Duncan > Garnett?

[TDfan2007]

You guys knew that this was comming since the Spurs play the T'pups today.

But seriously, who is better? I've heard pretty good arguments for both sides, but I'd have to go w/ Tim since he's taken his team all the way... twice. :king

[spurs_2108]

I am going to the game tonight. TD is way better.

Of the subject here. But where do we go and place bets? I went to the Vbookie thingy at the top but can't forget it out. I jsut registered here like 30 minutes ago, so i am a newbie to this site. So help me out. :lol

[Kori Ellis]

There will be a new vBookie posted for tonight's game in a few minutes. You'll see it as a thread that says "vBookie:" before the title.

[spurs_2108]

Cool. Thanks for you help Kori. I really like this site a lot. I finally came across a site for the spurs. I am even looking at spursreport now. Probably everyone is registerd at both. Is there any other spurs sites out there? Anywho. I am going to tonights game. Its gonna be fun.

[Kori Ellis]

Numbers point to Garnett as NBA's best
By John Hollinger, ESPN Insider

http://proxy.espn.go.com/nba/columns/story?columnist=hollinger_john&id=2014376


Let's play an ESPN.com version of Fact or Fiction: Kevin Garnett is better than Shaquille O'Neal.

Now here's the catch: Regardless of which player I think is better, how would I prove it? If I said Garnett, I would point out his outstanding perimeter skills for a big man, his tremendous rebounding effort and his superior shooting form. If I said Shaq, I would mention his domination around the basket and unusual athleticism for a man of his size.

But generalities like that don't make for an ironclad case. Whether we're debating Kevin Garnett vs. Shaquille O'Neal or Casey Jacobsen vs. Bostjan Nachbar, the conversation inevitably returns to the statistics compiled by the individuals in question.

Sure, we also take into account the intangible factors – teamwork, hustle and all those other things that don't show up in the box score. But the crux of the argument nearly always comes down to those things that are measurable.

Let's return to our Fact or Fiction example.
A KG fan would point out that Garnett gets more rebounds and assists while shooting significantly better from the free-throw line. Likewise, a Shaq devotee would highlight O'Neal's clear advantage in field-goal percentage and his nearly two-to-one edge in blocks.

But how do we objectively evaluate those differences? Shouldn't there be an easy way to settle the statistical part of the debate? To answer this question and others like it, I have developed a method to rate a player's per-minute statistical production across all categories, which I call the Player Efficiency Rating (PER). Let's review the logic behind PER and discuss both its strengths and limitations.


How does PER work?
The simplest explanation for how I arrive at a player's PER is this: I add the good and subtract the bad. I assign a value to all the positive plays (points, rebounds, assists, steals, blocks) and add them to a player's score. Then I evaluate all the negative plays (missed shots, turnovers, fouls) and subtract them.

I've seen other formulas that follow similar logic, but the key difference is that they fail to properly weight the different actions. Either they value all of the actions equally, like the NBA's Efficiency formula does, or they assign seemingly random weights to each stat with no rhyme or reason.

The PER doesn't do that. For example, it divides the points from every made basket between the scorer and the assister (if there was an assist). It values an offensive rebound as exactly opposite the value of a missed shot. It penalizes turnovers more severely than missed shots, since the offense can't rebound a turnover.

Additionally, it adjusts every player's rating for his team's pace. Teams that play a breakneck speed like Orlando and Phoenix must have some of the air let out of their stats, because it's easier to average more points and more rebounds than another player if your team's games have several extra possessions. Similarly, tortoises like Detroit and Indians require a boost.


Why is PER a per-minute rating?
Per-minute ratings are the most accurate way to compare players. If we just use per-game stats, the numbers are so heavily influenced by how many minutes a player gets that they become nearly worthless as comparison points. A salient example would be a comparison between Tim Duncan and LeBron James.

Because his team has so many blowout wins, Duncan averages only 34.8 minutes per game, while James, whose team has been in a dogfight nearly every night, averages 41.5. If we don't adjust for that difference, James gets nearly a 20 percent advantage on Duncan – an advantage resulting from his circumstances rather than his skill.

Moreover, per-minute evaluations are the most useful way to answer questions like whether backup Player X should take the place of starter Player Y or if a team should trade Player A for Player B. Looking at the per-game stats without adjusting for minutes played is like Spike Lee watching a Knicks game without his glasses. While he can see some basic outlines, most of what he's looking at is fuzzy and distorted, and a few things he thinks he might be seeing don't really exist.


