One of the main themes we have written about over the past few weeks is absurdity of some of the Wolves' recent rotations. How weird are these rotations? I believe there are two solid ways for average fans to break down team play: comparison of 2 and 5 man units. 5 man plus/minus stats give us an idea of how well entire rotations play together. 2 man plus/minus stats give us an idea of how substitution patterns can be better constructed and how different player combos can affect the greater lineup when subbed in with one another (or for one another). By comparing how smaller rotations are worked into greater 5 man units, we can make decisions about who plays better with who, who is effective with other 2 man pairings, who doesn't work well together, who makes the biggest impact on 5 man rotations, and so on and so forth.
- Brewer/Flynn, Brewer/Jefferson, Flynn/Jefferson, Brewer/Gomes, Flynn/Gomes, and Gomes/Jefferson are the heavy lifters, each with more than 1000 combined minutes. Each duo also has a negative efficiency margin, the least offensive being Gomes/Jefferson.
- The highest rated duo on the team is Gomes/Sessions, with an eff margin of +3 in 528 minutes.
- Of all the duos with an eff margin more than the team average, only one pairing is a duo from the starting lineup (Gomes/Jefferson). (Note: This could have a lot to do with the relative level of competition; we'll have more on that angle in a later post; the bad news here is that there is data that suggests the reserves play just as much better against opposing starters as do the starters. This angle is also diminished in 2 man pairings because it's easier to compare how well a Starter A/Starter B combo does in relation to Starter A/Bench Player A or Starter B/Bench Player A than it is with similar comparisons in large 5 man groupings.)
- One of the most reliable indicators of the Wolves performing above their own efficiency average is a double digit positive possession margin. In plain English, when a Wolves duo creates 10 or more possessions per its time on the court than its opposition, the Wolves have a fighting chance. This is yet another reminder of just how inefficiently the Wolves' offense has played this year and how poorly they shoot the ball from the floor.
- The Wolves' starting front court duo (Jefferson/Hollins) is the 2nd worst duo on the team next to Jefferson/Pavs. The 3rd worst duo are yet another starting pairing, Flynn/Hollins...as are the 4th (Gomes/Hollins) and 5th (Brewer/Hollins).
- The Wolves' best post duo is Love/Jefferson, and it's really not even close.
- Sessions/Gomes > Flynn/Gomes, Brewer/Sessions > Brewer/Flynn, Jefferson/Sessions > Flynn/Jefferson...you get the picture.
- Kevin Love does not have a single 500+ possession pairing below the team average. Ramon Sessions only has one, when he is paired with Pavlovic.
- Ryan Hollins is a black hole of suck.
- Jonny Flynn plays relatively well with Wayne Ellington compared to Sessions. Ellington plays relatively well with Kevin Love compared to Sessions. How many minutes do you think the Ellington/Flynn/Love combo has seen the court? Would you be shocked to learn that the lineup of Flynn, Ellington, Gomes, Love, and Jefferson has actually been somewhat effective in very limited action?
But in the Mavs’ infamous first-round loss to Golden State two seasons later? “I got burned when following the advanced stats,” Johnson says. Winston’s numbers showed that during the regular season, the Warriors had smoked the usual Dallas starting line-up, which featured Dampier at center. In a decision he now regrets, Johnson adjusted his starting line-up for Game 1 by benching Dampier and starting Nowitzki at center. The Mavs lost. Johnson, though, stands by the decison. “It was the right move,” he insists. Still, he reversed course for Game 2 and went back to the normal starting line-up. Dallas won, and Johnson believes the Mavs played better because they were–psychologically–more comfortable with Damp at center. “Everybody had freaked out” at the Game 1 line-up change, Johnson says.
And that represents the closest thing to a consensus that emerged from this panel: The best decisions will be made when coaches consider advanced stats not alone, but alongside everything else—what their eyes tell them, what the film shows and the psychology of each individual player.
Nate Silver suggests that numbers are crucial, but that in-game context matters. If the stats say a football team should go for it on 4th down and five 62 percent of the time, the crucial thing will be figuring out if a particular 4th down falls under that 62 percent. That might have to do with the personnel available, the health of the team, the weather and other variables, Silver says. “But if every time 4th and 5 comes up, you’re looking for excuses not to go for it, you’re probably not making the best decisions,” he says. (Silver in general believes coaches should experiment more with different line-ups. “You have some regular-season games against the Clippers,” he says, becoming at least the fourth person to insult the poor Clips today).
Back to hoops: It’s clear Johnson wants to learn from the Wayne Winstons of the world, but he believes such information should form just part of the decision-making equation. In Dallas, for instance, Cuban hired a full-time sports psychologist to talk to the players and attend every practice. Johnson thought it intrusive at first but grew to like the arrangement, because the psychologist–like Winston–gave him more information. “It gives me an edge beyond knowing what happens when Jason Terry gets the ball on the right side of the floor,” Johnson says. “What is he thinking in a pressure situation? How is his family situation? Is he tough enough mentally to fight through fatigue?”
The quest is for more information, and the best coaches, everyone seems to agree, will be open to all types of information.