This is not a typical post, but I was thinking about basketball the other day and I got kind of interested in a few topics.
This started when I recently saw that Kobe Bryant is averaging over 30 points per game, at the age of 34. This seemed strange to me, and admittedly being somewhat biased against him, I chalked it up to the fact that he was a ball hog. But I wanted to look at the data. I was interested in two items: (1) How has Kobe Bryant’s shooting (at least in the regular season) evolved over time? (2) Has it become more or less variable over time?
Rather than look at all field goal attempts and all field goals made and then simply coming up with a regular season average, I thought it would be better to take field goal percentage from each game and weight each game equally. I also thought it would be interesting to look at the median and the 25th/75th percentiles for each regular season. I see lots of basketball statistics, some of them great, but I don’t see lots of graphing, which is unfortunate as they can be great at conveying information. I used data from http://www.basketball-reference.com/.
First, I plot Kobe Bryant’s median field goal percentage for each regular season from 1998-1999 to 2012-2013 (31 games so far). The green bars give you the 25th and 75th percentile of his shooting. In other words, 50% of the games will fall between the two bars. I do not include Kobe Bryant’s first two seasons, as he was not a starter and had strange minutes/shooting, compared to the rest of his career.
A few things stand out to me. First, his median field goal percentage is very consistent over time, right around 46%. There is of course one exception: the current 2012-2013 season, where Kobe Bryant is shooting 50%. The interquartile range in the current season is on the lower end of his career, but it is consistent. The standard deviation of his field goal percentage is also roughly the same. But his shooting is much better. Kobe Bryant, at age 34, is playing the best season of his career if we are to look at accuracy. For a shooting guard, this is a good (but not perfect) measure to use.
The thought then occurred to me: maybe Kobe Bryant always shoots this well in the first 31 games of a season, and then it drops off. So I produced the same chart looking only at the first 31 games of each season.
Here the picture is a bit more nuanced. This season is still impressive in a few ways. He has shot this well for the first 31 games a few times before, but he has been much more consistent if we pay attention to the inter-quartile range. If we put weight on both the field goal percentage and the consistency, then this is still the best season of Kobe Bryant’s career. Of course, things could change dramatically over the course of an 82-game season, and none of this takes into account the many other factors that might be having an influence here. In any event, I’ve misjudged Kobe Bryant, at least this season.
Comparing Kobe Bryant and Lebron James maybe doesn’t make a lot of sense. Different players, different roles on the court, doing different things for their teams. But superstars are superstars.
What we see is pretty consistent improvement in shooting percentage over the course of Lebron’s 9 seasons. Recall that this area around the median is where 50% of his games have fallen into. What’s most remarkable is that this region, which I’m using as a measure of consistency, has narrowed considerably from the 2011-2012 season. Prior to the current season, the interquartile range (IQR, the top green bar minus the bottom gren bar) ranged from about 12 to 16 percentage points. In the first 31 games of the 2012-2013 season, this has dropped to about 9 percentage points.
Again, to check to see if this is not a feature of the first 31 games of every season, I look at only the first 31 games of each season.
There are many other things one has to look at: turnovers, assists, and of course wins and playoff outcomes. I’m aware there are many complicated statistical measures that do a good job at measuring the total impact of a player, but I just wanted to look at some basic questions on my own. I hope to explore more basketball data soon.