Home > Bolts > Friday Bolts – 8.6.10

Friday Bolts – 8.6.10

An interesting look from Tom Haberstroh on whether or not Carmelo Anthony is a max player: “In the end, Anthony’s game demonstrates why it’s important to strip away the biases that color our perceptions of elite players. In Anthony’s case, the excessive shot volume, his team’s stat-padding tempo and the lack of a true 3-point game makes his 28.2 ppg seem far less impressive than his sparkling reputation would suggest.”

Ziller ranks the world: “That the United States is the overwhelming Vegas favorites at the Worlds is a sign the betting public has no short- or long-term memory. Hasn’t Team USA already proven chemistry can’t be conjured in two months? Remember, the first travails of the Redeem Team ended up defeat (and some would say humiliation): in 2006, when LeBron James, Dwyane Wade, Carmelo Anthony and Kobe Bryant first got together, Greece knocked them out of contention in the semifinals of the World Championship.”

ESPN’s future rankings are out and OKC is third, down from one: “The Thunder have fallen from their perch atop the Future Power Rankings to third even though they’ve done little wrong. But they didn’t have the offseason Miami did, obviously, and the Lakers slid past them, too, as a result of their continued success putting the right pieces around their core. Can the Thunder do the same? Kevin Durant is an MVP candidate, and having him under contract for five more seasons should make Oklahoma City a much more desirable destination for other players — at least for those not angling for good weather or a bustling nightlife. Although the Thunder have plenty of good young players around Durant, it remains to be seen whether any will join him on the journey to stardom. To match up against the likes of L.A. or Miami, they’ll need somebody like Russell Westbrook, James Harden or Serge Ibaka to become an elite-level player.”

Durant is the second biggest bargain in basketball: “I’m guessing most people would argue that Kevin Durant is by far the biggest bargain in the game. But even though LeBron gets paid $11 million more than Durant, he still provides about $5 million more in surplus value. The difference of $11 million would generally buy about 5 wins on the open market so the 7.1 extra wins that LeBron produced wins out.”

The Thunder has re-branded Loud City as “Love’s Loud City”. I think I’ll have some more thoughts on this later.

The NY Times calls Patrick James’ Thundershark column “cute-but-weird”: “In other cute-but-weird N.B.A. news — if you consider sharks cute — the Oklahoma City Thunder is celebrating Shark Week by comparing Thunder players to different varieties of the species.”

Remember our summer t-shirt contest? Well we need more submissions!

Finally, I’m working on an upgrade to the commenting system. I don’t know if that’s something the gallery would be interested in, but we might test run it for a week or two and see what everyone thinks.

Share:
  • Print
  • Digg
  • Facebook
  • Twitter
  • del.icio.us
  • Fark
  • StumbleUpon

Bolts

Post comment as twitter logo facebook logo
Sort: Newest | Oldest

At this point I like offensive SPM a good deal. On the defensive side I'll look at dSPM but I'll rely more on defensive Adjusted +/- and counterpart and team defense with player on the court and on/off.

The rigid "implicit, simplifying" assumption.

Not saying that folks necessarily believe it rigidly but, if you assume it in a metric, then you either have to subjectively adjust later or accept making decisions off a database that assumes most or all of defense is equal from all players (and in several cases using data not limited to when they are even on the court).

A comparison of the defensive factors of Adjusted +/- with counterpart defensive data might be informative in thinking about and untangling the individual and help defense of players. A strong correlation might improve the perceived validity of each, a weak one could do the reverse. The comparison, at this point for me, would have to be multi-year since that is the only dataset I currently have of the former. Teams should do this. The team that hired Joe Sill can do it and on 1 year data if it wants. May a few others have, maybe not. I've spot checked a few players and the datasets generally agree and that agreement gives me more information than ignoring it and more confidence in these datasets than the rigid assumption that most or all of defense is equal from all players.

