WoW review of Atlanta's Year

[hussamga]

I know alot of people dont like the statistical analysis/projections out there, but theres still some good reads (btw he has Horford as the most productive rookie statistically)

"So how much did atlanta improve.."

Basically our starting rotation of bibby/johnson/williams/smith/horford was slightly above league avg, which should project us out to around 50wins...however our bench outside of childress was horrid statistically, bringing us below this avg to his projected win total of 36wins vs our actual of 37.

[Admin]

I am going to make this a sticky throughout tomorrow so that everyone gets a chance to read it. I'm especially interested in hearing what statman has to say about Childress' WP48.

[hussamga]

Theres alot of analysis on that website if you take the time to read through it....methodical team by team analysis, from the 1/3 mark, 1/2 way, to the end of season.

Heres a link to his Top 15 players at each position post and Most Improved Player

Looking over the lists and our current theme of trading to improve the team...the RFA's we should target according to efficiency would have to be Biedrins and Calderon who both rank high, Mike Miller is also worth mentioning. Trying to get these players is another matter altogether...

Possibly Smith for Biedrins+Azubuke? And Marvin for Calderon? Then try to send Bibby off for veteran shooter and lesser expiring? Either way our team will be on the brink of luxury tax, so owners are going to have to come around.

Loss of Marvin W -2.5wins

Loss of Mike Bibby -2.4wins

Loss of Josh Smith -6.7wins

Gain of Biedriens +14.5wins

Gain of Calderon +16.0wins

Net gain of +30.5-11.6 = 18.9 wins = 56 win team next year? Yah i doubt we would be that good or it would translate like this, but fixing our PG production would go a long way, +10wins is reasonable IMO

[exodus]

I am not familiar with this win score stat so i looked at some of the rankings by position based on this guys analysis.

http://www.wagesofwins.com/15PG0708.html

Looking at the PGs he has Kidd as the second best pg in the league even though NJ missed the playoffs and Dallas struggled when they got him.

He also has Rondo ranked ahead of Andre Miller.

Looking at shooting guards

http://www.wagesofwins.com/15SG0708.html

he has JJ ranked 11th behind guys like Ronnie Brewer and Anothony Parker.

Carlos Delfino is ranked ahead of Richard Hamilton and Wade.

For small forwards

http://www.wagesofwins.com/15SF0708.html

he has Childress ranked ahead of Camelo.

The results are self explanatory.

[adrian]

Does biedrens really fit what we are trying to do if miami doesent get rose they will probably throw around 8 mil at calderon whats up with all the get rid of smoove crap all of a sudden a good smoove trade would be mo williams and bogut to atl. and smoove and chills or marv to milwauke villanueva would help us also in to years or so we would have the best frontcourt in the leugue the we could look to move bibby in a deadline deal

[DrReality]

Quote:

---

I know alot of people dont like the statistical analysis/projections out there, but theres still some good reads (btw he has Horford as the most productive rookie statistically)

"So how much did atlanta improve.."

Basically our starting rotation of bibby/johnson/williams/smith/horford was slightly above league avg, which should project us out to around 50wins...however our bench outside of childress was horrid statistically, bringing us below this avg to his projected win total of 36wins vs our actual of 37.

I'll say it again: we have more talent on the bench than what was utilized. Solo, Salim, West can all be utilized in roles that help give us depth. Almost every team in the league utilizes players like this. Woody is horrible.

[Guest]

"He" isn't ranking him, his formula is. He regressed wins over certain stats to calculate linear weights and from this he simply creates a formula of .xx*Points + .yy*Rebounds... and this gives you the win score. So if you believes statistics gives you a legitimate answer for certain things then you should at least agree with his methods of figuring out wins.

[TheDude]

i would have liked to see more comments on there regarding the CHANGE in the Hawks once Bibby arrived. I mean it's not like he's a forward, he's the starting point guard which is a HUGE change for a team.

[exodus]

Quote:

---

"He" isn't ranking him, his formula is. He regressed wins over certain stats to calculate linear weights and from this he simply creates a formula of .xx*Points + .yy*Rebounds... and this gives you the win score. So if you believes statistics gives you a legitimate answer for certain things then you should at least agree with his methods of figuring out wins.

I am a little suspect of any formula that has Delfino ranked ahead of Wade.

[JackB1]

great read and I agree with most everything.

the major reason for our improvement was #15, plain and simple.

the major reason we underachieved was our weak bench.

One of our priorities over the off season has got to be bolstering our bench.

[adrian]

That is the most bull sh!t statistic I have ever seen. We do not need to make blockbuster deal unless we vastly underachieve next year. For now we just need to add a few of many good available bench guys.

[hussamga]

I know the statistic/ranking favors rebounding (possessions) and efficiency over scoring total...but funny thing is it generally works....at the beginning of the season or 1/3 through i remember his stat projections had us at 35-37 wins...

