2016-17 MVP Discussion Thread Pt 2

[laika]

Speaking of actual value, here are the stats that are most important.

Earlier I was referring to On/Off ratings, but called it Net rating because 82games calls it Net on their site. It's a little confusing because NBA.com calls Oncourt "+/-" and Team Offensive Rating minus Team Defensive Rating "NetRtg". As far as I know, no site shows current On/Off numbers. You have to compile them. Here are the current leaders-

I don't know where you are compiling these from but they are readily available and adjusted for pace in Basketball Reference on/off.

Here is a link to Curry's who is +17.5/100 poss on, +2.8/100 off for a net rtg of +14.7/100 poss.
http://www.basketball-reference.com/players/c/curryst01/on-off/2017

Almost all of these are well off BBRef

They are from NBA.com. NBA.com, basketballreference, espn and probably every other site have the same discrepancy between point differential and Ortg minus Drtg.

Here is the link at nba.com showing exactly the same values you would find at basketballreference.
http://stats.nba.com/team/#!/1610612744/onoffcourt-advanced/

So which one should we use?
I care about actual score differential, not some formula that approximates things. I am extremely confident that if I added up the actual scores for the year then point differential would be the correct one.

One possible source of the discrepancy is that the Warriors use 103 possessions per game. But if you adjust it to 100 possessions then that only closes part of the gap. On a related note, per possession is a terrible way to judge players. If a team has the same efficiency but uses twice as many possessions then they are going to win by twice as much. Per minute is the obvious choice when comparing players, with a moderate adjustment for high minute vs low minute players.

[laika]

Laika, you made a great post by compiling the on-off. Kudos for that.

Then you started using words like 'clear', 'way ahead', 'it's not even close'. Often these are the words that will turn a great post into a mediocre one because they hurt the credibility of a stance, if not used properly.

Well that depends on philosophical differences.

If you are using only On/Off stats then Curry is way ahead. In theory properly adjusted On/Off is the definitive stat, with absolutely no reference to box score stats needed to accurately determine how valuable a player is. Of course in practice no one has ever been able to produce adjusted On/Off numbers that don't have some weird and unlikely values. So people combine the two because the combined values seem to be more sensible. But no one knows precisely why the statistical On/Off adjustment models all seem to be getting the "real" values wrong. And no one really knows the precise weighting we should give to box score vs On/Off to get the most accurate results.

A couple factors in favor of my interpretation of On/Off though- The more minutes a player has played then the easier it is to adjust the raw On/Off values. Curry has played a decent number of minutes. The highest and lowest value on each team are the easiest to adjust for since there are much fewer confounding factors. Curry has easily the highest On/Off on the Warriors. The more a player has played with a variety of players then the more accurate the raw On/Off number is. The Warriors don't really use a platoon system like the Clippers or Wizards do. By playing with a variety of players it makes it more likely that Curry is actually extremely valuable like his raw On/Off shows.

[HurricaneKid]

Speaking of actual value, here are the stats that are most important.

Earlier I was referring to On/Off ratings, but called it Net rating because 82games calls it Net on their site. It's a little confusing because NBA.com calls Oncourt "+/-" and Team Offensive Rating minus Team Defensive Rating "NetRtg". As far as I know, no site shows current On/Off numbers. You have to compile them. Here are the current leaders-

I don't know where you are compiling these from but they are readily available and adjusted for pace in Basketball Reference on/off.

Here is a link to Curry's who is +17.5/100 poss on, +2.8/100 off for a net rtg of +14.7/100 poss.
http://www.basketball-reference.com/players/c/curryst01/on-off/2017

Almost all of these are well off BBRef

They are from NBA.com. NBA.com, basketballreference, espn and probably every other site have the same discrepancy between point differential and Ortg minus Drtg.

Here is the link at nba.com showing exactly the same values you would find at basketballreference.
http://stats.nba.com/team/#!/1610612744/onoffcourt-advanced/

So which one should we use?
I care about actual score differential, not some formula that approximates things. I am extremely confident that if I added up the actual scores for the year then point differential would be the correct one.