For 2 straight seasons the counterpart data suggests White is very effective at individual and team defense levels when presumed to be guarding the center. But that seems to be little known and given little weight. I'd explore it more against lineups with a lot of subs and suggested such last summer/fall. Maybe his next coach will try it more, if he gets either a free spirit or simply someone willing to listen to the numbers and experiment a bit until it either suggests going on or turning back.

Without counterpart data used it might be hard to get folks to believe DJ White was a positive impact defender this season. With it you can show that White's defense was far better in 08-09 in far bigger minutes than in 09-10. Whether you use that to say he slid and no longer deserves to be out there or use it to say he just needs time to shake off the rust / jitters is another thing, but looking at the counterpart data seems worthwhile to me. So I will.

The variation in the quality of individual shot and shot help defense appears to be often enough to potentially move players from average on their individual traditional box-score based data to above or below average or down from being above average or up from being below average. Not having it included can be considered akin to another "error" term.

This post illustrates http://courtsideanalyst.wordpress.com/2010/07/26/m...

Actually there is counterpart data available at Synergy (if you pay for it), 82 games and http://courtsideanalyst.wordpress.com/
queencityhoops and basketballprospectus , though the later 3 are based on boxscores not play by play.

If you don't want to use counterpart data exclusively you can blend with team level data.

Or just do what Sammy and others have suggested, which is to use team defensive ratings from when a player is actually on the court (as opposed to the average which happens with and without him, with it being hard to argue the latter is appropriate to include).

It is my understanding that Synergy provides counterpart or defensive data presumably based on the videotape and presumably because the NBA execs who are the main financial support for the technology and the data asked for it.

If you want to use it, you can find it and use it. It will always be a caveat (at least for some users) for metrics that don't have it.

82 games has had play by play matched up by assumed position counterpart data available and utilized for simple net counterpart match-up results for 7 years. It took a few more years for the Roland Rating to show up, change and stay mysterious and little used by others but the data is there and despite its imperfections it is still probably better in most cases than assuming every defensive player on a team is equal and gets the same team defense score.

Counterpart data isn't "difficult" to integrate. It is just data. It has been done. I've suggested an even easier way before that makes less pretenses about precision. You just have to decide it is worthwhile.

I like the randomness of Okie weather. Any weather for too long gets old for me.
I went to Cali/Vegas in June/July and it was 110 every day for 2 weeks. I went to Florida in May and it rained on us for 9 of 11 days. I went to Antarctica and it was cold and the penguins jumped me. (OK. that last one was a fib.)

But the point is give me a nice/fair day, a cloudy/overcast day, and hot/sunny day then a cold/rainy day in the same week...now THAT'S why I'm still here. Tornadoes, ice-storms, hail-storms, flash foods, ect...it just kinda exciting! Yes, I chase tornadoes, sometimes ;)
If you don't like the weather, wait 5 minutes, go inside, or vacation frequently to a desired climate. Go THUNDER!!

Sammy :

DSMok1 :

Sammy :@DSMok1

Does this definition from B-R still hold for your formula or are there other inputs?
Statistical Plus/Minus is an estimate of the player’s contribution to the team’s point differential per 100 possessions, using his boxscore stats as inputs.

If so, how is this different from a modified version of PER?

Significantly different, in general. Read the thread on APBRmetrics about it. PER is on that spreadsheet, also, for comparison.

I understand PER and SPM weight the inputs differently, but the isn’t the data from which the metrics are derived is essentially the same (i.e. box score stats)? Doesn’t this inherently skew the top defensive performers toward those who get a lot of blocks and steals, and underrate good positional defenders who might not have high block/steal totals?

Correct on the defensive issues. I can add charges taken to the mix at the same rate as steals and that helps quite a bit. Still, counterpart data such as Crow likes would be quite difficult to integrate.

California = sublime weather and much more fun outdoors ... that's why I moved.

Weather sucks here. Ice storms. Heat that's insane. Born and raised here but it's worse than half the country

@Todd

You're being a homer about the weather.

"Kevin Durant is an MVP candidate, and having him under contract for five more seasons should make Oklahoma City a much more desirable destination for other players — at least for those not angling for good weather or a bustling nightlife."

