Basketball is a fluid game with a variety of different interactions, inherent randomness, and nuances that aren't immediately obvious from only looking at statistics or only watching games. In this piece we'll look at the lineup data and some game tape to understand why Andre Miller is important to the Nuggets.

First The Fundamentals

In order to lay the foundation for what we're about to discuss, it's helpful to distill the game down to its fundamentals. Not to state the obvious, but the goal in basketball is to score more points than your opponent. It doesn't matter how you get there or how ugly or pretty it looks - if you score more points than your opponent you'll win the game, period.

There are various ways we can accomplish this. We'll start by looking at the extreme cases: elite offense and elite defense. Imagine a team that plays really good defense and gives up 0 points per game. In this scenario, their offense can be atrocious and score 2 points and still win every game. You can also imagine the opposite scenario where a team plays extremely good offense and scores 200 points per game while giving up 198 points per game and wins every game. If you invert the scenarios above you can come up with the extreme cases of really poor defense and really poor offense.

Good basketball teams usually fall somewhere between the two extremes of good offense and good defense with one common theme - they score more points than they allow over the course of a game. The best teams in the league have the highest point differentials rather than the best records because point differential measures teams on a possession-by-possession basis.

The media draws narratives heroic performances and glorifies individual stars, which makes us basketball fans also think about the game in terms of standout or poor individual performances. For example, when LeBron passes up the game winning shot we tend to think wow, why did a superstar pass up a game winner rather than wow, what a great basketball play. We're advocating a different way of looking at basketball and that's viewing offense and defense from a TEAM perspective rather than an individual perspective. In order to do this, we need to understand the different ways in which points are scored and look even further at how quality uncontested shots are generated (we'll refer to these as good looks).

Offense

In the past few years there's been a bifurcation of opinions where the traditional media has emphasized individual shot creation while the analytics community has emphasized efficiency. Both sides of the argument are both right and wrong and the two go hand-in-hand.

First, let's look at the components of how points are scored. The most basic concept on offense is a possession, or an opportunity to score a basket. You can score more points by either increasing the number of possessions (rebounding, turnovers), using possessions more efficiently (addition of talent, better shot selection, or better looks), or increasing the value of the possession (getting to the FT line). The most important component of good offense is using possessions efficiently. Why? If you averaged 1.1 points per possession instead of 1 point per possession you'd score 10 more points per game over 100 possessions. You'd need to gather 10 more offensive rebounds per game at 1 point per possession to makeup the efficiency differential of that extra 0.1 points per possession. We tend to think that you can improve offense by adding more individual talent to a roster. How many times have you heard the argument, if we had a star like Carmelo or Kobe we'd score more points? A superstar is valuable because he brings several things to the table: usually more efficient use of possessions and some combination of a better than average ability to get to the line and/or rebound the basketball.

Second, we need to understand the importance of good looks and how they're created. If we view offense from a team perspective, we realize that one of the best ways to maximize efficiency of possessions is to create good looks - wide open or quality shots for your existing players. If you gave every player on your team a wide-open layup or even a wide-open jump shot most players in the NBA would be shooting at an elite level way above 50%. People think of basketball as a one-on-one isolation game and assume that superstars have to create these good looks for themselves or for their teammates when in reality you can create good looks via individual talent, scheme, or some intermediate combination of the two. At the heart of creating good looks is the ability to create 2-on-1 situations or gain a numbers advantage. We can't stress this enough. Good team offense is a direct result of a team's ability to get good looks while the genesis of a good look is a team's ability to create a numbers advantage somewhere on the court at some point in the possession.

The most fundamental illustration of this concept is the 2-on-1 fast break. One player can't possibly defend two players, no matter how big the talent gap, and George Karl understands this. The goal of half court offense is exactly the same! In the half court, the goal is to create a numbers advantage somewhere on the court and force the defender to help or rotate.

The first way to create a numbers advantage is by utilizing superior talent. LeBron James is so talented that teams commit multiple defenders to wall off the paint against him. As soon as the defense commits more than one guy to stop LeBron, one of his teammates is wide open as one guy is drawing 2 or more defenders and 3 defenders are defending 4 of his teammates. The second way to create 2-on-1 situations is using scheme - or offensive designs and plays. For example, setting a pick on a guy's defender allows one offensive player, the screener, who isn't going to get the ball to occupy two defenders (his defender and the screened defender). This leads to a wide open shot for the player coming off the screen or a wide open shot for one of his teammates, assuming good ball movement. The aim of the pick and roll is to get a fast PG on a larger defender, forcing the defense to commit help (2 players to defend 1 penetrator) and leaving a wide open shot somewhere on the floor. Sometimes you'll see a guy get a pass, pump fake, and drive the lane, effectively taking one defender out of the play and forcing help defense. This in turn creates the type of numbers advantage we're talking about where there are 4 defenders for 5 offensive players, assuming the floor is spaced properly. Floor spacing is just a fancy way of saying spread the defense out and make them defend the whole floor so that good scheme can create open looks when defensive rotations are stressed. If the floor isn't spaced well, the opposite happens - the offense creates a numbers advantage for the defense and fewer defenders can defend more offensive players.

