At-The-Break Study: Marion & Mavs Defense

As demonstrated again in Indy, 'The Matrix' is a Swiss Army knife of fingers ready to plug the leaks born of others' faults to perforate the dam … yes, even that disarmingly crooked left pinky. Yet his advanced-stat stuff isn't jumping off the page. We think we've discovered why - and why Dallas needs to use the All-Star Break to make a Marion-related change.



An oversight as egregious as it is surprising

Shawn Marion is one of the most underappreciated players in the NBA. His effective versatility has been a genuine rarity. Whether it's guarding the opposing team's biggest offensive threat, almost regardless of position, or scavenging points without the necessity of plays being designed and called for him, the myriad avenues he can impact a game remains impressive.

He's one of only four players to total at least 17,000 points, 9,700 rebounds, 1,600 steals and 1,100 blocks, joining three who will undoubtedly share their final basketball homes in the Hall of Fame: Hakeem Olajuwon, Karl Malone and Kevin Garnett.

He's never made an NBA All-Defensive team in any capacity, an oversight as egregious as it is surprising to those that have been privileged enough to watch him play. The fact that his worthiness of Hall-of-Fame induction will be debated, with some (many?) doubting his credentials, will eventually show us who hasn't really been paying attention.

The Phoenix Suns escape from relevance as a contender began the moment Marion was dealt in 2007-08, though they would return once more the Conference Finals without him in 2009-10 (their sole playoff appearance in a full season without Marion since his departure), as their defensive rating steadily climbed from the 105-107 range to fall in the area of 110-111. Steve Nash was the offensive catalyst, but Marion was the glue.

The Matrix is a crutch, a Swiss Army knife of fingers ready to plug the leaks born of others' faults to perforate the dam … yes, even that disarmingly crooked left pinky. His awkward shot releases, ability to jump seven times in two seconds (only a slight exaggeration), and once frightening athleticism that can still challenge players born a full decade later, come together to form a rather unique weapon at both ends of the court.
Smile
Yet, going back to the 2010-11 season, Marion routinely ranks near the bottom of the Dallas Mavericks roster in per-game plus/minus and/or net rating, including a team-worst per-game plus/minus of -2.2 this year (the only player on the roster currently posting a negative).

Why? And as the sixth-in-the-West Mavs enter the All-Star Break off the strength of a Marion-led defensive gem in Indiana (an 81-73 victory detailed here), do the numbers tell us something negative about Dallas' team defense and its chances of "playoff noise''?

We'll look at the obvious reasons of "why,'' such as who he is or isn't sharing the court with, but we'll also look into how he's being used with a special focus on this season. Considering the context of the options available to Rick Carlisle, and the team's performance with Marion in certain roles, is it time for change?

The Dirk/power forward factor

Marion is 35, playing on a roster that has had little else to offer in the terms of perimeter defense prior to the return of Devin Harris, who is still finding his defensive legs … and with Samuel Dalembert's inconsistent spryness, often little in the form of a safety net behind him. And, perhaps most impactful to his numbers such as his net rating or plus/minus, he's acted as Dirk Nowitzki's primary backup at power forward.

Of the Mavs starters, Marion has played the highest percentage of his minutes without Dirk, at 44.5 percent – the next closest is Monta Ellis, 37.2 – and the third highest percentage of all players (minimum of 250 total minutes), exceeded only by Vince Carter (56.6) and Brandan Wright (45.1).

As we see below, this plays a significant factor in Marion's overall ratings.

per 36 +/-O RatingD RatingNet Rating
Marion w/Dirk1.2110.0107.32.7
Marion w/o Dirk-7.0102.1106.8-4.7


Playing without the lift of Dirk is certainly a factor, and the obvious first spot to look. It's logical. It's not a coincidence that the only season that Marion didn't finish last or second-to-last on the Dallas roster in per-game plus/minus was the season he played the lowest percentage of his minutes without Dirk, only 24.3 percent in 2009-10.

This directly hints at another current, perhaps more impactful, issue.

When looking though a bevy of other numbers, research originally intended to more clearly define Marion's defensive impact ... something else began to stand out.

The Mavs have struggled this season when Marion is asked to play power forward, particularly in certain circumstances, as we'll attempt to show.

Dallas has nine five-man units that have played over 50 minutes together. Only four of those have negative net ratings, three of those four, including the two worst, feature Marion at the four.

