of both charts, I wouldn't worry. Firstly, Ginobili was on fire to begin the season and he did a lot of his scoring with the a starting lineup that included Jefferson. Second of all, his points per shot number is still fantastic without Jefferson -- much higher than his career rate of 1.41.

The bottomline is, without even considering what Stephen Jackson will bring to the table, the Spurs should be a better team simply by trading away Richard Jefferson. The team-wide offensive improvement should be noticeable going forward and certain individuals, including Kawhi Leonard and Tim Duncan, should become much more effective players overall.

[slick'81]

didnt read all but then again doesnt take a rocket scientist to figure out rj was a bad fit from day 1 with the spurs

[Nathan89]

Not sure if you want to change it but...
1st paragraph --Because should be Become

Wow Splitter improve big time just from a lack of RJ. He should be even better with Sjax. I feel like the connection from Sjax to Splitter in the first game is the sign of many good things to come.

Thanks for the hard work

[timvp]

Not sure if you want to change it but...
1st paragraph --Because should be Become

Thanks :tu


Wow Splitter improve big time just from a lack of RJ. He should be even better with Sjax. I feel like the connection from Sjax to Splitter in the first game is the sign of many good things to come.
Yeah, it was great to see Jack have success on his first pick-and-roll with Splitter. Hopefully that is the first of many.

[Dr. John R. Brinkley]

To me, Splitter looks to greatly benefit with the subtraction of RJ. And as almost all of us are clamoring for Splitter to get more minutes, this can only be a good sign. As long as Leonard doesn't struggle with the addition of SJax, the Spurs should clearly be headed in the right direction.

And in regards to Blair, though I want him to do well, if he fails without RJ then at least he fails quickly and can move to the bench and allow more time for Splitter. That's the best we can hope for at this point.

[baseline bum]

Hey LJ... I think you were wrong when you compared RJ to Brandon. They both might like balls on the face, but at least Brandon got shit done before getting got. To use a favorite word out of Carver's lexicon, Jefferson is more shitbird than anything tbh.

[Russ]

Points Scored Per 100 Possessions - Team
http://oi41.tinypic.com/2vnnvop.jpg

Not to nitpick, but because RJ was a starter, the sample size for the pairings at the top would seem to be quite a bit smaller.

[SpursRock20]

Great write-up!

When Jefferson comes to mind, I think of a one-dimensional player that only contributes when his 3-ball is falling (Bonner, anyone?).

When Jackson comes to mind, my mind floods with hope. Although we did not address 2 of our primary needs at the trade deadline, I and most fellow spurs fans, alike, were excited by the swap. Granted, I was only 11 when Jackson was last here, but I have faint memories of an emotional and balls-to-the-wall type player.

He is not the prototypical Spurs player but I believe a championship team needs a dark horse of sorts that is different than the rest. A player that is not here just to fit in and ride on the Big 3's coattails, but one who rises to the occasion and feels free to do whatever it takes is a neccesity. This is what I think of when Jackson comes to mind, not Jefferson.

His first possession against Dirk was priceless. You could tell that Dirk was thinking, "What the fu**, are we playing the Spurs?". I love his pitbull mentality and respect how he understands that he must be a team guy that does whatever it takes.

[Russ]

He is not the prototypical Spurs player but I believe a championship team needs a dark horse of sorts that is different than the rest.

Ellie comes to mind in '99.

Jack in '03.

Horry in '05 and '07.

[Obstructed_View]

RJ looked like a different player his first game in GS, which is why I'm glad he's gone.

[analyzed]

Great stuff, timvp. It would be interesting to see if you did the same numbers analysis on Bonner, what would it look like, I've always wanted to determine Bonners value or lack of value, but stastistically speaking I have yet to see a stat that trully determines this

[jesterbobman]

Some of that is double counting. Points per shot ignores the fact that points come on FTA, with so many extra FTA with Jefferson out of the lineup, PPS is bound to increase.

The significant increase in offense isn't that shocking, mostly because of who's playing those minutes. RJ is most likely to be out with Green or Kawhi playing SF, and both of those guys are way better than him. (WP48, Kawhi is at .287, Green at .173, RJ at .08). Basically, swapping an elite player/semi elite player in for RJ is going to make the team a lot better. Swapping Kawhi in for RJ explains 6.4 points per 48 minutes, which is close to the effect in these numbers.

