Should KG and Melo start in the All-Star game?

On Thursday the NBA announced the starters for the 2013 NBA All-Star game that will take place on February 17th in Houston. The players are voted in by fans across the globe, often leading to some interesting outcomes. This year, instead of voting for the usual two guards, two forwards, and a centre, voters were asked to select two guards and three front court (forwards and centres) players. The starters are:

Western Conference
Player	Position	Votes	Wins Produced
Kobe Bryant	Guard	1591437	5.0
Kevin Durant	Front Court	1504047	10.3
Chris Paul	Guard	929155	9.4
Dwight Howard	Front Court	922070	4.8
Blake Griffin	Front Court	863832	4.7
Totals		5810541	34.2

Eastern Conference
Player	Position	Votes	Wins Produced
LeBron James	Front Court	1583646	8.5
Carmelo Anthony	Front Court	1460950	3.3
Dwayne Wade	Guard	1052310	4.5
Rajon Rondo	Guard	924180	5.6
Kevin Garnett	Front Court	553222	1.5
Totals		5574308	23.4

We here at the Wages of Wins understand that for many fans, the All-Star game is about supporting their favourite players. But what if the most productive players were selected to the All-Star game? What would those starting lineups look like? (numbers up to and including Thursday night’s games)

Western Conference
Player	Position	Votes*	Wins Produced
Kevin Durant	Front Court	1504047	10.3
Chris Paul	Guard	929155	9.4
Andrei Kirilenko	Front Court	DNQ	6.4
Kenneth Faried	Front Court	DNQ	6.3
James Harden	Guard	485986	5.9
Totals		2919188	38.3

Eastern Conference
Player	Position	Votes*	Wins Produced
LeBron James	Front Court	1583646	8.5
Tyson Chandler	Front Court	467968	8.5
Anderson Varejao	Front Court	149246	6.1
Jose Calderon	Guard	DNQ	5.7
Rajon Rondo	Guard	924180	5.6
Totals		3125040	34.4

* only the top ten guards and top fifteen front court players qualified for the voting leader boards, and we don’t have the voting numbers for those who did not qualify.

In both cases the Western Conference has an edge on the Eastern Conference, but it’s clear that the voters made some particularly poor choices when it comes to the Eastern Conference. Kevin Garnett, who was voted to his 15th All-Star game, is no longer the player he once was. After leading the league in Wins Produced during the 2003-04, 2004-05, and 2005-06 seasons, Garnett has fallen to 146th in the current season. This is not overly surprising, as Garnett is 36 years old and playing in his 17th season. But Garnett gets a pass from fans for being an all-time great on one of the league’s most popular teams. The other poor choice for the Eastern Conference is Carmelo Anthony. Despite scoring a lot of points, Anthony has never been very productive. As a matter of fact, this has been the most productive season of Anthony’s career to date, and he’s only the third most productive player on his own team. But NBA fans are mesmerized by the high scoring totals that Anthony puts up, and he does currently play in the league’s biggest market, so the fans vote him in year after year.

The real travesty of all this is that Tyson Chandler, Anthony’s teammate, is tied with LeBron James for the most wins in the Eastern Conference. While Anthony gets most of the credit for the Knicks’ early success, it’s actually Chandler who leads the team in wins. The other front court player who has been overlooked the most is Anderson Varejao. Although he recently suffered an injury and will be out for the next one to two months, Varejao put up some monster games in the early part of the season. It seemed like this was the year that the underrated forward/centre would finally make an All-Star roster, but injuries continue to hound him and it seems unlikely he’ll make it as a reserve.

Finally, the perpetually underrated Jose Calderon is slightly more deserving of starting than perennial favourite Dwayne Wade. Calderon has produced the fourth most wins of any player — tops among guards and just ahead of Rajon Rondo — in the Eastern Conference. Wade, on the other hand, is twelfth in the East and sixth among guards. But the difference is close enough to excuse the fans for voting for the more exciting player (although the fans should have at least voted Calderon into the top ten). Here’s hoping that Calderon gets voted in as a reserve (fat chance).

In the Western Conference, voters did a pretty decent job of picking productive players. The two top players in the conference — Kevin Durant and Chris Paul — were both voted in. Additionally, Kobe Bryant (12th), Dwight Howard (15th), and Blake Griffin (19th) are all among the West’s top players. Andrei Kirilenko (3rd), Kenneth Faried (4th), and James Harden (7th) would all be better options, but the difference is somewhat forgivable. What isn’t forgivable, however, is the fact that both Kirilenko and Faried didn’t even make the top fifteen front court players in West. For shame, voters; for shame.

- Devin

PS: Over at The NBA Update, Arthur North independently came up with a very similar players for both the West and East starting lineups.

11 Responses to "Should KG and Melo start in the All-Star game?"

Tommy Grand says:

January 19, 2013 at 10:11 am

I agree with Arthur North. Tim Duncan was robbed. T.Chandler got it worse.
ccagrawal says:

January 19, 2013 at 2:36 pm

Another great post! I was especially surprised by the KG vote since his performance has obviously dropped this season.