One possible source of the discrepancy is that the Warriors use 103 possessions per game. But if you adjust it to 100 possessions then that only closes part of the gap. On a related note, per possession is a terrible way to judge players. If a team has the same efficiency but uses twice as many possessions then they are going to win by twice as much. Per minute is the obvious choice when comparing players, with a moderate adjustment for high minute vs low minute players.

I'm rather shocked at how disparate the NBA.com and BBRef stats are. The NBA ones you are using are ALSO /100 possessions FYI.

BBref has GSW at 116.6 ORTG, NBA has them at 113.8. That is a chasm.

Does anyone have an explanation for this chasm?

[bondom34]

I don't know where you are compiling these from but they are readily available and adjusted for pace in Basketball Reference on/off.

Here is a link to Curry's who is +17.5/100 poss on, +2.8/100 off for a net rtg of +14.7/100 poss.
http://www.basketball-reference.com/players/c/curryst01/on-off/2017

Almost all of these are well off BBRef

They are from NBA.com. NBA.com, basketballreference, espn and probably every other site have the same discrepancy between point differential and Ortg minus Drtg.

Here is the link at nba.com showing exactly the same values you would find at basketballreference.
http://stats.nba.com/team/#!/1610612744/onoffcourt-advanced/

So which one should we use?
I care about actual score differential, not some formula that approximates things. I am extremely confident that if I added up the actual scores for the year then point differential would be the correct one.

One possible source of the discrepancy is that the Warriors use 103 possessions per game. But if you adjust it to 100 possessions then that only closes part of the gap. On a related note, per possession is a terrible way to judge players. If a team has the same efficiency but uses twice as many possessions then they are going to win by twice as much. Per minute is the obvious choice when comparing players, with a moderate adjustment for high minute vs low minute players.

I'm rather shocked at how disparate the NBA.com and BBRef stats are. The NBA ones you are using are ALSO /100 possessions FYI.

BBref has GSW at 116.6 ORTG, NBA has them at 113.8. That is a chasm.

Does anyone have an explanation for this chasm?

They calculate possessions differently.
I used to think NBA.com was more accurate but apparently I was incorrect, BBR is the more accurate data.

[K_chile22]

They are from NBA.com. NBA.com, basketballreference, espn and probably every other site have the same discrepancy between point differential and Ortg minus Drtg.

Here is the link at nba.com showing exactly the same values you would find at basketballreference.
http://stats.nba.com/team/#!/1610612744/onoffcourt-advanced/

So which one should we use?
I care about actual score differential, not some formula that approximates things. I am extremely confident that if I added up the actual scores for the year then point differential would be the correct one.

One possible source of the discrepancy is that the Warriors use 103 possessions per game. But if you adjust it to 100 possessions then that only closes part of the gap. On a related note, per possession is a terrible way to judge players. If a team has the same efficiency but uses twice as many possessions then they are going to win by twice as much. Per minute is the obvious choice when comparing players, with a moderate adjustment for high minute vs low minute players.

I'm rather shocked at how disparate the NBA.com and BBRef stats are. The NBA ones you are using are ALSO /100 possessions FYI.

BBref has GSW at 116.6 ORTG, NBA has them at 113.8. That is a chasm.

Does anyone have an explanation for this chasm?

They calculate possessions differently.
I used to think NBA.com was more accurate but apparently I was incorrect, BBR is the more accurate data.

And why is that

[bondom34]

I'm rather shocked at how disparate the NBA.com and BBRef stats are. The NBA ones you are using are ALSO /100 possessions FYI.

BBref has GSW at 116.6 ORTG, NBA has them at 113.8. That is a chasm.

Does anyone have an explanation for this chasm?

They calculate possessions differently.
I used to think NBA.com was more accurate but apparently I was incorrect, BBR is the more accurate data.

And why is that

TBH I forget . I know it was posted on APBR and the PC board, iirc it is something involving how offensive rebounds are credited. One counts an extra possession somewhere, I forget which.

Edit: http://www.apbr.org/metrics/viewtopic.php?f=2&t=9231&hilit=possessions+nba.com

Per Kevin Pelton:

As explained in the above link, using the stats.nba.com formula overestimates the number of possessions because it doesn't account for team offensive rebounds. In an individual game, that (as well as the estimate for free throw possessions that any estimator uses) mean the net ratings can get a little wonky. Over larger samples, they're likely to even out.