Although most of you would have a tough time believing this, the Nuggets are actually fairly efficient on offense. When viewed from a holistic team offense perspective we do a great job of getting transition opportunities and increasing the value of a possession by getting to the line. In the half-court, our offense depends on Nene (traded), Ty Lawson, and to a lesser extent Danilo Gallinari, to create the 2-on-1 situations mentioned above. Nene because he was a very efficient post scorer, often requiring double teams, which he did a great job passing out of and Ty and Gallo because of their ability to get past their initial defender and cause a rotation. One player who does a poor job of creating 2-on-1 situations is AAA - and it's obvious when you watch him. He never gets much separation, his man is always in front of him, and when he takes a jump shot that's not assisted, it's almost always contested.

Defense

You can think of defense as the opposite of offense. A defense does a great job when it, first and foremost, contests shots and forces an offense to be less efficient with its possessions, limits possessions by grabbing defensive rebounds, and finally, doesn't foul often or go into the bonus early. A defense does its job when each player understands their responsibilities, rotates, and defends as a unit.

On defense the only thing that matters is limiting points per possession allowed as a unit. While it's possible, though not ideal, to view an offense as the sum of all the individual player's contributions it's hard to evaluate defense by looking at an individual player's defense in a vacuum. A defensive unit may perform well as a unit even though some of the players on the unit may be poor individual defenders. Additionally, a defensive unit may do a great job defending the initial shot but allow so many offensive rebounds that it's a poor and ineffective defensive unit. It bears repeating that good defense involves forcing bad looks without fouling and, when possible, conceding inefficient shots while also ending the offensive possession as quickly as possible.

Generally, the soundest way to do this is with good individual defenders, but you can also scheme defense. We mentioned that it's important to create a numbers advantage on offense as it leads to good looks. On defense, it is also possible to create a numbers advantage by trapping and doubling. You may be thinking, doesn't this lead to a 2-on-1 situation for the offense? Yes, it does but it's one that the defense has dictated and has preset rotations, zones, and scheme behind the trap to continue to play solid defense whereas the defense is forced to adjust when an offense forces the 2-on-1 situation. If a defense traps indecisively, the offense has the ability to quickly pass out of it and create the opposite - a 2-on-1 situation for the offense with a stressed defensive rotation, or what you've seen all year with opponents creating wide open 3 point opportunities against the Nuggets.

Lineups

George Karl has always received a lot of criticism about his lineups, especially from Matt and I. The goal is to play lineups, regardless of individual talent, that maximize the differential between points scored and points given up. The hard part of lineups is you need to know how all the various strengths and weaknesses of each lineup interact with your opponent's strengths and weaknesses and figure it all out while dealing with the inherent randomness in a basketball game. The chart below shows all Nuggets lineups this year with more than 45 possessions played.

We've analyzed the Nuggets lineups in depth and have come to some interesting conclusions.

Mozgov isn't the answer for how this team is currently constructed

If you look at all the lineups containing Mozgov they are score less than league average on offense and allow way more points than league average on defense. The league average is 101 points per possession on offense and on defense. The best offensive team in the league, the Miami Heat, scores 107.7 points per 100 possessions while the worst defensive team, the New Jersey Nets, give up 108.2 points per 100 possessions. The worst offensive team in the league is the Charlotte Bobcats at 92.6 points per 100 possessions while the best defensive team, the Philadelphia 76ers, allow 93.8 points per 100 possessions. Recall that it doesn't matter if a lineup is individually bad on offense or defense as long as it scores more points than it allows. A lineup containing Timofey Mozgov can be expected to score less than 100 points per 100 possessions while giving up more than 105 points per 100 possessions.

According to Synergy Sports, Mozgov is a really good individual defender allowing 0.78 points per 100 possessions so why are our lineups with Mozgov so bad defensively? If we evaluate Mozgov's defense from a team perspective we get a better perspective of what's happening. Firstly, George Karl is scheming an aggressive doubling and trapping style of team defense. This defense needs athletic bigs to rotate to help and Mozgov simply isn't quick enough. Second, our point guards are getting beat on the perimeter, requiring our bigs to help, and Mozgov isn't rotating quickly or decisively enough. Third, his defensive positioning and ability to recover and help on team defense is also bad. We've provided two clips below illustrating the difference between Mozgov and Koufos. Notice how Koufos maintains good position, forces the pick and roll ball handler to make a decision, then recovers to challenge LaMarcus Aldridge's shot. Mozgov gets blown by and then ends up in no man's land, unable to recover to be a factor on defense.