5-Man Units w/ Negative Net Ratings (min 50 minutes)

MinsO RatingD RatingNet RatingReb%Per 48 +/-
Calderon-Ellis-Carter-Marion-Dalembert16398.2106.1-7.649.1-11.5
Calderon-Ellis-Carter-Marion-Blair135102.7104.9-2.345.4-5.7
Larkin-Carter-Crowder-Dirk-Blair6191101.8-10.952.4-7.9
Calderon-Ellis-Carter-Marion-Wright55101.1116.8-15.849.5-12.3


In itself, perhaps this is simply another reminder that this team functions better with Dirk on the floor. That it is, but there may also be more to it.

With Dirk on the floor, the Mavs have 110.0 and 105.0 offensive and defensive ratings respectively, and a 48.7 rebound percentage. To see the points scored per 100 possessions drop with him on the bench is entirely expected, but it comes somewhat as a surprise that the defense with Marion in place of Dirk actually slips in two of the three lineups above, with the third being only negligibly better.

It's also strange to find no significant bump in rebounding percentage in the most used lineup (49.1 with Marion at power forward compared to 48.7 with Dirk), but a noticeable drop in the second (45.4 with Marion at power forward compared to 48.7 with Dirk).

If you switch out Marion for either Brandan Wright or Jae Crowder with the same surrounding four as the lineups above, you get very different results. You also wander into the zone of shrinking sample sizes, a realm where outside influences, such as selective deployment by Carlisle, play a larger factor.

Take these with a grain of salt, but the change in result is significant.

5-Man UnitMinsO RatingD RatingNet RatingReb%Per 48 +/-
Calderon-Ellis-Carter-Wright-Blair2695.894.21.652.4-3.1
Calderon-Ellis-Carter-Crowder-Wright18138.6112.825.856.318.3
Calderon-Ellis-Carter-Wright-Dalembert9118.572.745.853.836.5


As the numbers below will show, perhaps it isn't that Marion has performed worse at power forward, but that the shift it causes in others has carried a negative impact, while also minimizing Marion's biggest advantage over opponents: defense and rebounding … though cross-matchups from offense to defense muddy some of these numbers for Marion, who may be listed at one position while guarding another.

How Marion as performed at small and power forward (per 82games.com):

Marion% of Teams MinutesPEROpp PERReb/48Opp Reb/48Net PERNet Reb/48
SF3113.815.597.9-1.71.1
PF2817.716.912.512.30.80.2


While Marion technically has performed better at power forward by these measurements, his two greatest strengths almost disappear, allowing opponents a 16.9 PER and essentially being equalized on the boards.

As a small forward, Marion is elite on the glass, ranking in the top 10 in rebounding percentage. As a power forward, his current rebounding percentage would rank somewhere in the area of 40-45.

Beyond this, Marion at the four more often than not means Vince Carter at small forward and/or Brandan Wright at center, where this season has shown both often to be outmatched.

Here are Wright's numbers at power forward and center, and Carter's at shooting guard and small forward:

% of Teams Minutes atPEROpp PERReb/48Opp Reb/48Net PERNet Reb/8
Wright at PF423.221.39.9131.9-3.1
Wright at C1424.725.910.117.7-1.3-7.6
Carter at SG1718106.55.980.6
Carter at SF3113166.37.2-3-0.9


Take this with the qualifier that Carlisle and the Mavs have been hesitant to play Wright at power forward, where he doesn't show the range they prefer from the position in an offense largely designed around the presence of Dirk.

It's also worth noting that these numbers were almost exactly reversed last season for Carter, though 2011-12 saw similar results to those above (better performance at shooting guard).

Perhaps these numbers shouldn't be viewed without the additional context of how they individually perform with Marion on the court, where Carter/Marion has surprisingly rated as the worst duo on the Mavs with a minimum of 150 minutes together (-6.1 per 36 minutes).

Per 36 +/-O RatingD RatingNet RatingMinsTotal Mins
Carter w Marion-6.100102.2107.4-5.25761233
Carter w/o Marion9.600110.41028.46571223
Wright w Marion-5.600112.8116.3-3.5141530
Wright w/o Marion10.000119.4106.612.9389530


In each of these categories, Carter and Wright perform far better when Marion is on the bench, though it's hard to say all of this is due their being forced into their less productive positions.

