Wages of Wins = Anti-Diesel

[lethalweapon3]

That Chillz article by Billy... er... WoW, for those that didn't see it. Just another log to add the fire:

The Steal of Free Agency: Josh Childress

http://wagesofwins.c...josh-childress/

Childress could score, was an elite player and in his last season he was on the only team that gave the Boston Celtics — the eventual champions — a problem in the playoffs.

Childress has been a great player his whole career.

Before he left for Europe, Childress was a monster. In fact, given the progression of the Hawks, had they kept him around, they might have gotten even farther in the playoffs.

WoW.

~lw3

Edited August 16, 2012 by lethalweapon3

[High5]

That website is a hoot.

[AHF]

Advanced stats are only cool when they support people's preconceived notions. The same stat will tell you that Smoove led the league in defensive winshares and people will say "Hell yea! That sounds absolutely right!" but then they'll tell you that Marvin led the team in overall winshares per 48 and it's "What kind of bull shit stat is this?!" Moral of the story, don't be hypocrites.

For me, I find value in advanced metrics and they definitely have their place but I'm not going to stand by and watch them be violently misused either. When you look at Marvin's advance stats, you can reconcile a guy who #2 on the team in WS/48 and #4 in PER with the play you see on the court by noting that Marvin wasn't asked to handle the ball (low turnovers which increases both WS/48 and PER), wasn't asked to score on his own or take contested shots (higher shooting % which increases WS/48 and PER on similar shot volumes), and who played limited minutes (able to maximize effort during his time on the floor which increases per minute production). You can see that his limited role is why he was high in WS/48 but was only fifth in total WSs (behind four players with worse WS/48). When you look at the total picture - including the advanced metrics - you see a productive but limited role player. You don't see a star or the best player on the team and anyone using a single metric to arrive at that conclusion isn't using the number validly in much the same way that anyone who says "the guy who averages the most pp36 is the best player in the league" isn't using pp36 in a valid way (for reference, Brook Lopez was #4 overall in the NBA in pp36 and Patrick Mills was #10). Edited August 16, 2012 by AHF

[MaceCase]

You're going too deep into the analysis, AHF. Obviously more sensible people like you and I can delve into numbers and parse them out over numerous metrics but too often with most if a stat confirms their bias it's useful and if it doesn't it's not.My argument is don't take any stat by itself at face value be they negative or positive, this article proclaiming Marvin to be elite is as equally bad as the ones that proclaimed Smoove the best defender and Al at an elite level because they all are based off the same WS metric.

[Guest]

I always laugh when I see "I don't trust stats, I use my eyes". Really? Then how the f--- do you know who won? I have suggested this before, but watch a game without the sound or a score on screen. Then I want you to answer me "who won at the end of the 1st, 2nd, and 3rd quarter" (the 4th quarter is obvious). If that is too easy, then start telling me who won each individual quarter. You quickly start to realize you need the score because your qualitative assessment is not always right.People resisting statistical analysis are the equivalent of Luddites. Maybe it is because of innumeracy that they refuse to delve in? However, it is clear that to make basketball decisions you need analytics in order to drive the quantitative side of things. Hearing all the qualitative arguments of "Lebron is physically better than Marvin" is all fine and dandy, but don't we want to know *how much better* Lebron is than Marvin? Of course we do, from the basketball perspective we do not have binary decisions of "pick Lebron or Marvin for your team". What we have are contracts that teams offer to Lebron based upon quantitative evidence and contracts that teams offer to Marvin based upon quantitative evidence that the GMs are handing out based upon multitudes of information. It is not likely that every team has the same contract offer in mind for Marvin and that is where the beauty of analytics comes in to play (with salary cap restrictions, clearly all teams have the same contract in mind for Lebron).EDIT: I said teams base contracts on quantitative evidence...well hopefully they do in this day and age.