[jjktkk]

Thanks for the writeup Tim. I wouldn't even care if stats showed that RJ was the better option. Sometimes,(this instance) you have to throw stats out the window. Jack provides that "it" factor, which stats can't measure.

[timvp]

Points Scored Per 100 Possessions - Team
http://oi41.tinypic.com/2vnnvop.jpg

Not to nitpick, but because RJ was a starter, the sample size for the pairings at the top would seem to be quite a bit smaller.Not to nitpick, but because RJ was a starter, the sample size for the pairings at the top would seem to be quite a bit smaller.

Good thought but the sample sizes are relatively close. Leonard has played 389.2 minutes with Jefferson and 569.8 minutes without Jefferson. Splitter (326.4 with, 471 without), Bonner (316.2 with, 577.4 without) and Neal (353 with, 475.5 without) are also pretty much in line.

I guess the one good thing about Jefferson playing a lot of minutes is it allowed for good sample sizes for each player pair :lol

[timvp]

Hey LJ... I think you were wrong when you compared RJ to Brandon. They both might like balls on the face, but at least Brandon got shit done before getting got. To use a favorite word out of Carver's lexicon, Jefferson is more shitbird than anything tbh.

:lol Then maybe RJ is that guy who tried to stab Omar in jail . . .

[timvp]

Great stuff, timvp. It would be interesting to see if you did the same numbers analysis on Bonner, what would it look like, I've always wanted to determine Bonners value or lack of value, but stastistically speaking I have yet to see a stat that trully determines this

Don't have the stats in front of me but the theme of Bonner's comparable stats is simple: He numbers look great in the regular season and bad in the playoffs.

[DeadlyDynasty]

I see the Honeymoon phase of RJ's departure hasn't ended yet:lol

[Mel_13]

I see the Honeymoon phase of RJ's departure hasn't ended yet:lol

That honeymoon lasts until Jackson makes his first contract extension demand.

[pgardn]

Lots of hard work so props to you.

But this really requires more data.

[RodNIc91]

Great analysis as always Timvp. You should patent this ish already.

Also, I'd like to point out one reason I think the offense became so stagnant with RJ. Let's face it, he was always a great athlete who could play basketball, not a basketball player with athletic ability. Thus, he was never going to be able to create a decent shot for himself or others . He became a good spot-up shooter but he was never gonna thrive since the offense nowadays is more perimeter oriented and requires more ability from said players (a reason why GHill was beggining to thrive). All in all, what I'm trying to point out is that when the ball was in RJ's hands, the offensive posession was very predictable. This, I believe, is somewhat backed up by the statistics.

And that is why Bonner posts good +/- numbers in the regular season. Like Timvp has pointed out, he makes quick good decisions with the ball, thus resulting in great ball movement. He is not a hesitant shooter anymore and he ocasionally makes a drive which opens shots and driving lanes for others.

[SenorSpur]

Another great writeup LJ.

We can all take shots at RJ - and deservedly so. He was a terrible fit, a mental midget and he rarely played with any passion. The Spurs never should've traded for him in the first place.

Having said all that, I have to give the guy credit for his consistent professionalism. He never created any waves, wasn't a lockerroom cancer and was never a distraction. As a matter of fact, I was surprised to hear that he volunteered to workout with Kawai Leonard over the offseason, after he was drafted by the Spurs. That says a lot.

I'm so glad the guy is gone, but I do wish him the best in Golden State.

[timvp]

Some of that is double counting. Points per shot ignores the fact that points come on FTA, with so many extra FTA with Jefferson out of the lineup, PPS is bound to increase.PPS includes free throws but I don't understand what's "double counting" about it. The PPS I used was for each individual so it's not like the increase in team free throw attempts was inflating the individual stats. I used PPS because it highlights the fact that it's easier to get to the line when RJ isn't on the court, for whatever reasons, and that happens to be the main reason why the overall offense is much improved.

I could use eFG% and that would show similar improvements but 1. it misses out on a big reason why the offense is improved without RJ 2. I don't see how eliminating free throws makes it "single counting" instead of "double counting".