For one, due to the substitution patterns Carlisle has adhered to this season, Dirk is often pulled somewhat early, with Marion shifting to power forward prior to the opponent making any lineup changes … meaning, subs (most often Carter and/or a backup center) are fairly often playing against the opponent's starting five. Even if this only occurs two minutes a half, after 52 games that would add up to 208 total minutes (give or take after the odd influences that may pop up, such as foul trouble or injury), enough to make an impact.

Further widening the gap is the fact that the trio of Carter/Wright/Marion has only shared the floor with Dirk for a total of two minutes, outscoring the opponent by three points, while the duo of Carter/Wright without Marion has 174 minutes with Dirk, posting a plus/minus of +12.7 (per 36 minutes), a net rating of +17.0.

While this influence is undeniable, Dirk is not the sole source of the difference. Carter and Wright, playing without Marion or Dirk posts a net rating of +8.4 … add only Marion to that, with no Dirk, and they go to -5.5 … still significantly different levels of production.

One could argue that Brandan Wright, via his play and contract, has earned the right to play; allowing him more minutes at power forward is an easy avenue to grant him that without removing DeJuan Blair from the rotation. It could create a knot in guard minutes (if Carter's suddenly getting more time at shooting guard), but it shouldn't be too difficult to spread that hit around… at least in the case of Monta Ellis's heavy-minutes load, this could carry its own positive ramifications if he's kept a little bit fresher.

Smile

The Mavs have shown moderate success with Wright at power forward. He's seen 116 minutes with either Blair or Dalembert, showing enough positive results to grant the notion consideration, posting a net rating of +4.3 in 99 minutes with Blair and +6.9 in 17 beside Dalembert.

So it's about more than 'Trix

This is not directly a comment on Marion, but on how the trio of Carter/Marion/Wright or duo of Carter/Marion as the forwards has fit together this season, along with the positions they are most likely to fill when sharing the court … leaving both Carter and Wright in spots they've been less effective this season, while stealing some of the versatility Marion brings – if there is only one other "big" on the court with Marion, it makes it much more difficult for him to provide perimeter defense without forcing someone with significantly less length to cover a power forward or center.

There is one significant caveat to the assertion that the team has struggled with Marion at power forward: Jae Crowder.

Previously, we showed the poor performance of 5-man units with Marion at power forward and at least 50 minutes together. If we lower that minimum to 20 we find another interesting trend.

5-Man UnitsMinsPer 48 +/-Net RatingReb%
Calderon-Ellis-Crowder-Marion-Blair3412.714.350
Ellis-Carter-Crowder-Marion-Blair3217.911.655
Calderon-Ellis-Crowder-Marion-Dalembert22-28.9-18.250
Calderon-Ellis-Crowder-Marion-Dirk212.3-238.6


In three of the four lineups featuring only one "big" and Marion, Dallas either held their own or dominated. There was also a nice uptick in rebounding with three of the four lineups.

There are two common threads throughout the four: Crowder is included. Wright is not.

The presence of Crowder provides another thick body more capable of banging inside than Carter, able to share in the physically taxing dirty work that is otherwise left solely to Marion and the center … a negative trait that is exaggerated when Wright and Marion form the Mavs interior without a third "big."

All of the physical characteristics that make Wright a better fit at power forward than center are only compounded when paired with a power forward who has a body more aptly designed to play small forward, in Marion.

Problem: Identified

As Carlisle stated earlier this season, we in the media get to be "problem-identifiers" while he and his staff must be "problem-fixers." His job is far tougher than ours. Here, we believe we've identified a problem ... and believe the Mavs should attempt to fix it.

Smile

When Carter, Marion and Wright are positioned at small forward, power forward and center respectively, or Carter and Marion are paired as the forwards, the Mavs have struggled. Making rotation changes to alter that won't necessarily result in nightly efforts like we saw in Indy, where Marion was holding Paul George to 4-of-17 shooting and the Mavs were holding the Pacers to a season-low 32-percent shooting and an all-time series low of 73 points.

But we do believe we've discovered advance stats that demonstrate the best way to keep Marion as a "stopper'' and the best way to milk passable defensive performances out of many of his teammates.

There's the data. It's Carlisle-and-company's job to work around it, to be the "fixer."



(All stats current through Feb. 10, 2013, pulled from NBA.com, SynergySports, NBAWOWY.com, 82games.com, and Basketball-Reference.com ... lead photo courtesy of Stevallica.)