Dallas-Mavs.com Forums - View Single Post

Fidel · 10-10-2009, 11:23 PM

Stubled across this on espn. Winston is the Mavs stats guy. He basicly says that his +/- stats and analysis tell him Durant is a "terrible" player, his performance bottom 10 % of the league. Asked if he´d sign Durant if the Mavs could get him for free he says No. This is an interesting read. I don´t know how much we rely on Winston in player desicions but this is defenitly not a conventional approach. So what do you think. Are we screwed with Winston? Is he a genius?

The Kevin Durant Conundrum

October 9, 2009 3:25 PM

There is almost no one like Kevin Durant. You can make a case he was the best freshman player in the history of the NCAA, and in the two years since he has only gotten better.

He's 6-9. He moves like a gazelle. No one can stop him getting his shot. He's a tremendous teammate and an unbelievably hard worker. He also just turned 21, but has already spent two years feeling out the superstar role.
Kevin Durant
There is almost no surer superstar prospect in the NBA, and yet for some reason Kevin Durant's team plays better when he's on the bench.
(Andrew D. Bernstein/NBAE via Getty Images)

For players his age, his scoring is top-five all-time.

Given that he's a half-decade from the age players typically start to peak, it's not hard to picture him becoming an MVP. The most respected single number to express a player's total contributions -- PER -- ranked him as the 20th best player last season, while projecting him to leap to fifth in what would be his senior year in college.

The expectations are now approaching ridiculous; the future the Thunder have been grooming Kevin Durant for has made some cameos, and he just spent an off-season marinating in his legendary work ethic.

NBA General Managers have selected Kevin Durant as the player most likely to have a breakout season, and he's everybody's pick to be the next player to join basketball's two most exclusive clubs: Team USA and the All-Star team. It's no surprise that Durant is on billboards and the covers of video games. (A surprise of the off-season was word that the Thunder won't be on national TV much; Durant has the trappings of a player the world will crowd around television sets to watch.)

And yet, a shocking piece of news: The Thunder have, over the last two years, consistently performed worse than normal when Durant is on the floor. Any way you slice the +/- numbers, he's one of the Thunder's worst players.

You read that correctly. Kevin Durant, uniformly regarded as an out-of-this-world NBA player, has been killing his team.

Sometimes +/- can punish players simply for being on bad teams, but this is more than that. Mavericks' statistical expert Wayne Winston's in-depth lineup data shows that every one of Durant's key teammates -- Russell Westbrook, Jeff Green, Nenad Krstic, Nick Collison -- gets better, in many cases far better, results playing with less heralded teammates Thabo Sefolosha or Kyle Weaver while Durant sits.

In fact, almost nobody on the Thunder has a +/- rating as poor as Durant's. Winston rates Durant's performance "in the lowest 10% of all NBA players."

An Assault on Conventional Wisdom
The first line of analysis, of any player, are real experts: People who assess talent for a living.

Three out of three I talked to shrug. If you are trying to tell me that Kevin Durant is somehow a terrible player, they say, go ahead. But don't expect me to listen to you.

They are unanimously emphatic that Durant is an absolute gem of a keeper. You can't teach size and mobility. It's hard to teach that kind of feel for getting the ball in the hole. There are times in games when nothing matters more than being able to reliably create your own shot and he already has that. The things he doesn't have -- and nobody denies they exist -- can and will be learned. Players who start their careers like Durant, and keep working, tend to improve dramatically, they say.

Meanwhile, the things that make him inefficient -- mediocre passing, forcing some shots, turnovers, not making teammates better, and of course bad defense -- are all things that improve not just with work but also with better teammates. With the mere passage of time, thanks to an armada of draft picks, cap space, and high-potential young players to develop with him, these things will get much better on their own.

Surrounded by better shooters, for instance, Durant would face fewer double teams and more assists, and less of a need to create for himself and turn the ball over. With better teammates, his team would score much more easily. On defense, meanwhile, veterans have a long track record of getting more consistent results -- the referees may be a factor, but probably even more important is the physical development that allows players in their prime to take nightly poundings and keep on fighting.