Edited August 16, 2012 by hawksfanatic

[HawkItus]

I always laugh when I see "I don't trust stats, I use my eyes". Really? Then how the f--- do you know who won? I have suggested this before, but watch a game without the sound or a score on screen. Then I want you to answer me "who won at the end of the 1st, 2nd, and 3rd quarter" (the 4th quarter is obvious). If that is too easy, then start telling me who won each individual quarter. You quickly start to realize you need the score because your qualitative assessment is not always right. People resisting statistical analysis are the equivalent of Luddites. Maybe it is because of innumeracy that they refuse to delve in? However, it is clear that to make basketball decisions you need analytics in order to drive the quantitative side of things. Hearing all the qualitative arguments of "Lebron is physically better than Marvin" is all fine and dandy, but don't we want to know *how much better* Lebron is than Marvin? Of course we do, from the basketball perspective we do not have binary decisions of "pick Lebron or Marvin for your team". What we have are contracts that teams offer to Lebron based upon quantitative evidence and contracts that teams offer to Marvin based upon quantitative evidence that the GMs are handing out based upon multitudes of information. It is not likely that every team has the same contract offer in mind for Marvin and that is where the beauty of analytics comes in to play (with salary cap restrictions, clearly all teams have the same contract in mind for Lebron). EDIT: I said teams base contracts on quantitative evidence...well hopefully they do in this day and age.

Avanced statistics. I don't think anybody is railing against scores and rebounds etc. My thing is before the advanced statistics boom began the NBA was just fine in the 80's. People are looking for an edge and that is great. As a fan though I don't need or want it. If that floats your boat have at it.

[AHF]

Physically better is also a tough standard to articulate. Who were more physically gifted? Reggie Miller or Isiah Rider? Larry Bird or Tracy McGrady? John Stockton or Dee Brown? Mark Price or Nate Robinson? etc.Guys like Darko always look good in terms of length and physical gifts but nobody relies on that alone as a proxy for greatness as a player.On the normal box score statistics, that is how a lot of GMs ended up paying too much for one-dimensional scorers. Nobody has to used the advanced statistics but I think any team not using them is leaving out a very important tool.

Edited August 16, 2012 by AHF

[Guest]

Avanced statistics. I don't think anybody is railing against scores and rebounds etc. My thing is before the advanced statistics boom began the NBA was just fine in the 80's. People are looking for an edge and that is great. As a fan though I don't need or want it. If that floats your boat have at it.

This sounds like innumeracy. So you are OK with the statistics you can understand (raw #s, FG%, even eFG%), but as soon as they become too confusing/twisted to understand they are...worthless? I probably set up a straw man with the first paragraph of my last post, but these advanced statistics can provide solid foundation if they are grounded with reason. I do not disagree there are some advanced statistics out there that are worthless (NBA Efficiency Rating, PER and most of RealGM stands out in particular) but that is because of their foundations. These have ad hoc weights in combining raw data. They are not worthless because they are advanced statistics. Something like wins produced actually has a great foundation for its statistic, it is grounded in finding what is correlated with wins and then reconstructing these weights to create a normalized value to then compare across players. That is all fine, it makes sense and it is hard to argue with the method for its particular use (you can argue about implementation of estimating these weights as a problem, but first explain to me what they are currently doing). Part of the reason why WoW is a crappy blog is that they hail the statistic as the holy grail, which its not. However, that does not mean wins produced is worthless it is just that its creator (and his minions) don't know how to correctly use this statistic. Maybe it is fair to say that most people who blog using advanced statistics are worthless. This is true, part of the issue is that if someone was blogging in a way that was using advanced analytics in a creative/new/solid way then they wouldn't be blogging long...soon enough an NBA team would gobble them up to work for them.

[Sothron]

Avanced statistics. I don't think anybody is railing against scores and rebounds etc. My thing is before the advanced statistics boom began the NBA was just fine in the 80's. People are looking for an edge and that is great. As a fan though I don't need or want it. If that floats your boat have at it.

My problem with the advanced stats people is not that I don't understand the methodology I just refuse to accept it has any basis on reality. Who hit the most three point shots in the third quarter on a full moon? Why that's a great way to judge every player in the NBA because on those full moon games in the third quarter THAT GUY is the best player in the NBA. Dare to doubt me? Why I have this handy "advanced" stat to show you. Haha Luddite! Marvel at my Bill James impersonation for the NBA! The eye test by someone who knows the game is enough for me. I don't need "win shares" or "defensive win shares" to tell me whose more valuable. eFG? TS? Now those are some advanced stats that have some legs. But the vast majority of them are random and arbitrary compilations designed to derive at a conclusion and not a true evaluation of the game itself. If the game was really just who had the best PER or who had the best metrics in the pre draft combine then every center with wing span and reach would be awesome, every wing player with a good shuttle speed or vertical would be the next MJ.