But I had another question/thought. Have you guys ever considered using Dean Oliver’s DRtg along with the individual defense statistics (steals, blocks, defensive rebounds) to calculate WP. From what I understand (which may be wrong, correct me if so), you guys use the individual statistics, and then use the team defense adjustment. But that team defense adjustment is the same for each player on a team.

When I look at Dean Oliver’s statistics for the current Raptor’s roster (http://www.basketball-reference.com/teams/TOR/2013.html#advanced::none), I see that Jose Calderon is one of the best offensive players on the team — which you guys agree with — but that he’s also the worst defensive player on the team. I’m sure this wouldn’t make too big of a difference for most of the players WP, but it could have a significant effect on the players that are on either extreme on defense. Imagine how good Tyson Chandler would be now!
Brandon Owen says:

January 19, 2013 at 4:24 pm

Even if they had given it serious consideration, the problem then becomes…how would you weigh it? You have to compare it to other players on the team, not just the number itself (Ryan Anderson, as an example, had a DRtg of 109 before joining the Howard led Magic, a 103, 101, and 105 with Howard, and back to 109 this year, so context matters). So you have to find a way to normalize it in context of the team. Secondly, you have to adjust it year in and year out to what the League average is, because that matters as well.

The main problem, though, is that you have to somewhat arbitrarily decide how to weigh DRtg if you’re going to “use it” to calculate a “new” WP. I don’t think the WP guys are terribly concerned with adding arbitrary calculations into their formula.

Personally (and this is me, not putting words in anyone else’s mouth here), I do tend to take things such as that into consideration somewhat, but even I don’t think there’s really a place for it in a “new” WP formula. With Calderon specifically, he ALREADY gets very low steal/block/rebound numbers though, which is WP’s way of measuring “Defense”. In Calderon’s case, I think WP probably regards him pretty well: his steals/blocks/rebounds lead you to believe he’s a piss-poor defender, and his DRtg says the same. I would argue there’s probably not that many players that this isn’t true for, overall: the stats do, for the most part, tend to correlated pretty well with actual defensive ability. There are a few cases I feel it’s not entirely correct (for instance, based purely on stats, one might believe Kevin Durant to be a superior defender to LeBron, or at least quite close: 6.8 rebounds per-36, 1.4 steals, 1.2 blocks to LeBron’s 7.6, 1.6, and 0.9…but I don’t personally believe their defensive value is that close; subjectively, LeBron’s still further ahead than Durant, although the gap is closing), but again, they require subjective considerations, and subjective considerations have a couple problems: they’re incredibly hard to model into an objective formula…and they’re prone to being very, very wrong. I subjectively think WP probably doesn’t give LeBron enough credit defensively, and I honestly and truly believe that. I also realize it has a high probability of being wrong, because it is entirely subjective.

Point being WP tends to already cover the general bases of DRtg via the stats that it uses. This used to be my main criticism of WP, until I really thoroughly looked at players I thought were probably being over/under-valued subjectively…and realized that, due to their defensive stats, their value was probably right in line. Now my only major qualm with WP is with regards to the positional adjustments (using LeBron as an example again, his WP numbers suffer greatly this year because he’s listed as a PF, despite the fact that he’d probably be putting the same exact numbers up as an SF, and the truth is that the Heat have been so ambiguous with their SF/PF distinction this season anyway that it’s hard to really justify pigeonholing LeBron in particular into one of those spots). And even that one is a fairly minor qualm, and doesn’t apply to 99% of players.
ccagrawal says:

January 19, 2013 at 7:00 pm

Brandon, from what I understand, WP incorporates individual defense through steals/blocks/rebounds as well as a team defense adjustment (http://wagesofwins.com/how-to-calculate-wins-produced/). I have no problem with using the steals/blocks/rebounds, but I’m not sure how well those alone can measure defense. I tried testing by using data from teams this year.

Correlation of Steals vs Points Against Per Possession = -0.34
Correlation of Blocks vs Points Against Per Possession = -0.24
Correlation of DRebounds vs Points Against Per Possession = -0.42

So the WP team definitely acknowledges this, and they supplement it with a team defensive adjustment. And I think the problem you described with Ryan Anderson in DRtg should still be present with WP. When WP incorporates team defense, Ryan Anderson will be getting a boost just for playing alongside Dwight Howard, a great defender. At least with DRtg we can easily see that Dwight Howard deserved much more of the credit than Ryan Anderson did for Orlando’s good defense in 2011-12 (http://www.basketball-reference.com/teams/ORL/2012.html#advanced).