Those three experts all had different theories to to dismiss the +/- numbers:

Theory #1: Any player playing long minutes on a bad team would have a bad +/-.
This makes sense at first. Think about +/-. You're on the court when the team gets slaughtered, and you get a bad +/-. Kevin Durant plays a lot. The Thunder have been killed a lot. There's nothing to see here, right?
Kevin Durant and Russell Westbrook
Russell Westbrook doesn't have nearly Durant's reputation, but he has had a much more positive effect on the Thunder.
(Larry W. Smith, NBAE via Getty Images)

As it happens, most Thunder players do have bad +/- numbers. But the Thunder perform particularly poorly when that one player is on the court.

In fact, the Thunder has plenty of lineups that perform well. "In most situations," reports Winston, "holding some things constant and putting Durant in makes things worse, not better."

The Thunder's big three, for instance, are Russell Westbrook, Kevin Durant, and Jeff Green. Last season, in the 387 minutes Westbrook and Green were on the floor without Durant, the team outscored opponents by a dozen points. In the 1,683 minutes all three of them were on the floor together, they were outscored by 245 points.

Combinations of Green, Westbrook and Nenad Krstic -- without Durant -- are uniformly good. Meanwhile, it's hard to find any common combinations of players with Durant that stand out. An exception: When Durant plays with Green, Westbrook and Collison. Those four played nearly 700 minutes together during which they outscored opponents just a little, by 18.

One mitigating factor: Durant missed some games last season, including a very soft part of the schedule where the team rolled against weakling opponents like the Timberwolves, Kings and Wizards. So, while his teammates were looking good in +/-, Durant was holding constant.

Theory #2: Any player with those teammates would have a bad +/-.
Again, sound good, but his teammates manage to be part of many good lineups and player combinations.

Of particular interest here is Russell Westbrook. The Thunder are very good when Westbrook's on the court, and generally pretty bad when he's off. (Wayne Winston says this effect is so pronounced he could build a strong case that Westbrook should have been rookie of the year over Derrick Rose.)

How does Westbrook do it?

In some ways, he's the opposite of Durant. He is not a gaudy scorer -- in fact, he's a pretty bad shooter. He also has a high turnover rate, which is typically murder on team efficiency.

Yet Westbrook is a good defender. He draws a ridiculous number of fouls on both ends of the floor. He gets the team in the bonus, which in turn earns all of his teammates free throws. He rebounds well for his position, particularly at the offensive end, which can often lead to easy baskets.

Westbrook also has a statistical advantage over teammates like Durant and Green: He didn't start for the first 13 games last season under P.J. Carlesimo, when the Thunder were terrible. Carlesimo wanted to run and gun, but they were poorly equipped for the task. They lost badly, again and again.

By the time Westbrook started, the team was playing a more sensible style under Scott Brooks.

On the other hand, it's not like the first month of the season gave Durant wholly wacky +/- numbers. Winston keeps his +/- ratings per month. Durant had a -11 for the first month of the season, and then improved for three months in the middle (-6, -2, -7) before finishing about where he started (-10, -13).

Theory #3: It's hard to play with a superstar.
Just as it's undoubtedly a challenge for young Kevin Durant to perform the duties of a superstar, it may also be hard for his young teammates to know how to account for his abilities.

In other words, when Durant is benched, Westbrook, Green et al can make basketball decisions more or less as they have their entire basketball lives.

When Durant, superstar-in-the-making, is on the court, his teammates would presumably be aware of that. Hesitant jumpers, seams in the defense unexploited and shot-clock wasted finding Durant when he doesn't have the team's highest percentage shot ... you can find examples of all that on video.

This could explain why the Thunder are not just (far) worse on defense when Durant plays. They're also, surprisingly, slightly worse on offense.

Theory #4: Anyone that young, playing those kinds of minutes, is likely to see his team get outscored.
There may be something to this.

Kevin Durant has played 5,653 minutes in his first two NBA seasons. Good teams generally don't play 19- and 20-year-olds those kinds of minutes. In fact, bad teams don't often do it, either.

