Basketball Prospectus: Unfiltered Everything Else is Fluff.

January 1, 2010

More on Duncan v. Garnett

Filed under: Uncategorized — Kevin Pelton @ 7:26 pm

A couple of follow-up notes on the debate (Tim Duncan or Kevin Garnett) at the top of the column on the decade’s top players. First, a really good e-mail from RR, who noted this in Duncan’s favor:

I think pace is accounted for in WARP but how about Pop’s decision to play Duncan less over the years? Duncan is known as a fantasy middle of the road guy because while per minute, he is and was a monster, Pop’s tendency to keep his minutes under forty and lately under 36 since 2003 (actually 36.7 and approx 34) always pulled his counting stats down. Garnett just started cutting his minutes down to 30 the last three (coincident with boston being a dominant winning team and injury).

Yes, that’s a good point. I ran a quick and dirty experiment to show the effect of playing time. If you give Garnett Duncan’s minutes per game, he would have nearly 10 fewer WARP over the last decade. Now, Garnett may have paid for that late in the decade with last year’s knee problems. Duncan has played more minutes per game the last three years, and that study doesn’t account for the additional games Duncan played last season because he was healthy. Still, the small difference between Duncan and Garnett disappears or maybe even reverses when we consider how Gregg Popovich has used Duncan.

On the other hand, another reader wondered whether the argument I raised about Duncan’s advantage in late-game situations was really backed up by the numbers. So let’s take a look. Here are statistics for the entire decade for close and late situations (five minutes or fewer remaining in the game, five point differential or smaller) for our top 20 players.

Player      ORtg   eFG%    TS%    Usg   Ast%   TO%

Nash       129.6   .553   .649   .276   26.7   9.3
James      125.8   .506   .578   .383   15.1   8.2
Billups    122.9   .482   .601   .288   18.3   7.1
Wade       119.9   .441   .544   .377   16.4   9.0
Bryant     119.6   .431   .543   .397    9.6   8.2
Nowitzki   116.9   .477   .585   .295    7.9   5.7
Allen      112.6   .501   .579   .274   11.2   7.7
Pierce     112.3   .434   .551   .329   10.0   8.9
Carter     111.9   .443   .535   .345   10.6   7.9
Iverson    106.9   .403   .500   .316   16.5   7.6

Player      ORtg   eFG%    TS%    Usg   Ast%   TO%

Kidd       105.7   .412   .540   .240   25.5  10.5
O'Neal     105.0   .569   .574   .225   11.5   8.1
Miller     104.7   .388   .518   .260   20.9   9.8
Duncan     103.3   .461   .536   .286   12.0  10.7
Garnett    102.3   .459   .537   .273   13.0   8.6
McGrady    102.3   .415   .482   .318   16.0   8.6
Gasol      101.0   .473   .565   .234   11.2  12.0
Marion      99.8   .528   .588   .184    6.0   6.7
Brand       87.5   .457   .504   .214   11.2  10.1
Wallace     71.3   .477   .482   .089   10.1  10.8

This group is ranked by an adjusted Offensive Rating. It takes into account the scoring and passing portions of WARP (basically all of offense except offensive rebounding) and adjusts for usage rate using the rule of thumb that one additional percent of usage is generally equivalent to one point of Offensive Rating. All are adjusted to an average usage rate of 20 percent. (That is, a player with a usage rate of 25 percent has five points added to his Offensive Rating.)

What is quickly evident is that perimeter players dominate this list. In general, WARP favors guards on offense and big men on defense (consistent with the results from offensive and defensive adjusted plus-minus). However, this is even more dramatic here. Dirk Nowitzki is the only big man in the top 11, while just one guard (the surprisingly ineffective Tracy McGrady) is in the bottom seven. Elite wing players like Kobe Bryant, LeBron James and Dwyane Wade are able to increase their usage rates to superhuman levels down the stretch without sacrificing a great deal of efficiency. Nash, on the other hand, has achieved incredible efficiency in these situations to go along with the highest assist rate of any of these players.

As for Duncan and Garnett, the former has an advantage, but it’s a slim one. Duncan has the higher usage rate of the two, and derives a significant advantage from the fact that fewer of his baskets are assisted (57.5 percent of Garnett’s baskets were assisted, as compared to just 40.3 percent of Duncan’s, showing his ability to create his own shot). Garnett, however, is much less prone to turnovers in these situations. So late-game play probably should not be considered a point of difference between Duncan and Garnett, though it does potentially hold up as a reason to support Bryant.

No Comments

No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.

Powered by WordPress