[Guest]

My problem with the advanced stats people is not that I don't understand the methodology I just refuse to accept it has any basis on reality. Who hit the most three point shots in the third quarter on a full moon? Why that's a great way to judge every player in the NBA because on those full moon games in the third quarter THAT GUY is the best player in the NBA. Dare to doubt me? Why I have this handy "advanced" stat to show you. Haha Luddite! Marvel at my Bill James impersonation for the NBA!

You are referring to a conditional statistic, which may or may not be advanced (your specific example is not advanced). Not all advanced statistics are conditional. These advanced statistics are meant to enhance your knowledge of the game and correct mistakes your eye test makes. When you immediately duck your head into the sand and refer to the "eye test" you are showing Luddite tendencies. Are these advanced statistics supposed to replace human scouting? f*** no, only a Luddite would reach that conclusion to then bash the engine of growth. Analytics enhance scouting, they also make scouting more efficient and help predict players' performance better. It is a fallacy to think that "statheads" only go by their numbers. Some do, but the good ones look at the advanced statistics as more knowledge and use this to get a better grasp on the situation at hand. If you want to criticize an advanced statistic, first explain to me what the statistic is capturing and how it is calculated. Then you can start your criticism.

[PSSSHHHRRR87]

http://www.peachtreehoops.com/2012/8/27/3268851/atlanta-hawks-evaluated-by-wages-of-wins-revisited

