drza wrote:therealbig3 wrote:Well, I think Duncan is a better low post player, because he puts more pressure on a defense by drawing fouls. From 98-07 (Duncan's prime), he averaged 7.5 FTA/game in the regular season, and 8.9 FTA/game in the playoffs. From 99-07 (KG's offensive prime), KG averaged 5.9 FTA/game in the regular season, and 5.3 FTA/game in the playoffs. The fact that KG was a better FT shooter than Duncan made up for the difference in attempts, which is why KG's TS% in the regular season is practically equal with Duncan's. But in the playoffs, KG's FTA go down, while Duncan's go up. That's a huge difference, and it's the main reason why Duncan's TS% is able to stay the same in the playoffs, while KG's takes a big dip. And it's their style of play that dictates this. KG tends to take fadeaways when he posts up, while Duncan tends to go into the defense and either draw a foul or take a short hook shot. And the fact that their FG%'s are nearly identical as well indicates that KG's fadeaways aren't higher percentage shots than what Duncan does. So the fact that Duncan is able to draw significantly more fouls makes him a clearly better low post player. He didn't have to sacrifice efficiency for volume in the playoffs like KG did because of this. And that makes him the better overall offensive player in my mind.
Thank you. Those are logical points. But, as I'm sure you expected, I do have a rebuttal. And it deals with the conundrum of how to gauge stats in the playoffs. We want to emphasize the playoffs more because it is championship time, and I both understand and endorse that. The problem is, for any stat to be of use (in anything, not just basketball) you need to have a
large sample size and similar conditions. The further you get away from those rules, the weaker your statistical conclusions. The problem with how to judge the postseason is, not only is the sample size smaller but also the conditions are now different. Instead of playing every team multiple times (like in the regular season), in the postseason you play the same team a bunch of times. Which means that who you play is going to make a BIG difference in your results, which therefore is going to bias your stats. Let me give you the example that took me on this tangent.
In the Top 100 project, as I'm sure you remember, we had a discussion about why KG's TS% wasn't in any way reflective of the impact he was having on games. While researching this, I pointed out that from '99 - '01 in the playoffs KG was matched up against Duncan/Robinson, Sheed/Sabonis, and Duncan/Robinson in back-to-back-to-back years. And I showed how Duncan and Sheed had their scoring totals and/or percentages fall through the floor against KG, only to bounce back in a huge way in the rest of the playoffs. Remember? If so, keep following me, because it ties back into their postseason stats.
KG and Duncan faced each other in 8 games over the 1999 and 2001 playoffs. In those 8 games, Garnett averaged 21.4 ppg on 52.5% TS. Now, over the 1999 and 2001 playoffs, Duncan averaged 23.7 ppg on 55.3% TS. So, if we just look at the postseason stats, we'd say that in those two years Duncan produced a little more volume on a bit more efficiency than KG. In fact, those numbers are very representative of their career postseason averages, where KG's TS% is 51.9% and Duncan's is 55.0%...about the same 3% difference in TS%. Not a huge amount, but enough that Duncan supporters can hang their hats on it.
But wait a second.
Because KG compiled all of his 21.4 ppg on 52.5% TS in eight games against one of the two best defensive power forwards of all time. And point of fact, in the eight playoff games that Duncan was facing one of the two best defensive power forwards of all time, HE "only" averaged 20.7 ppg on 51.4% TS himself. The difference, then, between Duncan's and KG's true shooting percentages in the '99 and '01 postseasons was NOT that Duncan was operating out of the post, or drawing more fouls, or opening things up, or any of the logical suggestions you made above...the difference in those 2 years was that while KG and Duncan performed similarly against similar levels of competition, Duncan ALSO had another 22 games in those years NOT against the best defensive power forward in history, and in those 22 games he averaged 24.8 ppg on 56.6% TS. So this makes it a textbook example of how the
small sample sizes/different conditions gives a weaker (and in this instance misleading) conclusion about Garnett's and Duncan's shooting efficiencies.