Table of Atlanta Production Dec27 has us at 37.8 wins

I think there are some things you could dismiss, like relative rankings of players...but the body of work is pretty strong.

Someone mentioned New Jersey, he wrote a post about how RJ has regressed greatly over last 2years in his efficiency.

He Review of the Bibby Trade at the deadline... basically saying his shooting efficiency is slightly above avg, but he doesnt round out the rest of stats and turns it over more than AJ etc...

About Joe Johnson, at the all star break he had him only producing 2.2wins, at end of season he has him at 6.9 wins...the improved play after the all star break a stark contrast every hawk fan noticed. If Joe had played like this all year round he would be around 12wins and our most efficient player and #3 on that list of top15 sg's behind Kobe and Manu...Exactly where he belongs

[Admin]

Sounds like Sabermetrics for basketball. This guy might really be on to something. It's interesting because if we simply replace one of our negative performers for someone who can have an impact similar to Childress or Al we would be among the top 4 teams in the east according to these stats.

[hussamga]

The way i see his review...we dont have to change the starters. Marvin has been improving over the years, where he hurt us alot his rookie year now he is slightly helping. Joe playing like he did post all star break is an extra 6 wins alone. Add a couple positive play bench members and we could easily be a 45-50 win team. It also shows that our weakest starters are at PG and SF, im sure most will agree with and that our bench is horrible sans childress. Also shows the regression of Pachulia.

Basically he confirms from a data point of view everything we saw. The data doesnt care if we add a shooter or slasher to the bench, just that the guy is efficient with his possessions.

[bronnt]

Quote:

---

Sounds like Sabermetrics for basketball. This guy might really be on to something. It's interesting because if we simply replace one of our negative performers for someone who can have an impact similar to Childress or Al we would be among the top 4 teams in the east according to these stats.

I've been looking at this site rather infrequently this year-mostly I was watching to keep track of their take on the ROtY-but seeing this now confirms something I've believed pretty much all season: Drafting Al Horford was the main reason this team got better. And it's nto really that tough to fathom because he's a legitimate post man who can defend and rebound, and he scores efficiently when called upon. He also proved to be capable of handling the ball and a good passer. We've seen run the fast break on occasion. If he experiences some improvement as a player (and is used more) this team is going to be really good.

I'm feeling starting to feel hopeful about the Hawks' chances to improve next year even by keeping a poor coach in charge. We need to give more minutes to Acie Law off the bench and let him improve as a player. He should be a guy who has a positive effect on the team, and we've certainly seen flashes of it at times, even though his playing time has been limited. As a rookie, it's just damn pessimistic to doubt that he'll show improvement in his second year, and if he improves while playing more minutes, it should help the poor showing by the Hawks' bench. Also, JJ played really poorly in the first half compared to how well he's played over the past two seasons. He played really well in the second half after we acquired Bibby. If Bibby can get back to shooting like he did for us in the regular season, he'll continue to spread the floor for JJ.

And one other thing I'm hopeful about is David Andersen. We still don't have any other shooters to spread the floor-Salim is inconsistent and underused. Acie isn't a shooter, really, and Marvin is inconsistent enough to be a non-factor at times. But if David Anderson can shoot around 40% from three in the NBA, he could be the other shooter off the bench we need to spread the floor. And he has the added benefit of being able to play inside. Even if he's not really up to NBA standards for a player who can contribute in the post, he might be an improvement over Zaza. He'll add some versatility to the line-up.

[Plainview1981]

I don't know if you will ever be able to prove anything through these type of stats. There always seems to be a hole somewhere.

[frosgrim]

The problem is that the data probably doesn't meet the normality assumptions. He should have used a Poison regression due to the non-linear nature of the model. When you try to squeeze in non-linear data into something like OLS you get garbage results. He also needs to present to the AIC calculations so we can actually compare the effectiveness of the modeling.

Bottom-line. Bad statistical methodology.

[Admin]

How can it be bad if it's an extremely accurate predictor? Isn't that the bottom line to analyzing statistics?

[LOL_Hawks]

Quote:

---

I know alot of people dont like the statistical analysis/projections out there, but theres still some good reads (btw he has Horford as the most productive rookie statistically)

"So how much did atlanta improve.."

Basically our starting rotation of bibby/johnson/williams/smith/horford was slightly above league avg, which should project us out to around 50wins...however our bench outside of childress was horrid statistically, bringing us below this avg to his projected win total of 36wins vs our actual of 37.

So it's basically telling us something we already know: Our starting line-up is ok, our bench, aside from Chill, is one of the worst in the league. So now we know exactly what to fix.

[frosgrim]

I was refering to the stats that have Carmlo rated below Childress, not the other regression analysis. Sorry for the confusion.