Mozgov Pick and Roll (via Chris Chan)

Koufos Pick and Roll (via Chris Chan)

The truth about Andre Miller

"Andre Miller is amazing. He's one of the few guys I've ever coached who can will you to victory. Just look at that Philadelphia game and what he did. He just took over. It may not always look pretty, but Andre Miller wins you basketball games." -George Karl, The Association

If you've watched the Nuggets for any amount of time this year you've probably thought to yourself, wow, Andre Miller sucks. He dribbles around way too often and forces up shots, gets blown by on defense, and doesn't close out on 3's. What the hell does George Karl see in Andre Miller? He's always been a very positive player when viewed from the perspective of regularized adjusted plus/minus, over a significant sample size. This suggests that he's doing things that can't be measured by box score statistics, such as defense. However, if you've watched him play defense, you know he's not an elite defender. What exactly does Andre Miller bring to the table?

Diving deeper into the lineup data we start to get a sense of what Andre Miller brings to the table. If you look at all the lineups containing Andre Miller, you notice a few things. Specifically, our offense tends to struggle when Miller isn't playing - almost all lineups without Miller score less than 100 points per 100 possessions while some lineups with him score significantly more than 100 points per possession. We've broken down and categorized the various lineups he's played with this year below.

1. Andre Miller lineups that are elite offensively and obviously positive lineups (score more than they give up). Surprise, surprise, our good lineups are almost entirely the 2 point guard lineups that we complain about and often 3 guard lineups.
   
   
2. Andre Miller lineups that are elite offensively but are so bad defensively they're net negative. There are exceptions, but these lineups are bad because they're poor at defensive rebounding and give up extra easy opportunities. While they're not elite defensively, they don't need to be since they're so efficient offensively.
   
   
3. Miller lineups that are poor in offense, but elite in defense. These lineups tend to score 90-95 points per 100 possessions and give up 85 points per 100 possessions - they're hugely net positive. These lineups almost always consist of Andre Miller, good enough 3-point shooters, athletic wings and bigs. While the athletic wings and bigs aren't efficient scorers, Andre Miller seems to single handedly score enough to make sure that we outscore opponents with these lineups.

If you've watched games, when Miller isn't in, you see a bunch of guys start to dribble into the lane, get stopped, then pass around the perimeter and throw up a contested jump shot when the clock winds down. We've provided two clips below to contrast our offense with Miller and without. In the first clip without Miller, various players try to get a good look, fail, and pass back out to the perimeter, eventually settling for a terrible contested long jumpshot. In the second clip, Andre Miller dribbles around, gets past his defender, forces a rotation, and passes out to AAA for a wide open three.

Afflalo No Look (via Chris Chan)

Andre Miller Good Look (via Chris Chan)

The data suggests that while Miller has deficiencies, they're disguisable to the point where we're able to outscore our opponents with him on the floor. He's so good at one skill that you can hide his defensive weaknesses, or at least play enough team defense to win games with him. What is his skill? Creating good looks on offense! You only recognize this when you 1. understand how teams score, 2. understand the importance of good looks, and 3. evaluate him from a team perspective rather than an individual perspective. Miller not only creates layups and uncontested shots for himself but he does so for his teammates as well. Mozgov has the opposite problem where he may be a better all around basketball player than Miller, but we can't mask the whole of the weaknesses and he becomes a net negative player.

We can conclude that when playing with solid wing defenders and athletic bigs, the defense becomes solid and we tend to outscore opponents. When he's surrounded by bigs who don't rotate well on defense or poor wing defenders, defense suffers tremendously. Also, when Karl plays 3 guard lineups for offense we're able to score but defense suffers because we're not able to get a rebound to end the possession - the lineup ends up being net negative. Finally, he needs floor spacing on offense as he tends to draw multiple defenders and excels finding wide open teammates for lobs and open 3's.

The synergistic effects of all this are interesting because it fits pretty well with the recent Nene trade and Wilson Chandler signing. Chandler is the exact kind of player that Andre Miller needs to succeed - he is able to hit wide-open 3's, defend the perimeter well, and play good individual and team help defense. He's also the type of player who rebounds well for his position, allowing Karl to play the two guard lineups more often without sacrificing defensive rebounding. Finally, if Javale McGee works out, I wouldn't be surprised to see us make a run at Steve Nash as a backup point guard this offseason. Steve Nash is a similar player to Andre in that he creates good looks for teammates while being somewhat of a liability on defense. He's also an exceptional pick and roll player and McGee is an upgrade over Nene in the pick and roll game. You protect Steve Nash on defense the same way you protect Miller - you surround him with great perimeter help defenders and athletic bigs to defend against the blow-bys.