There may be an inevitability to his poor performance. Since the mid-1990s only a handful of players as young as Durant have played the kinds of minutes he has over two seasons -- and only LeBron James has more field goal attempts. Durant has been in an unusual situation that almost never produces good results until the player matures.

At the same age as Durant, some had better bottom line results than Durant -- LeBron James, Carmelo Anthony, Chris Bosh, Dwight Howard -- but they all had more veteran teammates. Young Kevin Garnett in Minnesota, on the other hand, had Thunder-style youth around him, and performed only slightly better than Durant (according to Win Shares) in terms of helping his team win while he was on the court.

But again, if you're in Kevin Garnett territory, the player Winston calls the best of the decade, in terms of adjusted +/-, it's hard to worry too much about Durant.

Kevin Durant, Great Player?
It shakes out as something of a debate between what we see with our eyes and what the team actually does. Based on how others have played at his age, it strikes me as likely that Durant will blossom precisely as predicted. But there's no denying that years one and two have proved little-to-nothing about his ability to help a team, despite his glossy reputation.

There is an undeniable reality that flashy scorers, like Durant, have long gotten the benefit of the doubt when it comes to assigning greatness. Scoring is considered, essentially, a "big thing" for a player to be able to do. Durant's good at it, too. Durant's effective field goal percentage last season was a healthy 51%, even as he shot a ton. Garnett, James, Kobe Bryant, Amare Stoudemire ... at the same age, they all made a lower percentage.

But if scoring is a big thing, what about the "little" things, like, say, defending the pick and roll, closing out on shooters and finding open teammates in rhythm? It could be that a young Durant has been doing the little things poorly enough that they overwhelm the big thing he does well.

Knowing that just about any NBA general manager would trade his own children for a prospect of Durant's caliber, I asked Winston if he'd advise his team to accept if the Mavericks were (in some alternate universe) offered Durant for free. "I'd say probably not," he replied. "I would not sign the guy. It's simply not inevitable that he'll make mid-career strides. Some guys do. But many don't, and he'd have to improve a lot to help a team."

And when I relayed Winston's comment to one of the NBA's most respected talent evaluators, his response was simply: "He's crazy."

Over the next few years, one of them will be proved wrong.

