The Novice's Guide to Advanced Stats Part III: Basketball

Posted by Taylor Nigrelli on February 20, 2015 · 7 mins read

If you are a basketball fan, you might have come across this recently.

https://www.youtube.com/watch?v=NZf9NFaCQHQ

Barkley is wrong 40 ways from Sunday, but the rant had one false, central theme: the idea that advanced stats are some witchcraft, unrelated to the actual game being played. This is just not true.

As stated in Part I, stats are just a log of what happened in a given game, season or career. His quote “Daryl Morey is one of those idiots who believes in analytics” is essentially saying “Daryl Morey is one of those idiots who uses information to make decisions.”

See how insane that sounds when you phrase it that way? And that’s the issue really; he has no idea what analytics are or how using advanced statistical analysis would manifest itself on the basketball court. That’s clear when he asks “what analytics do the Spurs have?,” as though a team that properly uses advanced metrics doesn’t also need a group of talented basketball players.

Straw men and obvious unwillingness to be reasonable aside, Barkley does raise an interesting point: what effect does advanced statistical analysis have on the NBA and the way it’s covered?

Let’s take a look at a few ways. (The NBA has taken to advanced stats almost as well as the MLB, so I won’t be able to cover all the innovations. I’ll just outline a few ways players, coaches, scouts and analysts view the game differently in the 21st century.)

Starting with something relatively simple that ties in with Barkley’s rant, per possession numbers are an easy way to add context to a team’s offensive or defensive stats. When showing the standings, most websites show each team’s points per game and points allowed per game. But that can be misleading because teams play at different paces, which can lead to more or fewer possessions. Thus, it’s wise to consider points scored per 100 possessions and points allowed per 100 possessions as measures of efficiency.

This can extend down to individual players to. Looking at a team’s points per 100 possessions and points allowed per 100 possessions with and without a certain player on the court can tell you a great deal about that player.

Barkley calls Houston’s defense “the worst in the NBA”, partially based on them giving up 118 points in one game. Forgetting Barkley’s ridiculous sample-size issues, one would be misled if they looked strictly at raw points totals for Houston. The Rockets allow 99.8 points per game, which ranks 18th in the NBA. But they’re eighth in points allowed per possession at 100.3 per 100 possessions because their games tend to feature more possessions due to their or the opponent’s pace. That’s a major difference, and it illustrates how important it is to consider the right statistics.

PER – “Player efficiency rating” was invented by former ESPN employee John Hollinger, who currently works in the Memphis Grizzlies front office. PER attempts to measure a given player’s entire on-court impact in one stat. The formula behind the stat is fairly complicated but it measures a player’s per-minute performance and adjusts for pace. Hollinger admits PER has some shortcomings when measuring defensive performance, but is generally considered the best measure of offensive efficiency.

Three-point barrage – One of the major on-court innovations advanced stat analysis has led to is an increase in three pointers taken and made. One of the earliest analytics theories in basketball was that teams should take significantly fewer mid-range shots and replace them with threes. The idea is that it’s only slightly harder to hit a three-pointer than a long two, and the three-point obviously counts for more. Since anywhere from inside the arch counts for the same, it stands to reason teams should look for an open shot as close to the basket as possible.

Houston Rockets general manager Daryl Morey – he of Barkley rant fame – was one of the driving forces behind this style of play. His NBA team is built on threes, drives to the basket and low-post play. He took this a step further with Houston’s Development League team, doing away with mid-range shots all together.

Quite a few teams rely heavily on three-pointers now as the number of made three pointers has increased drastically in recent years. As the previous link cautions, however, defenses will likely begin to adjust to this soon, possible making mid-range shots slightly more important than they are now.

So, to answer, Barkley’s question from the rant, that is the “analytics” that the Spurs “have”. They move the ball well, finding open shots in high-percentage, high-efficiency areas. Of course, they need talent to do this, but they get the most out said talent by being smart.

Real Plus-Minus - This stat, which ESPN introduced prior to last year’s postseason, is an adjusted version of the regular plus-minus statistics. The issue with traditional plus-minus is that it can be skewed by the relative strengths and weaknesses of teammates and opponents. According to ESPN “real plus-minus estimates how many points each player adds or subtracts, on average, to his team’s scoring margin for each 100 possessions played.” This is a good example how to apply advanced stats: improve on an old statistic to get a better idea of player’s relative strengths.

SportVU – Much like Pitchf/x in baseball, SportVU is a system of high-tech, data-collecting cameras. The cameras, located strategically throughout the arena, track the movements of teams, individual players and the ball and collects a piece of data on each 25 times per second. The cameras both tape the game and record the stats themselves.

The cameras identify objects on the court and get “positioning data” for the players and the ball. In plain English, the cameras record each thing a player does and the exact path the ball takes. Each piece of data recorded is immediately uploaded to a database with a time stamp and the player in question’s name. This enables the cameras to track such information as field goal percentage by location, the percentage of passes that lead to a field goal and location tendencies by team among dozens of other innovations. This creates an incredible data base to draw information from and has completely revolutionized the way basketball is viewed by those who know most about the game. There’s no longer any need for guesswork or assumptions, there’s an infinite base of information available.

(Information courtesy of stats.com/sportvu)