Does PER pass the smell test?
As you'll see in the chart of this season's top 10 players in PER, the answer is yes – all 10 are recognized as legitimate stars. Last year's top player, Kevin Garnett, won the MVP award. PER also had Shaq as the top-rated player from 1997-98 to 2001-02, an era when most acknowledge he was deprived of several MVPs due to our unwillingness to cheer for Goliath.


What about defense?
Defense is the one dark area that PER has trouble illuminating. We have a few shafts of light in the form of blocked shots and steals, but the rest of it is a blind spot. Thus, PER tends to understate the efforts of outstanding defenders who compile relatively few blocked shots and steals (Bruce Bowen, for example).

On the other hand, players whose defensive contribution consists mostly of steals and blocks, like Jason Williams or Shawn Bradley, get somewhat overrated. We're not talking about massive discrepancies here, but a player like Williams or Bradley will be a couple of points higher than he deserves, while somebody like Bowen will be a few points lower.


What does the rating mean?
The PER approximates how many points a player adds to his team's bottom line if he plays 40 minutes. I say "approximates" because I set the league average PER at 15.00 in every season. That makes it easy for us to make comparisons between seasons, which otherwise might become difficult because of changes in scoring over time.

In any given year, most serious MVP candidates have PERs near 30, while an All-Star usually has a PER of at least 20. On the other end of the scale, anybody lower than 12 better hope he has a guaranteed contract, and 10 is the PER's version of baseball's Mendoza line.


So who's No. 1?
The top-rated player this season is Minnesota's Kevin Garnett, who was also numero uno last season. He's barely holding off Tim Duncan at the top, however, and the margin among the top six players is very small. You'll notice that five of the six have been mentioned as MVP candidates, with Stoudemire's non-candidacy resulting more from his teammate stealing the spotlight than any deficiency on his part.

Big men dominate this year's list, but that hasn't always been the case. Tracy McGrady was the top-rated player two years ago, and some guy named Jordan had the top rating several times in the '90s.

PER: 2004-05 Top 10 (min. 500 minutes)
PER Player Team PPG RPG APG FG%
28.07 Garnett Minnesota 22.4 13.8 5.7 .505
27.33 Duncan SAS 21.2 11.6 2.8 .500
26.82 S. O'Neal Miami 22.5 10.3 2.9 .593
26.81 Stoudemire Phoenix 26.1 8.5 1.5 .565
25.89 Nowitzki Dallas 26.8 10.1 3.1 .452
25.84 James Cleveland 25.6 6.9 7.5 .484
24.00 Kirilenko Utah 15.7 6.6 3.2 .480
23.69 Wade Miami 23.8 5.3 7.1 .484
23.40 Bryant LAL 28.4 6.1 6.4 .425
23.09 J. O'Neal Indiana 24.8 9.0 2.0 .454

What about Nash and Iverson?
Those are the two names I imagine people are looking for most that don't appear on the list. Allen Iverson ranks 11th and Steve Nash is 14th. While they're the top-rated point guards, Iverson is hurt by his field-goal percentage, while Nash's scoring average holds him back from the highest echelon.


Should Garnett be the MVP?
Not necessarily. Like I said, this is the summary of a player's statistical accomplishments, but there's much more that goes into MVP than that. Garnett isn't defending with nearly the ferocity of a year ago, so that hurts him, especially in a comparison with a shut-down defender like Duncan.

Of course, the other "unmeasurables" also don't factor in: attitude, coachability, leadership, clutch play, etc. That's why McGrady lost out to Tim Duncan (as well as several others) two years ago, and it's why you can still build a legitimate argument for Nash, Duncan, Shaq or several others this year.


The big picture
One player isn't necessarily better than another simply because he has a better PER. But other factors being equal – defense, intangibles and the like – the player with the higher PER is definitely outproducing his counterpart. This can be incredibly useful information for teams (not to mention fantasy players), who have to make choices among players for playing time, contract dollars and trades. It's helpful for fans, too, if they're trying to learn more about how their team's players are producing. Either way, PER is a helpful tool to separate fact from fiction.

[timvp]

Cool. Thanks for you help Kori. I really like this site a lot. I finally came across a site for the spurs. I am even looking at spursreport now. Probably everyone is registerd at both. Is there any other spurs sites out there? Anywho. I am going to tonights game. Its gonna be fun.


www.spurscentral.com

[spurs_2108]

The forums on that site look to blah. I'll pass. lol. Thats just my view about the site. I am not trying to be mean.

[smeagol]

www.spurscentral.com
Aren't there more? :angel

[adidas11]

Shaq will probably win the MVP this year.

[td4mvp21]

I really hope not, but your probably right.