[K_chile22]

They calculate possessions differently.
I used to think NBA.com was more accurate but apparently I was incorrect, BBR is the more accurate data.

And why is that

TBH I forget . I know it was posted on APBR and the PC board, iirc it is something involving how offensive rebounds are credited. One counts an extra possession somewhere, I forget which.

Edit: http://www.apbr.org/metrics/viewtopic.php?f=2&t=9231&hilit=possessions+nba.com

Per Kevin Pelton:

As explained in the above link, using the stats.nba.com formula overestimates the number of possessions because it doesn't account for team offensive rebounds. In an individual game, that (as well as the estimate for free throw possessions that any estimator uses) mean the net ratings can get a little wonky. Over larger samples, they're likely to even out.

Thanks!

[Hero]

Exactly why Curry should be in the top-3:

Because his very presence warps and bends defenses in ways never seen before on an NBA court, the Warriors play out of their minds when he’s on the floor. With Curry off the floor, they have a respectable +3.0 net rating. When Curry is on the floor, however, the Warriors have a +16.8 net rating, yielding an on/off split of 13.8 points per 100 possessions. Considering the fact that the Warriors have three of the best 10 and four of the best 20 players in the NBA, it’s remarkable that any one player can produce such a monstrous swing in production.

While Durant is certainly the missing piece, his output is also at its best when he shares the floor with Curry. When Durant plays without Curry, the Warriors outscore their opponents by just 1.1 points per 100 possessions. When Curry plays sans Durant, the Warriors outscore their opponents by 9.8 points. However, when the duo play together, the Warriors outscore their opponents by a remarkable 21.1 points per 100 possessions.

He's the most valuable player on a historically great team.

You GSW fans need to make up your mind. Just a few days ago you all were clamoring for KD to get consideration. Now it's Curry...

[PeptoKlepto]

Exactly why Curry should be in the top-3:

Because his very presence warps and bends defenses in ways never seen before on an NBA court, the Warriors play out of their minds when he’s on the floor. With Curry off the floor, they have a respectable +3.0 net rating. When Curry is on the floor, however, the Warriors have a +16.8 net rating, yielding an on/off split of 13.8 points per 100 possessions. Considering the fact that the Warriors have three of the best 10 and four of the best 20 players in the NBA, it’s remarkable that any one player can produce such a monstrous swing in production.

While Durant is certainly the missing piece, his output is also at its best when he shares the floor with Curry. When Durant plays without Curry, the Warriors outscore their opponents by just 1.1 points per 100 possessions. When Curry plays sans Durant, the Warriors outscore their opponents by 9.8 points. However, when the duo play together, the Warriors outscore their opponents by a remarkable 21.1 points per 100 possessions.

He's the most valuable player on a historically great team.

You GSW fans need to make up your mind. Just a few days ago you all were clamoring for KD to get consideration. Now it's Curry...

Well, they both should get consideration. But Curry is the most important/valuable player on this team.

[makaveli_99]

Is there any way that curry/kd can win MVP. What if they finish the season with 13+ less losses than hou/cle/okc. I know that is extremely unlikely, but what if..

current loss chart:
gs 7
cle 15
hou 17
okc 22

[Fico92]

Exactly why Curry should be in the top-3:





He's the most valuable player on a historically great team.

You GSW fans need to make up your mind. Just a few days ago you all were clamoring for KD to get consideration. Now it's Curry...

But Curry is the most important/valuable player on this team.

No, he is not. For regular season wins, maybe. KD is the one who will be the key player against the Cavs, and take home the Finals MVP. If you value regular season wins more than the championship, then sure, Curry might be more important and valuable to the Warriors. Highly doubt you see it that way, though.

[PeptoKlepto]

You GSW fans need to make up your mind. Just a few days ago you all were clamoring for KD to get consideration. Now it's Curry...

But Curry is the most important/valuable player on this team.

No, he is not. For regular season wins, maybe. KD is the one who will be the key player against the Cavs, and take home the Finals MVP. If you value regular season wins more than the championship, then sure, Curry might be more important and valuable to the Warriors. Highly doubt you see it that way, though.

I don't care if we're playing in the finals or backyard basketball, Steph Curry is the engine and the fire starter of this team.