On August 15th I posted about an article on wagesofwins.com evaluating the Hawks off-season. Not being familiar with their site I was unable to locate the stats that were the basis of their article (in my defense the links on the page take you to basketball-reference.com). Vivek, the author of the article, has kindly provided me a link to their stats, so I thought I should revisit the article and look at the Hawks summer through the prism of their stat system. Also, Vivek was obviously using stats he believes in as the basis of his opinion and is not a Marvin Williams stalker. I hope he understands that any system that declares Marvin Williams as the best Hawks player in going to be met with a truckload of skepticism from anyone what watches more than a handful of Atlanta games. Let start the review with Marvin Williams. Key comments in the original article were "best player from last season" and "star forward". It's hard to believe such claims after seeing Marvin lose his starting job during the year and seemingly disappear even when on the floor. Their stat page clearly shows that Marvin Williams had the highest Wins Produced at 6.32 edging out Jeff Teague's 6.29. In third place is Josh Smith with 5.61 and then Joe Johnson with 5.40. You can see by the per 48 numbers that Josh Smith brings it across the board. Unfortunately he's not that efficient or at least wasn't last year, so I take it that their system shows that while he appeared to be the best Hawk on the floor to my eyes his mistakes mitigated his overall impact. Joe is the other player that an observer of guys might brand as the best Hawk. The problem with Joe's numbers are two fold. As primary offense option he is exposed to a lot of potential turnovers due to handling the ball against multiple defenders. Additionally he doesn't rebound as well as expected for the position which could be due to his focus on scoring above everything else. Back to Marvin, he's clearly very efficient. He has low turnovers and higher shooting percentages. In their system efficiency = wins, it's simple. I guess he was efficient enough to overcome lesser minutes to lead the team in Wins Produced. His role on the team doesn't expose him to many turnover opportunities or even that many truly contested shots. You'd think that if he was so efficient and productive that he'd see more court time, not less. The problem with that is that I don't think Marvin's numbers would scale as logically expected because he defers to a fault. You also would have trouble succeeding with a bunch of Marvins on your team because his play relies almost exclusively on others to create and handle the ball. As for the "star forward" comment, Vivek was being kind to consider the top Hawk in wins produced a star. The truth is that the numbers were down for the individual Hawks (spread out). Look at the year before and you see Al Horford leading the team with 10.83 WP and a year before that it was Josh with 12.48 WP. The year before that it was Marvin with 8.15 WP (his best year ever). Basically, Marvin is around #50 in WP in the league which isn't exactly the same as being a star. Also as I already mused it is hard to imagine more playing time working for Marvin because of his nature to defer. Now to the Hawks off-season. I didn't have a problem with the C grade, but if wagesofwins.com add in the financial component to their system they'd probably grade it as an A. 25 million dollars plus to get the 11.7 WP by Joe and Marvin doesn't seem like effective use of salary. My first observation is that we should be jazzed that Al is back because he seems to be a certain double figure producer in WP. Also because of all the outside shooters in the new lineup Josh Smith has the potential to regain his form from 3 seasons ago (12.48 WP) by playing mainly inside. Teague's numbers should improve a little if for no other reason it being a full season. Conversely, Zaza should probably be lower as he loses minutes to Al. The returning core I see as approximately 11 + 10 + 8 + 3 or 32 WP. Kyle Korver's WP was 5 last year in similar minutes to Marvin. The difference seems minor. Devin Harris had a WP of 4 last year and I don't see why he can't match that. Lou Williams only produced 3.5 WP with Philadelphia. I guess he lacks the efficiency to grade out higher, but let's just count those 3 subs at 12 WP. You might notice that the negative WP of Jason Collins is off to Boston, but we can't celebrate it because Johan Petro produced negative numbers as well (so did DeShawn Jackson). You can also count on about 2 WP from backup PF/C going by last year's numbers for Ivan Johnson and Jordan Williams, respectively. While it is just an estimate at this time, the 2012-2013 Hawks project out about the same as last year's team in wins with a greatly reduced long term payroll. I could see that grading out as a C on pure talent in the roster changes, but an A in bang for the buck from the roster. Some final observations on WP versus what you see on the floor. Wins Produced seeks to evaluate players in a way to capture things that often get overlooked while watching the game. It's a laudable goal. Efficient shooting and rebounding help your score more than just raw stat production and turnovers drop you quickly. If you click on the individual players you on their site you can see a players weak areas exposed in red. For instance, while Joe Johnson had impressive points and assists his rebounding, steals and blocks numbers were low in comparison to the average per 48 player. Furthermore, turnovers drag down his numbers, but that is common for the focal point of the offense. Jeff Teague's red marks correspond to lack of aggressiveness of offense. Josh Smith is dinged for turnovers and offensive rebounding, which could be seen as an extension of his positioning on offense. Marvin Williams gets a boost from per 48 minute rebounding numbers and his low turnovers, but his red marks on scoring, assists, blocks and steals are what you'd expect from some who lacks the assertiveness to achieve even average usage. These individual player sections are interesting to me because they actually do correlate with what I see watching the games and I feel help me understand how they got to their WP numbers. Have I drank the kool-aid and buy into wagesofwins.com now? Still not so much. I see what they are trying to do, but also I see enough questionable bottom lines to think their system is flawed. If I did buy in and was upset that the Hawks dealt their best player to Utah, I could at least find solace in the same stat system says Jordan Williams produced essentially the same wins as Deron Williams did last year. What an under the radar steal on a throw-in player! Also, Josh Childress is still out there in free agency and they love Josh Childress. What about you? Does seeing the numbers make the claims about Marvin Williams more reasonable? Or is the concept of Marvin as the best Hawk according to their stat system too hard to resolve with a season of watching him play? Is Wins Produced the advanced stat for you?

Do the authors of these article keeps tabs on what goes on the Squawk? lol

[Guest]

I seem to remember an "Evildallas" poster from RealGM. I haven't posted on there for YEARS. I would not be surprised if they ventured to the Squawk, but in general I wish the RealGMers stay over there. Too many posts about of hypothetical trades, just give them all a copy of NBA2KXX and let them have fun.Peachtree Whoops still does not actually know what is going on with WoW. The blogger just looked at the results for a while and then said "oh ok, I see these red areas and these green areas and now I see they combine to make an overall green impact!" Not the same as understanding the underlying regression for wins produced. Another fail, but that is to be expected when someone does not do any research (not that I was expecting anything insightful from the analysis anyway).

[DrReality]

I say the writers from this site choose their Allstars based on their numbers and let them play the real allstars. Case closed

[Guest]

WoW with a videocast, makes fun of the Peachtree Whoops some with their podcast:http://www.youtube.com/watch?feature=player_embedded&v=H0-8P8Bjt5EIt starts around 7 minute mark where they talk about the Hawks. Definitely a poke at the Whoops staff about how they comment on the WoW article without even reading the website to understand the stats.Too bad they fail to realize when you point out faults at someone, you've got 3 fingers pointing right back at you. Their arguments using statistics outside of wins produced are unfounded. To me, this has turned into a 'tard fight. Both are wrong, it is just hilarious to see them fighting back and forth.