The significant increase in offense isn't that shocking, mostly because of who's playing those minutes. RJ is most likely to be out with Green or Kawhi playing SF, and both of those guys are way better than him. (WP48, Kawhi is at .287, Green at .173, RJ at .08). Basically, swapping an elite player/semi elite player in for RJ is going to make the team a lot better. Swapping Kawhi in for RJ explains 6.4 points per 48 minutes, which is close to the effect in these numbers.In theory, I agree that WP48 would explain the overall improvement after taking out RJ. However, in this particular case, I think WP48 is more of a coincidence than the reason:

1. Leonard's WP48 has been great virtually all season but it hasn't corresponded well at all to the actual play. The Spurs have struggled with Leonard on the court even though WP48 predicts the Spurs would be much better with Leonard.

2. The WP48 equation leans heavily on stats that have little to do with offense (namely defensive rebounds, steals and blocks) and thus the reason why Ben Wallace led the league in this stat a few years and why even an extremely washed up version of Marcus Camby is still elite. I'd argue that the WP48 difference from RJ to Leonard should manifest mostly on the defensive end ... but that's not what happened.

3. WP48 has been a poor predictor of Spurs basketball this season. Going by WP48, Leonard has been responsible for more wins than any other players on the team ... and we know both objectively and subjectively that isn't true. It also tells us that Danny Green has been responsible for more wins than Tim Duncan this year, which we obviously know doesn't hold any water.

4. WP48 has a poor history of actually "finding" the key players on the Spurs. According to WP48, Bowen was worthless, Parker is mediocre and Duncan is about as good as Brent Barry. I'm a big fan of WP48 in a league-wide sense and to easily find overrated/underrated players, but the correlation between a high WP48 number and actual production over the years has been extremely low. PER is basically antiquated these days but it actually correlates between for the Spurs over the years.

[RGMCSE]

Excellent write up! Why did RJ never fit in? Is it because he didn't really feel like he fit in? It seemed like he always referred to the "Spurs" instead of "we" in some of his interviews. He was a better player than this, but for some reason he shutdown.

[Cant_Be_Faded]

Good stuff lj, although I still think a decent increase in offense does not outweighan even tiny decrease on our already feeble defense.


Good thing we play the lakers a bunch in the next few Weeks, the competition against that size will truly tell us where this team is at both defensively and offensively

[maverick1948]

Many things can come from stats. After a game, we all wonder why Bonner has a great +/-, yet attack his defense. Blair's on court play is the one unexplainable fact. He seemed to be the RJ connection. It would make you think that our second team is better than the starters, as each had a better ranking with and without Jefferson. All these can be gathered from the above charts.

My question is this: "What will be the 'Jackson effect'?" Give him 10 games with 5 in the next 6 days starting Wed. By the 1st couple of games in April, we will see a difference in these charts with SJ.

[gospursgojas]

The drop in defense (or spike in opponents scoring....however you wanna look at it) with rj on the bench has everything to do with rj playing most his mins with tim.

[DPG21920]

Astounding that everyone but Blair improves without RJ

[bigfan]

I think one of the "intangibles" about Jackson over Jefferson that opposing teams must keep in mind is that much like the presidents they are named for, Jackson is much more likely to punch you in the face and break your teeth.

[baseline bum]

I see the Honeymoon phase of RJ's departure hasn't ended yet:lol

LOL, did Laker fan ever regret losing Kwame Brown? Likewise, I cannot picture a time when Spurs fan will ever look back and say "Damn, I wish we would have kept the cocksucker another year".

[Dex]

I'm betting someone is already planning a writeup on this, but I'd really like to see how the subtraction of Jefferson affects the Spurs from a financial standpoint. Obviously, taking his contract off the books and exchanging with Jackson's, which expires in 2012/2013, should change the outlook of the next couple summers.

Another factor which I'm sure has been mentioned, but hasn't really been highlighted, is how the Spurs saved their amnesty which 90% of people assumed they would use last offseason. Who will be the main target for amnesty now? Will Bonner get the axe if he flames out again in this years playoffs?