10-10-2009, 11:23 PM	#1
Fidel Diamond Member Join Date: Apr 2001 Posts: 3,283	Wayne Winston: I would not sign Durant. Stubled across this on espn. Winston is the Mavs stats guy. He basicly says that his +/- stats and analysis tell him Durant is a "terrible" player, his performance bottom 10 % of the league. Asked if he´d sign Durant if the Mavs could get him for free he says No. This is an interesting read. I don´t know how much we rely on Winston in player desicions but this is defenitly not a conventional approach. So what do you think. Are we screwed with Winston? Is he a genius? The Kevin Durant Conundrum October 9, 2009 3:25 PM There is almost no one like Kevin Durant. You can make a case he was the best freshman player in the history of the NCAA, and in the two years since he has only gotten better. He's 6-9. He moves like a gazelle. No one can stop him getting his shot. He's a tremendous teammate and an unbelievably hard worker. He also just turned 21, but has already spent two years feeling out the superstar role. Kevin Durant There is almost no surer superstar prospect in the NBA, and yet for some reason Kevin Durant's team plays better when he's on the bench. (Andrew D. Bernstein/NBAE via Getty Images) For players his age, his scoring is top-five all-time. Given that he's a half-decade from the age players typically start to peak, it's not hard to picture him becoming an MVP. The most respected single number to express a player's total contributions -- PER -- ranked him as the 20th best player last season, while projecting him to leap to fifth in what would be his senior year in college. The expectations are now approaching ridiculous; the future the Thunder have been grooming Kevin Durant for has made some cameos, and he just spent an off-season marinating in his legendary work ethic. NBA General Managers have selected Kevin Durant as the player most likely to have a breakout season, and he's everybody's pick to be the next player to join basketball's two most exclusive clubs: Team USA and the All-Star team. It's no surprise that Durant is on billboards and the covers of video games. (A surprise of the off-season was word that the Thunder won't be on national TV much; Durant has the trappings of a player the world will crowd around television sets to watch.) And yet, a shocking piece of news: The Thunder have, over the last two years, consistently performed worse than normal when Durant is on the floor. Any way you slice the +/- numbers, he's one of the Thunder's worst players. You read that correctly. Kevin Durant, uniformly regarded as an out-of-this-world NBA player, has been killing his team. Sometimes +/- can punish players simply for being on bad teams, but this is more than that. Mavericks' statistical expert Wayne Winston's in-depth lineup data shows that every one of Durant's key teammates -- Russell Westbrook, Jeff Green, Nenad Krstic, Nick Collison -- gets better, in many cases far better, results playing with less heralded teammates Thabo Sefolosha or Kyle Weaver while Durant sits. In fact, almost nobody on the Thunder has a +/- rating as poor as Durant's. Winston rates Durant's performance "in the lowest 10% of all NBA players." An Assault on Conventional Wisdom The first line of analysis, of any player, are real experts: People who assess talent for a living. Three out of three I talked to shrug. If you are trying to tell me that Kevin Durant is somehow a terrible player, they say, go ahead. But don't expect me to listen to you. They are unanimously emphatic that Durant is an absolute gem of a keeper. You can't teach size and mobility. It's hard to teach that kind of feel for getting the ball in the hole. There are times in games when nothing matters more than being able to reliably create your own shot and he already has that. The things he doesn't have -- and nobody denies they exist -- can and will be learned. Players who start their careers like Durant, and keep working, tend to improve dramatically, they say. Meanwhile, the things that make him inefficient -- mediocre passing, forcing some shots, turnovers, not making teammates better, and of course bad defense -- are all things that improve not just with work but also with better teammates. With the mere passage of time, thanks to an armada of draft picks, cap space, and high-potential young players to develop with him, these things will get much better on their own. Surrounded by better shooters, for instance, Durant would face fewer double teams and more assists, and less of a need to create for himself and turn the ball over. With better teammates, his team would score much more easily. On defense, meanwhile, veterans have a long track record of getting more consistent results -- the referees may be a factor, but probably even more important is the physical development that allows players in their prime to take nightly poundings and keep on fighting. Those three experts all had different theories to to dismiss the +/- numbers: Theory #1: Any player playing long minutes on a bad team would have a bad +/-. This makes sense at first. Think about +/-. You're on the court when the team gets slaughtered, and you get a bad +/-. Kevin Durant plays a lot. The Thunder have been killed a lot. There's nothing to see here, right? Kevin Durant and Russell Westbrook Russell Westbrook doesn't have nearly Durant's reputation, but he has had a much more positive effect on the Thunder. (Larry W. Smith, NBAE via Getty Images) As it happens, most Thunder players do have bad +/- numbers. But the Thunder perform particularly poorly when that one player is on the court. In fact, the Thunder has plenty of lineups that perform well. "In most situations," reports Winston, "holding some things constant and putting Durant in makes things worse, not better." The Thunder's big three, for instance, are Russell Westbrook, Kevin Durant, and Jeff Green. Last season, in the 387 minutes Westbrook and Green were on the floor without Durant, the team outscored opponents by a dozen points. In the 1,683 minutes all three of them were on the floor together, they were outscored