Okay, the "bustling nightlife" comment I can understand. But "not angling for good weather"...Huh? What, do they think it's 120 degrees six months a year with the other six months witness to a rampaging pack of F5 tornadoes attacking the state? Where do they come up with this stuff? Do they ever actually spend *time* here before writing idiotic comments like that? And does it have to be 72 with no clouds and no wind all the time for these people in order to qualify as "good weather?" Geez.

Using counterpart defensive data, imperfect as it admittedly is, would go further to try to address the issue. So far a variant of Wins Produced by a guy named Ty Willingham is the only system I know of which has done this.

@Sammy

It has been suggested elsewhere to use team defensive rating when a player is on the court in the formula instead of the team average defensive rating for all minutes to somewhat address the issue you raise.

http://nba.fanhouse.com/2010/08/06/isiah-thomas-re...

isiah thomas gets to do the one thing he is good at, talent evaluator

I took note of the fact that the Spurs spent time looking at him but I was slow to see why. So were they perhaps.

Tolliver did decently at Golden State. A role player or trade asset.

Anthony Tolliver to the T-Wolves.

DSMok1 :

Sammy :@DSMok1

Does this definition from B-R still hold for your formula or are there other inputs?
Statistical Plus/Minus is an estimate of the player’s contribution to the team’s point differential per 100 possessions, using his boxscore stats as inputs.

If so, how is this different from a modified version of PER?

Significantly different, in general. Read the thread on APBRmetrics about it. PER is on that spreadsheet, also, for comparison.

I understand PER and SPM weight the inputs differently, but the isn't the data from which the metrics are derived is essentially the same (i.e. box score stats)? Doesn't this inherently skew the top defensive performers toward those who get a lot of blocks and steals, and underrate good positional defenders who might not have high block/steal totals?

@Steve H
I'm not sure they'll go down much. D-Wade and LeBron are elite rebounders for their positions, but Big Z is pretty putrid at rebounding. So whoever they lineup at center (Z, Joel Anthony, etc.) I don't expect them to grab too many boards.

@Greg Bosh has some solid rebounding numbers- but he did it on a really soft team. No great center on the Heat squad, but I still think his rebounding numbers will dip this year.

I would love a threaded system, even a real time system. Sounds good...

Bosh is a good rebounder, in addition to being a great scorer. He should be well above Nene, Josh Smith, and Gerald Wallace. I'd also take him over Tim Duncan of today. If I had my choice of power forwards in in the league right now it would be between Gasol and Bosh. You can't go wrong with either.

Rumor is that Cho is trying to or will take Oklahoma City's Director of Pro Personnel Bill Branch. If so, either Hennigan gets a promotion or they bring in someone.

@DSMok1

Question: How does a player who averaged just shy of 11rpg qualify as someone who doesn't rebound? In 36.1mpg last year, Bosh averaged 10.8 rpg. That seems like a pretty decent number, considering 6th overall accorindg to espn on rebounds/game. Career rpg of 9.4. That's pretty decent considering he's not the biggest PF in the game.. only 230 pounds.

it is cool that patrick james got a mention in the NYT article

I still have trouble understanding the definition of or formula for some of the columns and how you get from one to another. A glossary of them or another step by step, slower than you might think necessary description, would help, at least for me. If you care to (at APBRmetrics or here).

@DSMok1

Among Thunder players, a few things I noticed from your new SPM:

All of Westbrook's positive impact is estimated as coming from the offensive side.

Green, Thabo, Harden and Collison all close to neutral.

The positive portion of Green's and Harden's impact are estimated as coming just from the defensive side but nearly offset on the offensive side.

White had the 2nd best SPM on the team (if you ignore Bowen) and was rated as near equal to Thabo on beneficial defensive impact, however it happens.

Weaver was pretty near neutral and had a defensive impact close to Ibaka's.

Krstic, Ibaka and Maynor were moderate negatives due to their negative impact on offense.

Sammy :
Whoa, why do Roy and Bosh grade out so low?