Now, we would hope that eventually it would even out over time...that the law of large numbers would kick in over the years as the sample sizes grew. Unfortunately, I'm not sure this really happens, though, because year after year KG and Duncan kept not playing the same teams in the same postseason...and playing very different numbers of games per year...with very different calibers of teammates. I've said many times that part of the reason why I believe KG's efficiency goes down is because of the ratio of his own teammate's caliber vs the caliber of his opponents. He's asked to do so much against such stacked competition that his efficiency suffers a bit, but his volume if anything goes up and (as I showed with the available +/- data in the top 100 project) his impact is absolutely massive, drowning out the efficiency question. The +/- data is probably the closest that we can come to normalizing for teammate caliber, but aside from that, today I wanted to follow my above idea and play it out. So, I did a quick test.
I did a quick rough-and-ready test this afternoon. From '99 - '09 (from the first year they both made All NBA until their obvious fall), KG and Duncan played in the same postseason in '99, 01, 02, 03, 04, and 08. We talked about '99 and '01 above, so here let's look at the other 4 seasons. We need a test criterion for competition level, so let's look at how KG and Duncan did against teams that won 55 or more games (reasonably contenders) vs teams that won fewer than that ("merely" playoff teams) over those years.
In '02, 03, 04 and 08 Duncan played 34 games against teams with 55 or more wins, while Garnett played 28 such games. Duncan averaged 22.9 points on 52.2% TS% in those games, while Garnett averaged 22.4 points on 52.0% TS. It's funny how similar those numbers look, isn't it? Meanwhile, in Duncan's 26 games against teams that won fewer than 55 games his TS% jumped to 57.6%, again opening up the ~3% advantage over KG and his 53.9% TS in 25 games against teams with fewer than 55 wins.
Now, I'm not in any way claiming this is rigorous. First of all, I didn't include Duncan's '05 - 07 years, when he was in the playoffs and KG wasn't, and at a glance his great/good opponent splits aren't as large in those years so I wouldn't be surprised if they would have boosted his numbers a bit in the efficiency department and allowed him to open a small gap over KG. Secondly, at their absolute peaks Duncan tended to have higher percentages than KG, while KG was more efficient once he got to Boston, so there was obviously a teammate effect that I don't account for. And third, this was a very quick test with a simple record-based criterion...there are probably better, more sophisticated ways to try to make an estimate like this. My point here wasn't to re-discover the wheel.
My point, is, that by just looking a bit closer I'm now seeing very reasonable evidence from the 1999, 2001, 2002, 2003, 2004, and 2008 postseasons to suggest that the Duncan's better shooting efficiencies appear to be in the "noise". That for those 6 postseasons, which span the majority of their primes, the shooting efficiency stats are weakened enough by the violation of the sample size/similar conditions rules of stats that I can't with any confidence say that Duncan was REALLY a more efficient postseason scorer than Garnett to any degree.
ESPECIALLY when you consider that over the much larger/more equal conditions sample size of the regular season, from 1999 - 2009 KG averaged 21.6 ppg on 55.1% TS (816 games played) vs Duncan's 21.4 ppg on 55.0% TS (817 games played). Almost rock-solid, spitting image EXACTLY the same.
I'm just saying. Duncan may be slightly more efficient than KG as a postseason scorer. But even if we chalk up everything I wrote above as garbage, even the 3% Duncan is better in career average is a pretty small number in the scheme of things with how much each was responsible for. And it seems to me that when one chooses a single aspect of the game and expounds on it, the way folks do with Duncan's post-oriented game vs Garnett's, it tends to skew the message. Because I could just as easily make similar paragraphs about how Garnett's high-post pick-and-pop ability is what opened things up for Sam, Chauncey and Hudson to have (to date) career years while playing with him. And Garnett's ability to draw attention and find open teammates is what allowed Wally to be as effective as he was on offense. And that his ability to help his teammates on defense with so much range is what has boosted Pierce, Rondo and Perkins into a higher stratosphere of attention than they otherwise would have received. And that it's because Garnett is doing so many things that the box scores don't catch that he has such a significant advantage on Duncan in the +/- stats over their primes.
Both Duncan and Garnett are history-caliber players, and both do a LOT of things both in the boxscores and not that allow them to make huge impacts. So just picking one of the areas that Duncan stylistically does differently, and expanding on it in the absence of balance, and using stats with small margins of difference under dissimilar conditions as the justification...I just don't think that you're covering the whole story, I guess is what I'm saying.
), but as of right now, I feel that Duncan was a more consistent offensive presence, while KG's offense could be slowed down easier (relatively speaking). And this goes back to their relative skillsets, I think Duncan's low post scoring ability allows him to maintain or even improve his regular season production (and impact) with more regularity than KG.