In Calderon’s case, not only should he be getting no credit for having a low steal/block/rebound amount, but he should be getting the worst team defensive adjustment of the players on the Raptors, since he’s the worst defender on the team. WP has him being one of the top players in the league, but I read a blog post that puts up a decent argument saying otherwise (http://ascreamingcomesacrossthecourt.blogspot.com/2012/10/jose-calderon-example-of-how-wins.html)

I think calculating the Team Defensive Adjustment with the DRtg statistic could help. The players on the team with the highest DRtg should get the highest adjustment, and vice versa for the players on the team with the lowest DRtg. Just to make sure the statistic was accurate on a team level, I made a “Team DRtg” by averaging each team’s players’ DRTGs weighted by the minutes played, and the correlation of Team DRtg vs Points Against Per Possession = -0.96. So this statistic does do a decent job in calculating defense, as difficult as it may be.
Devin Dignam says:

January 19, 2013 at 8:06 pm

ccagrawal,

I’d mostly echo what Brandon said. However, I’d recommend that you check out these links:

Let’s Talk Team Adjustments
Breaking Down Wins Produced

The first link uses the old Wins Produced formula, but the point remains the same: the team defensive adjustment isn’t very large at an individual level. The second link mentions the team defensive adjustment and may also be of interest to you.

As far as ORTG and DRTG go, both form the basis of Win Shares — Wins Produced’s sibling metric. So ORTG and DTRG are going to tell largely the same story as Wins Produced does (although as a matter of fact, the Win Shares formula uses an even larger team adjustment than Wins Produced). When looking at ORTG and DRTG, I’d focus more on the differential (ORTG-DRTG) than anything. In Calderon’s case, that means he’s third on the team, behind Quincy Acy (24) and Amir Johnson (14) and tied with Kyle Lowry. This generally agrees with WP48 — Acy’s at 0.278, Calderon’s at 0.249, Johnson’s at 0.174 (as a centre), and Lowry’s at 0.230. Likewise for WS48, although obviously I prefer Wins Produced over Win Shares.
TheNBAUpdate says:

January 19, 2013 at 9:14 pm

Thanks for the shoutout Devin! To me there’s no real excuse for Calderon to not be in the All Star game. While the general public doesn’t understand just how good Calderon’s shooting efficiency and ast/to ratio is, I don’t think anyone denies that he’s a great offensive player. The number one criticism has always been his defense. That’s a valid argument, and while I agree he’s a bad player on the defensive end, his offense itself makes up for it. Besides from a fan perspective don’t you want to see the best offensive players in the All-Star game? Everyone says that the most exciting, offensive players should play in the All-Star game to make an exciting game yet when Jose Calderon is brought up his poor defense becomes the issue, even though his offensive game makes up for it. Very hypocritical.
Statmandu (@lovethoseknicks) says:

January 20, 2013 at 8:42 am

I question the value of Varajao. If he was that valuable, the Cavs should be playing much worse in his absence and their rebounding (which is where Varajao generates much of his value) would be way down. But it’s not.
TheNBAUpdate says:

January 20, 2013 at 3:01 pm

Cleveland hasn’t become a disaster after Varejao’s injury because of improved production from his teammates. Alonzo Gee went from terrible to very good. Kyrie is above average now. Tristan Thompson is playing at a very high level. Dion Waiters went from Bargnani bad to Byron Mullens bad… an improvement nonetheless.
This infographic should pretty much explain why Cleveland is playing at the same level (or better) without Varejao: http://wagesofwins.com/2013/01/09/nba-2012-13-season-visualization-to-date/
Statmandu (@lovethoseknicks) says:

January 21, 2013 at 10:48 am

“Cleveland hasn’t become a disaster after Varejao’s injury because of improved production from his teammates. Alonzo Gee went from terrible to very good. Kyrie is above average now. Tristan Thompson is playing at a very high level.”

That’s exactly my point.

Varejao’s statistical value was coming from areas that were easily replaced by his teammates when their roles were expanded in his absence (most significantly rebounding). So while technically Varejao may have been valuable statistically because of what he was producing on the court, he wasn’t that valuable.

Try to replace James so easily. That didn’t work for the Cavs.
Dave says:

January 21, 2013 at 1:43 pm

Statmandu:
1. There’s no indication that once Varejao is back, the gains made by Kyrie, Gee, and Thompson won’t stick around. Unlike shooting, where if one player shoots a teammate can’t shoot, in rebounding it’s rare for teammates to be stealing rebounds from each other because they’re all positioned in different areas of the floor.

2. The Cavs have been noticeably less productive from the center position without him. This is somewhat made up by the gains at other positions, but the point spreads have gotten worse without Varejao. With Varejao, the Cavs were losing by about 5 points per game, without him closer to 12 points per game.

3. The only stat in which Anthony and Garnett are significantly better than Varejao is … no surprise here … scoring points. According to the stats, Varejao is a better passer, better at avoiding turnovers, better at stealing, as well as much much better at rebounding.

So really, what your argument comes down to is that you believe that scoring frequently is something rare and unique while rebounding well is not. In other words, Yay Points!
Reservations about the All-Star reserves says:

January 25, 2013 at 11:12 pm

[...] are supposed to represent the best players that the NBA has to offer. As with the selection of the All-Star starters, a few predictable mistakes were made. First let’s start with who was [...]

January 2013
M	T	W	T	F	S	S
« Dec				Feb »
	1	2	3	4	5	6
7	8	9	10	11	12	13
14	15	16	17	18	19	20
21	22	23	24	25	26	27
28	29	30	31

Share this:

Leave a Reply Cancel reply