[Fico92]

But Curry is the most important/valuable player on this team.

No, he is not. For regular season wins, maybe. KD is the one who will be the key player against the Cavs, and take home the Finals MVP. If you value regular season wins more than the championship, then sure, Curry might be more important and valuable to the Warriors. Highly doubt you see it that way, though.

I don't care if we're playing in the finals or backyard basketball, Steph Curry is the engine and the fire starter of this team.

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

[TheGOATRises007]

No, he is not. For regular season wins, maybe. KD is the one who will be the key player against the Cavs, and take home the Finals MVP. If you value regular season wins more than the championship, then sure, Curry might be more important and valuable to the Warriors. Highly doubt you see it that way, though.

I don't care if we're playing in the finals or backyard basketball, Steph Curry is the engine and the fire starter of this team.

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

Said this a few weeks ago and it's still true.

I see Durant having a monster finals and people will start to ask if he's better than Lebron currently.

[Hero]

I don't care if we're playing in the finals or backyard basketball, Steph Curry is the engine and the fire starter of this team.

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

Said this a few weeks ago and it's still true.

I see Durant having a monster finals and people will start to ask if he's better than Lebron currently.

People? Yeah maybe a few homers and GSW fans.

Nobody in their right mind will broach such a conversation even if he wins 3 straight.

[TheGOATRises007]

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

Said this a few weeks ago and it's still true.

I see Durant having a monster finals and people will start to ask if he's better than Lebron currently.

People? Yeah maybe a few homers and GSW fans.

Nobody in their right mind will broach such a conversation even if he wins 3 straight.

Lebron isn't a robot.

He won't be the best player in the league forever.

If KD outplays him in the final, it can be argued at the moment. Key word being: at the moment.

[Starboy]

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

Said this a few weeks ago and it's still true.

I see Durant having a monster finals and people will start to ask if he's better than Lebron currently.

People? Yeah maybe a few homers and GSW fans.

Nobody in their right mind will broach such a conversation even if he wins 3 straight.

Lol what? KD is having a better regular season than Lebron. If Warriors win the finals and KD plays on this level in playoffs, it's not really a conversation. The only argument for Lebron right now is what he did over the course of 3 games 7 months ago. Big stage and all that, so I'm keeping him at #1 for now.

[PeptoKlepto]

No, he is not. For regular season wins, maybe. KD is the one who will be the key player against the Cavs, and take home the Finals MVP. If you value regular season wins more than the championship, then sure, Curry might be more important and valuable to the Warriors. Highly doubt you see it that way, though.

I don't care if we're playing in the finals or backyard basketball, Steph Curry is the engine and the fire starter of this team.

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

Via ESPN:

“I just said to him, ‘Don’t worry about me,'” Durant told ESPN of his ongoing dialogue with Curry. “I said, ‘Just play your game. I’ll figure it out. I’ll figure it out around you. You’re the engine of this team, and I know that. I’m not trying to come over and feel like everything just revolves around me. Just do you, man. I’m going to play around you. I’ve played this game long enough. I know how to score. I know how to find the ball. Just go out there and play your game.’ And that’s what he’s been doing."

[B-easy]

Westbrook and Harden need to start winning, otherwise Kawhi or Durant will win MVP.

[righterwriter]

I don't care if we're playing in the finals or backyard basketball, Steph Curry is the engine and the fire starter of this team.

Last year, absolutely. Durant took over that and will be the reason you win, and take FMVP.

Via ESPN:

“I just said to him, ‘Don’t worry about me,'” Durant told ESPN of his ongoing dialogue with Curry. “I said, ‘Just play your game. I’ll figure it out. I’ll figure it out around you. You’re the engine of this team, and I know that. I’m not trying to come over and feel like everything just revolves around me. Just do you, man. I’m going to play around you. I’ve played this game long enough. I know how to score. I know how to find the ball. Just go out there and play your game.’ And that’s what he’s been doing."

It really is frightening to think that this team is still figuring itself out. Makes you wonder if the other guys will take on some fine tuning and improve their games, as well (Klay, Iggy, and Draymond can all play better, I believe). If that happens then what we are seeing now might pale in comparison to what will come later.