FiveThirtyEight data based Nets projection

[celticsclay]

I apologize if this has been posted, discussed elsewhere but haven't seen it. (Also haven't been around the forums much latetly). 538 did their stats based projection for the Nets and has them at 25 wins

https://fivethirtyeight.com/datalab/nba-brooklyn-nets-preview/

This has a lot of the same stuff the data based people have been saying for a while, and kind of runs in the face of the proud few that have maintained the Nets will be playoff contender. Kind of hard to argue this isn't an objective evaluation.

Note: The 25 wins projected are the same as they project for the 76ers

Looking forward to a few posters further digging in their heels...

[PhoSita]

I think their projections for the Nets are probably about 5 wins too low and the projections for the Celts are about 5 wins too high.

I'd love to be wrong on both fronts.

[Lucky17]

I'm still trying to wrap my head around the CARMELO tool, which is the basis for these team projections.

Brook Lopez' CARMELO profile says he is equally comparable to Keon Clark c. 2003 and Jeff Green c. 2014.

Also, good luck projecting Bojan Bogdanovic.

[celticsclay]

I think their projections for the Nets are probably about 5 wins too low and the projections for the Celts are about 5 wins too high.

I'd love to be wrong on both fronts.

You are just coming off as stubborn at this point to be honest. A few months ago you tried to put some numbers together to show the nets winning 32-34 games. I told you that didn't really jive with actual numbers or what anyone had out at that time. Here we are a few months later and one of the more renowned statistical projection models come out with them having 25 wins and you just say "meh, i think they win 5 more."

[tarheelsxxiii]

I'm still trying to wrap my head around the CARMELO tool, which is the basis for these team projections.

Brook Lopez' CARMELO profile says he is equally comparable to Keon Clark c. 2003 and Jeff Green c. 2014.

Also, good luck projecting Bojan Bogdanovic.

I wish 538 gave a little more detail on the models given it's a stats site.

It seems pretty simple though. They probably created a model using past data on players and teams over time and using them to predict win shares and whatever their other variable is (I think i recall RPM).

Their predictors are like draft order, physical attributes, and some hand picked offensive and defensive stats (%, chucking, DRPM). If your a stats person, it's just regressions...correlations with multiple variables

[Monkhouse]

I'm actually convinced the Nets won't even win 30 games or 25.

They look terrible. Young was the only one stepping up tonight, and when Sully was defending Lopez, you could tell how frustrated he was getting. He even at one point last night, ruffled Sully's hair and was mocking him, to which Sully put his hands up not wanting a fight.

If a player like Lopez can get so easily rattled by Sully, who was terrible defensively last night, I'm not expecting much out of them, considering Lopez and Johnson, is both of their best players, but were completely silent last night.

[Ilikesports17]

I think their projections for the Nets are probably about 5 wins too low and the projections for the Celts are about 5 wins too high.

I'd love to be wrong on both fronts.

That would put us at 43 wins and Brooklyn at 30?

I think that's about right.

Likewise I hope that I'm wrong. I think we grab the 7 seed and Brooklyn gives us a 7-12 range pick

[konkmv]

I think net are one injury away from giving us a top 3 pick...

[PhoSita]

I think their projections for the Nets are probably about 5 wins too low and the projections for the Celts are about 5 wins too high.

I'd love to be wrong on both fronts.

You are just coming off as stubborn at this point to be honest. A few months ago you tried to put some numbers together to show the nets winning 32-34 games. I told you that didn't really jive with actual numbers or what anyone had out at that time. Here we are a few months later and one of the more renowned statistical projection models come out with them having 25 wins and you just say "meh, i think they win 5 more."

Ah, see here I guess I missed the part where a statistical projection model in its first season of use is now the actual determining factor in how many games these teams are going to win.

Guess we don't even need to play the regular season.


The difference between 25 wins and 30 wins really isn't that large.  It will probably come down to injuries and how many opponents the Nets get to face in the last few months of the season that have already decided to pack it in.

Neither of which can really be accounted for in a statistical projection model.


For the record, I like what I've seen from CARMELO so far, but it reaches some questionable conclusions on a good number of players.

[sofutomygaha]

You're fivethirtyeight.com. Give us the confidence intervals! We can take it! We know you calculated them. We see them right there on the player projections. We know you propagated them!

It's a neat system. I love the way that they sidestep regression with age by turning every player forecast into a weighted aggregate of previous comparable players. The comparable players blow your mind, though, because the dimensions in which players are considered "comparable" are  different from what we'd tend to come up with when we are chatting about playing style and physical attributes.

I would think the confidence intervals would be humongous for young players and for superstars: (1) Young players because they are in flux and it's too early to pin comps on them, and (2) superstars because what comps you do find will either tend to be lesser players or all be drawn from the same very shallow pool.

The Nets look even more depressing in this light, because they are not young, nor star-packed. If this system was going to have anyone pegged, it would be Jack, Johnson, Lopez, and Young.

[sofutomygaha]

Did you guys notice how Amir Johnson, Sullinger, and Olynyk all project better for 2015 than Brook Lopez in this system?

[Lucky17]

Did you guys notice how Amir Johnson, Sullinger, and Olynyk all project better for 2015 than Brook Lopez in this system?

Ah, so CARMELO stands for "Celtics Are Really Magnificent; Especially Love Olynyk."

[celticsclay]

I think their projections for the Nets are probably about 5 wins too low and the projections for the Celts are about 5 wins too high.

I'd love to be wrong on both fronts.

You are just coming off as stubborn at this point to be honest. A few months ago you tried to put some numbers together to show the nets winning 32-34 games. I told you that didn't really jive with actual numbers or what anyone had out at that time. Here we are a few months later and one of the more renowned statistical projection models come out with them having 25 wins and you just say "meh, i think they win 5 more."

Ah, see here I guess I missed the part where a statistical projection model in its first season of use is now the actual determining factor in how many games these teams are going to win.

Guess we don't even need to play the regular season.


The difference between 25 wins and 30 wins really isn't that large.  It will probably come down to injuries and how many opponents the Nets get to face in the last few months of the season that have already decided to pack it in.

Neither of which can really be accounted for in a statistical projection model.


For the record, I like what I've seen from CARMELO so far, but it reaches some questionable conclusions on a good number of players.

See it is not just 538 though, it is the vegas posted win totals or any major sports forecaster that has released their projections. If EVERYONE is one side, and you are on the other, the onus is on you to try offer some sort of explanation of why you have a different opinion than everyone else. If I say I think Hilary Clinton has a good chance to be the democratic candidate, I don't really need a great explanation. If I say I think Lincoln Chaffee does, I do. You are Lincoln Chaffee in this argument.

[Eddie20]

Ah, the Nets and those that think they won't be horrible.

So the Nets who finished with a record of 38-44 lost the following players:
Williams
Anderson
Teletovic
Plumlee

and replaced them with:
Larkin
Ellington
Bargnani
Robinson

So these 4 will have to account for the following per game losses:
MPG 98.3
PPG 37.6
RPG 17.4
APG 9.8
3PG 3.9

Plus, they'll be leaning more on Lopez, who has missed 160 games the last 4 years, and 15 year veteran Joe Johnson, who came off a season where he averaged his lowest scoring total since his second season in the pros on just 43% from the field.

Vegas > A few C's Blog posters

[sofutomygaha]

Your Lincoln Chaffee in this argument