[Libri]

The most glaring difference is the fact that San Antonio has gone to the free throw line much more without Jefferson on the court. The overall percentage is basically the same but, per 48 minutes, free throws made (14.4 to 17.6) and attempted (19.8 to 24.3) go way up.

I found this observation interesting. His career average in FT attempts demonstrate a significant drop off from his days in New Jersey and Milwaukee to his first year with the Spurs. A small drop off was expected because he would no longer be the primary option on offense but to go from 6+ attempts with the Bucks to 1.2 attempts this season shows that RJ turned into a one dimensional player on offense, content with camping out at the 3-point line.

The Spurs FO didn't sign Richard to become a 3-point shooter. From what I remember, Jefferson was supposed get his points by the flow of the game, which included driving to the basket and making free throws.

[therealtruth]

I found this observation interesting. His career average in FT attempts demonstrate a significant drop off from his days in New Jersey and Milwaukee to his first year with the Spurs. A small drop off was expected because he would no longer be the primary option on offense but to go from 6+ attempts with the Bucks to 1.2 attempts this season shows that RJ turned into a one dimensional player on offense, content with camping out at the 3-point line.

The Spurs FO didn't sign Richard to become a 3-point shooter. From what I remember, Jefferson was supposed get his points by the flow of the game, which included driving to the basket and making free throws.

It's interesting he never really felt comfortable in the offense doing anything but shooting 3's. Meanwhile Jack is able to feel at home and he's already found Splitter for a dunk.

[Rapper]

After read this I can see that timvp doesn't like RJ at all

[analyzed]

Don't have the stats in front of me but the theme of Bonner's comparable stats is simple: He numbers look great in the regular season and bad in the playoffs.

So if we can't count on either Bonner and Blair in the playoffs. Are we saying we will either see a lot of Duncan and Tiago playing together and a forward line-up Jackson and Kawhi playing together ? If you ask me I would rather go to war with either combination, they have the defensive and competitive edgeness in them to succed in the playoffs

[Obstructed_View]

After read this I can see that timvp doesn't like RJ at all

Jefferson is averaging 6.5 free throw attempts in his first two games off the bench in Oakland. He hasn't averaged that many since he was in Milwaukee.

[Horse]

Ellie comes to mind in '99.

Jack in '03.

Horry in '05 and '07.
Jaren Jackson '99
Speedy Claxton '03

[jag]

PPS includes free throws but I don't understand what's "double counting" about it. The PPS I used was for each individual so it's not like the increase in team free throw attempts was inflating the individual stats. I used PPS because it highlights the fact that it's easier to get to the line when RJ isn't on the court, for whatever reasons, and that happens to be the main reason why the overall offense is much improved.

I could use eFG% and that would show similar improvements but 1. it misses out on a big reason why the offense is improved without RJ 2. I don't see how eliminating free throws makes it "single counting" instead of "double counting".

In theory, I agree that WP48 would explain the overall improvement after taking out RJ. However, in this particular case, I think WP48 is more of a coincidence than the reason:

1. Leonard's WP48 has been great virtually all season but it hasn't corresponded well at all to the actual play. The Spurs have struggled with Leonard on the court even though WP48 predicts the Spurs would be much better with Leonard.

2. The WP48 equation leans heavily on stats that have little to do with offense (namely defensive rebounds, steals and blocks) and thus the reason why Ben Wallace led the league in this stat a few years and why even an extremely washed up version of Marcus Camby is still elite. I'd argue that the WP48 difference from RJ to Leonard should manifest mostly on the defensive end ... but that's not what happened.

3. WP48 has been a poor predictor of Spurs basketball this season. Going by WP48, Leonard has been responsible for more wins than any other players on the team ... and we know both objectively and subjectively that isn't true. It also tells us that Danny Green has been responsible for more wins than Tim Duncan this year, which we obviously know doesn't hold any water.

4. WP48 has a poor history of actually "finding" the key players on the Spurs. According to WP48, Bowen was worthless, Parker is mediocre and Duncan is about as good as Brent Barry. I'm a big fan of WP48 in a league-wide sense and to easily find overrated/underrated players, but the correlation between a high WP48 number and actual production over the years has been extremely low. PER is basically antiquated these days but it actually correlates between for the Spurs over the years.

Nice work