On Monday, Tim Duncan announced his retirement from the NBA  after 19 seasons in the NBA. While Kobe Bryant announced his retirement early on in the season the NBA lost two of it’s greatest stars of the last 20 years in the same season. Together they won 10 NBA championships (both winning 5) and appeared in 13 different championship series since 1999. Only 5 NBA finals did not feature Duncan or the Mamba in the past 18 seasons.  But who was the more dominant player and who will be remembered more? Let’s take a look at the data, which I gleamed from basketball reference  (which had all their data in downloadable format!). Below we can see the 3 major offensive stat averages per 36 minutes below (broken down by season) along with a defensive summary measure, blocks+Assists.

36 min

All of these statistics produced significant p-values using a wilcoxon signed rank test (with Kobe having significantly more points and assists and Duncan having more rebounds and blocks/steals) but since they play different positions (SG vs PF/C), this doesn’t really make our comparison for us. Let’s take a look at the win shares (which attributes the number of wins a player adds to the team) for each player over time

WinShares

The wilcoxon signed rank test produces a p-value of 0.2549 so we don’t have a significant difference in win shares between the two, although Kobe had a bigger drop off towards the end of his career (partially due to injury). The stars are the years that they won the championship, which if you notice only Duncan won the ‘ship during his best season in terms of win shares. Kobe’s best season in terms of win shares was in 2007 (which many argue should have been an MVP year) as he averaged 32.8 points per game, but the Laker’s fell in 5 games in the first round against the Suns. Duncan had a higher career total of win shares with 206.4 while Kobe had a total of 172.7. Before we get into post season stats, let’s take a look at another stat, wins above replacement which estimates the number of wins each player contributed above a league average replacement player.

WAR

Once again, we see a p-value of 0.2727 for the wilcoxon test indicating no significant difference, which we can also see graphically. Duncan’s peak seasons were in 2002-2003 (same time as Kobe’s in 2003). Duncan leads in this category for his career with 241.11 WAR compared to Kobe’s 194.67. Again Duncan’s best season in term’s of WAR led to a Spurs title while Kobe’s best took place in Shaq’s final year (2004) with the Laker’s supersquad. For this stat, Duncan’s second best WAR outperformed Kobe’s best WAR season and Duncan had 3 out of the 5 best WAR seasons among both the players. Kobe actually had NEGATIVE WAR in 2014 (when he only played 6 games due to injury) and in his sendoff year of 2016, which is reflected in the Laker’s worst record in franchise history. Still, while it’s cool to see the trends, there’s not a significant difference here.

Now let’s talk about playoffs . First I’m going to do the same comparison as with 36 minutes, looking at the boxplots of points per game and other stats for each playoff year. I’m looking at per game here instead of per 36 because when you need them the great ones will show up for a whole game.

PlayoffStats

Once again we see the same trend with the Wilcoxon signed-rank test, namely that Duncan outperformed Kobe defensively and on the boards but Kobe outperformed Duncan significantly in points and assists per game. Now let’s take a look at the advanced NBA statistics in the playoffs. First let’s look at the win shares over each of their respective careers:

WINSHAREPLAYOFF

We see that both players had the most win shares in seasons where they won the championship as the true alpha dog. In all 5 of Duncan’s top playoff win share years, they won the championship while Kobe lost in the finals during his third and fourth highest win share seasons (2004 and 2008). Obviously this statistic is inflated by the fact that they made it to the finals, so by proxy this is also a measure of how far their teams went. Duncan also had the highest Win shares during the Spurs 2004 championship run, higher than any of Kobe’s win share seasons. With respect to this statistic, Kobe had 3 out of the 5 highest Win share playoffs out of the two players.Once again, the p-value here was 0.6642 indicating no significant difference. Let’s look at WAR again to compare.

WARPLAYOFF

Duncan by far has the highest WAR during his second championship while Kobe’s 3 year stretch to the finals (and 2 ships) in 2008-2010 were the second 3 best out the two stars. Again, we don’t have a significant difference due to the Wilcoxon-Signed rank test but can be in awe of their career numbers.

I’m going to take a look at one final advanced stat for the playoffs, that gives the biggest nod to Duncan with a p-value of statistical difference of 0.0857, the Box Plus-Minus. In short, it represents the number of points more than an average player per 100 possessions on an average team.

BOXPM

It’s no wonder Duncan has a slight significance for this statistic. Since this statistic incorporates both offensive and defensive efficiency, Duncan has four better playoff runs than Kobe’s best in terms of BPM. Kobe’s first two playoffs resulted in a negative BPM, but he hadn’t flourished to be the superstar talent he became.

So I don’t think you can make a definitive conclusion about who was better, at least statistically. Since they played different positions, the comparisons based on pure stats alone doesn’t do the comparison justice. The advanced stats show different years of dominance (for both regular season and playoffs) but no huge statistical difference (slight nod for Duncan in terms of playoff BPM). Together they have 3 MVPs, 5 finals MVPs and 10 NBA championships.

So while stats don’t point to one clearly (both basic and advanced stat) I’ll give my own person bias here. If I am drafting a player to build a team around and win, I’ll take Duncan. However, If I want to sell tickets, merchandise and win, I’ll take Bryant. Who was the better player, statistically it’s not clear, but I’m taking Duncan (even though the spurs weren’t as fun to watch).

If you enjoyed this post, take a look at some of my other Adventures in Statistics.

-Andrew G. Chapple

 

 

Advertisements

2 thoughts on “Fun with Data: Kobe or Duncan? What do the numbers say?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s