[AHF]

WoW with a videocast, makes fun of the Peachtree Whoops some with their podcast: http-~~-//www.youtube.com/watch?feature=player_embedded&v=H0-8P8Bjt5E It starts around 7 minute mark where they talk about the Hawks. Definitely a poke at the Whoops staff about how they comment on the WoW article without even reading the website to understand the stats. Too bad they fail to realize when you point out faults at someone, you've got 3 fingers pointing right back at you. Their arguments using statistics outside of wins produced are unfounded. To me, this has turned into a 'tard fight. Both are wrong, it is just hilarious to see them fighting back and forth.

That was just embarrassing as basketball analysis. (1) The analysis on Marvin is just ignorant and funny on several levels. They do nothing statistically to validate their Wins Produced metric - just cite it as a holy grail like Hollinger and PER. They discuss only one statistic and they do it in a way that is just disingenuous. They ostencibly rebut an argument that Marvin doesn't touch the ball much by using his 18% usage rate to say he touched the ball an "average" amount during his minutes on the floor and therefore his limited role doesn't contribute to his efficiency. 18% ranks Marvin 9th on the team in usage rate - below average. In the league Marvin ranks 247th in usage rate. This statistic powerfully shows how limited his role on the team was and the fact that it was cited in support of an argument that Marvin is a "star forward" with an ordinary usage rate for a star is just incompetent analysis. (2) They entirely failed to discuss salary considerations when discussing the trades. Wow. How can you be a basketball analyst and not even mention the financial side of that trade when discussing whether it was a good move for the franchise or not? hawksfanatic nailed this one. One person making the right argument for the wrong reasons (Peachtree Hoops) and the other side claiming the mantle of expert while doing lazy and incompetent analysis. As a third party seeing this, it is just sad. Edited August 28, 2012 by AHF

[MaceCase]

(2) They entirely failed to discuss salary considerations when discussing the trades. Wow. How can you be a basketball analyst and not even mention the financial side of that trade when discussing whether it was a good move for the franchise or not?

Because contracts don't get out on the court and wins-produce

Hey, it works for messageboarders why not fanbloggers?

Woa, just started listening. I had no idea Egon was running WoW!

Edited August 28, 2012 by MaceCase

[AHF]

Because contracts don't get out on the court and wins-produce

Hey, it works for messageboarders why not fanbloggers?

WoW is the new Hollinger for me...except with less substance.

We have our formula. We love it!

All we do is plug into our formula and report on the outcome. Don't ask us about the business or to watch a game (shudder).

My formula says Marvin Williams is a star and I stand by that because....I reran the numbers and my formula still says so! Winner, winner chicken dinner!

Edited August 28, 2012 by AHF

[Guest]

WoW is the new Hollinger for me...except with less substance. We have our formula. We love it!

The writers at WoW are worse than Hollinger, but Wins Produced is better than PER. They actually describe the methodology for creating Wins Produced. Hollinger has never described PER. The Wins Produced story is coherent and makes sense, the methodology of calculating this is where they falter. Dave Berri gets a plus in my book for describing wins produced, Hollinger still hides behind the Insider. The WoW blog has turned to crap these days, except every now and then they come up with a nugget of good information. You cannot do that when you are married to one statistic.

[MaceCase]

WoW is the new Hollinger for me...except with less substance. We have our formula. We love it! All we do is plug into our formula and report on the outcome. Don't ask us about the business or to watch a game (shudder). My formula says Marvin Williams is a star and I stand by that because....I reran the numbers and my formula still says so! Winner, winner chicken dinner!

Hey, I thought my bastardization of a famous meme by certain posters on this site would give away the tongue in cheek nature of my comment. Like what has already been brought up in this thread, there exist fellows on the edges of either extreme, factual and subjective. Their eyes or metrics can't be bothered by such menial details........it's just left up to us in the middle to laugh.

[AHF]

I laughed at your post, MaceCase. I wasn't trying to contradict you - I was just nodding my head and adding more color to my agreement.hawksfanatic - The reason I said "less substance" wasn't because PER is better than Wins Produced (and I agree that sharing the methodology as WoW is laudable)...it was because Hollinger will also talk about some of the business drivers for deals and take that into consideration in his articles. I was shocked by the lack of analysis on that point by WoW both in the article and (even moreso) in the podcast.