by 245 points. Combinations of Green, Westbrook and Nenad Krstic -- without Durant -- are uniformly good. Meanwhile, it's hard to find any common combinations of players with Durant that stand out. An exception: When Durant plays with Green, Westbrook and Collison. Those four played nearly 700 minutes together during which they outscored opponents just a little, by 18. One mitigating factor: Durant missed some games last season, including a very soft part of the schedule where the team rolled against weakling opponents like the Timberwolves, Kings and Wizards. So, while his teammates were looking good in +/-, Durant was holding constant. Theory #2: Any player with those teammates would have a bad +/-. Again, sound good, but his teammates manage to be part of many good lineups and player combinations. Of particular interest here is Russell Westbrook. The Thunder are very good when Westbrook's on the court, and generally pretty bad when he's off. (Wayne Winston says this effect is so pronounced he could build a strong case that Westbrook should have been rookie of the year over Derrick Rose.) How does Westbrook do it? In some ways, he's the opposite of Durant. He is not a gaudy scorer -- in fact, he's a pretty bad shooter. He also has a high turnover rate, which is typically murder on team efficiency. Yet Westbrook is a good defender. He draws a ridiculous number of fouls on both ends of the floor. He gets the team in the bonus, which in turn earns all of his teammates free throws. He rebounds well for his position, particularly at the offensive end, which can often lead to easy baskets. Westbrook also has a statistical advantage over teammates like Durant and Green: He didn't start for the first 13 games last season under P.J. Carlesimo, when the Thunder were terrible. Carlesimo wanted to run and gun, but they were poorly equipped for the task. They lost badly, again and again. By the time Westbrook started, the team was playing a more sensible style under Scott Brooks. On the other hand, it's not like the first month of the season gave Durant wholly wacky +/- numbers. Winston keeps his +/- ratings per month. Durant had a -11 for the first month of the season, and then improved for three months in the middle (-6, -2, -7) before finishing about where he started (-10, -13). Theory #3: It's hard to play with a superstar. Just as it's undoubtedly a challenge for young Kevin Durant to perform the duties of a superstar, it may also be hard for his young teammates to know how to account for his abilities. In other words, when Durant is benched, Westbrook, Green et al can make basketball decisions more or less as they have their entire basketball lives. When Durant, superstar-in-the-making, is on the court, his teammates would presumably be aware of that. Hesitant jumpers, seams in the defense unexploited and shot-clock wasted finding Durant when he doesn't have the team's highest percentage shot ... you can find examples of all that on video. This could explain why the Thunder are not just (far) worse on defense when Durant plays. They're also, surprisingly, slightly worse on offense. Theory #4: Anyone that young, playing those kinds of minutes, is likely to see his team get outscored. There may be something to this. Kevin Durant has played 5,653 minutes in his first two NBA seasons. Good teams generally don't play 19- and 20-year-olds those kinds of minutes. In fact, bad teams don't often do it, either. There may be an inevitability to his poor performance. Since the mid-1990s only a handful of players as young as Durant have played the kinds of minutes he has over two seasons -- and only LeBron James has more field goal attempts. Durant has been in an unusual situation that almost never produces good results until the player matures. At the same age as Durant, some had better bottom line results than Durant -- LeBron James, Carmelo Anthony, Chris Bosh, Dwight Howard -- but they all had more veteran teammates. Young Kevin Garnett in Minnesota, on the other hand, had Thunder-style youth around him, and performed only slightly better than Durant (according to Win Shares) in terms of helping his team win while he was on the court. But again, if you're in Kevin Garnett territory, the player Winston calls the best of the decade, in terms of adjusted +/-, it's hard to worry too much about Durant. Kevin Durant, Great Player? It shakes out as something of a debate between what we see with our eyes and what the team actually does. Based on how others have played at his age, it strikes me as likely that Durant will blossom precisely as predicted. But there's no denying that years one and two have proved little-to-nothing about his ability to help a team, despite his glossy reputation. There is an undeniable reality that flashy scorers, like Durant, have long gotten the benefit of the doubt when it comes to assigning greatness. Scoring is considered, essentially, a "big thing" for a player to be able to do. Durant's good at it, too. Durant's effective field goal percentage last season was a healthy 51%, even as he shot a ton. Garnett, James, Kobe Bryant, Amare Stoudemire ... at the same age, they all made a lower percentage. But if scoring is a big thing, what about the "little" things, like, say, defending the pick and roll, closing out on shooters and finding open teammates in rhythm? It could be that a young Durant has been doing the little things poorly enough that they overwhelm the big thing he does well. Knowing that just about any NBA general manager would trade his own children for a prospect of Durant's caliber, I asked Winston if he'd advise his team to accept if the Mavericks were (in some alternate universe) offered Durant for free. "I'd say probably not," he replied. "I would not sign the guy. It's simply not inevitable that he'll make mid-career strides. Some guys do. But many don't, and he'd have to improve a lot to help a team." And when I relayed Winston's comment to one of the NBA's most respected talent evaluators, his response was simply: "He's crazy." Over the next few years, one of them will be proved wrong.