Roy probably because of injury to some extent, Bosh just because all he does is score. No rebounding, defense, etc.

Sammy :@DSMok1

Does this definition from B-R still hold for your formula or are there other inputs?
Statistical Plus/Minus is an estimate of the player’s contribution to the team’s point differential per 100 possessions, using his boxscore stats as inputs.

If so, how is this different from a modified version of PER?

Significantly different, in general. Read the thread on APBRmetrics about it. PER is on that spreadsheet, also, for comparison.

@GAP

Smith had an amazing season last year, but that doesn't make him someone I would put at the top of my list of players to build a team around.

GAP :
@Mark!
I was just thinking to myself that I wouldn’t give Josh Smith a max contract before Melo.

Yeah, Smith's history is not as good.

@Mark!
I was just thinking to myself that I wouldn't give Josh Smith a max contract before Melo.

Whoa, why do Roy and Bosh grade out so low?

@DSMok1

I'll grant that's probably around where Melo falls on a list of best player in the NBA, but max contract has to take into account 5-6 years of future production w/ an increasing contract value each year.

I could be wrong, but I dunno if Bryant, Duncan, Nowitzki, Ginobili or Nash fetch max contracts. A max contract for any of them would be at least $105M (non-Bird).

And I think we can take Boozer off the list because he didn't fetch a max contract in a market that was giving them away. It would be a fair argument to say his ability deserves one, but his injury history is probably what prevented it. He got $75M/5Y and was eligible for between $90M (non-Bird) and $120M (Bird).

That leaves Carmelo at #11

James
Wade
Durant
Paul
Howard
Smith
Williams
Gasol
Rondo
Wallace
Anthony

That's a fair list. Might be an interesting conversation about whether or not Josh Smith or Gerald Wallace would be above Anthony on a priority list for players you want as the star of your team. They aren't on mine.

Other than those two, I wouldn't begrudge someone for preferring Rondo or Gasol over Anthony even though I'm not sure I would put them ahead of him.

As an addendum, Nene over Bosh is ludicrous in the context of who better deserves a max contract.

@Royce Young
That's exciting.

@Sammy
Yep I believe so and also, I think there is a real-time updating feature on it.

I'm working on maybe having it test some later today.

@DSMok1
Does this definition from B-R still hold for your formula or are there other inputs?
Statistical Plus/Minus is an estimate of the player's contribution to the team's point differential per 100 possessions, using his boxscore stats as inputs.

If so, how is this different from a modified version of PER?

Finally, I’m working on an upgrade to the commenting system. I don’t know if that’s something the gallery would be interested in, but we might test run it for a week or two and see what everyone thinks.

Yes, please! Threaded messages at last?

Mark and all stat geeks:

My new Advance Statistical Plus/Minus spreadsheet is available now:

http://sonicscentral.com/apbrmetrics/viewtopic.php...

Actually, there aren't 15, because some of the players that are above him are also well on the downside of their career. If we're talking for 1 year, the list is (my top 20):

LeBron James
Dwyane Wade
Kevin Durant
Chris Paul
Dwight Howard
Kobe Bryant
Tim Duncan
Dirk Nowitzki
Josh Smith
Deron Williams
Pau Gasol
Rajon Rondo
Manu Ginobili
Steve Nash
Gerald Wallace
Carlos Boozer
Carmelo Anthony
Brandon Roy
Nene Hilario
Chris Bosh

However, over a multi-year contract, the numbers above Carmelo would drop quite a bit.

@DSMok1

I'd love to hear your list of 15 players you would pay max money before Carmelo.

love's loud city is pretty corny, but it's a whole lot better than hertz loud city. that would be awful.

yea, if it means more money for the team, bring it on.

Cole Aldrich finally signs his rookie deal. Only unsigned first-rounders now are the Memphis guys, Henry and Vasquez.

By the way, I hate Aldrich's twitter. Talks too much about decorating, moving and petty complaints about living in a hotel. Doesn't sound very motivated to me, but I'll judge him for what he does on the court.