NBA Analytics aren't the end all
By Jamie Wianecki
John Hollinger came out with his Player Efficiency Rating (PER) he earned himself a front office position with the Memphis Grizzlies in 2012. Since then, there has been an abundance of basketball analytics. And they are really fun to look at. Plus-Minus shows the team’s point differential when each player is on the court. Offensive real plus-minus shows each player’s offensive impact per 100 possessions. Defensive real plus-minus shows each player’s defensive impact per 100 possessions. Real plus-minus (RPM) is the individual player’s difference offensive real plus-minus and defensive real plus-minus. I am not here to argue against all advanced stats, but what I am arguing is there are some serious flaws when only considering advanced stats.
For example, let’s go through RPM. Here are the top 10 players in RPM:
As you notice, there are a lot of names who you would expect to have made an impact on the game. Some people wouldn’t expect to have Jokic or Golbert in the top 10, but nothing too crazy. You’d probably expect Kevin Durant to be in there but he’s coming up and with two Warriors already in there, his impact is slightly muted.
Now here is the next 10 (11-20):
Other than Durant being this low, the next biggest surprise is Jae Crowder being 20th. Crowder is a very nice player, but he is not in the top 20. If you were to start the NBA over and do a draft, Jae Crowder would not be the 20th guy drafted. Crowder hasn’t made and most likely will never make an all star team. Going by this list and considering 5 teams have 11 of the top 20 guys, Crowder would be the best player on 17 teams. Crowder is a nice player, but he’s not that good.
The next 10 is where it gets crazy and unreliable. (21-30)
Otto Porter Jr.
In this group we have a number of role players. Amir Johnson, Robert Covington, Cody Zeller and Zaza. All of those guys have made a positive impact on their team, but for a stat to put them in the top 30 in the NBA doesn’t pass the smell test. All of them earned their way to the top 30 because of their defensive performance. Let’s dive into Robert Covington, who finished 16-17’ with the 25th best RPM. The 76ers won 28 games in the Eastern Conference last season. No other player on the 76ers roster made the top 40 in RPM. Covington’s defensive rating, which is the number of points given up per 100 possessions, was 105—which is above average, but not even in the top 20. Out of all the players on the 76ers roster who played at least 50 games (10 players) he was the best in defensive rating. 6 of those 10 players had defensive ratings over 110. 4 of those 6 guys were guards, Timothe Luwawu-Cabarrot, Sergio Rodriguez, Gerald Henderson and Nik Stauskas.
In addition to the lack of defensive prowess, the 76ers were 30th in team offensive rating, scoring 103.2 points per 100 possessions. If they had a better offense, opposing teams would have needed to score more points. Considering the 76ers lost 54 games, 11 by 20 points or more, had they played closer games the opposing team would have needed scored more points, which would have raised Covington’s DRPM.
So my big issue with Covington being in the top 30 is his impact wasn’t more important than Damian Lillard, who was 28th in RPM, or Gordon Hayward who was 29th in RPM, or Paul George who was 37th in RPM. All these guys were on playoff teams and had to shoulder the bulk of the offensive load while their teams played more meaningful games. Covington may have not been a detriment to the defense on a 28-win team, but he also shot 39.9% from the field with a 53% true shooting percentage. So you are telling me, the defense that Covington played on a 28-win team was more impactful than Paul George’s overall game on the 42-win Pacers? Or Gordon Hayward’s overall game on a 51-win Jazz team? I can’t buy that.
Another issue I have with advanced stats is estimating numbers per 100 possessions and not factoring in minutes per game. So looking at the top 20 guys, there was only one player who played less than 30 minutes per game. Nikola Jokic, who is the advanced NBA statistic fan’s darling. The 22-year-old had a great second season and is off to a promising career, but what he did was not as impressive as Jimmy Butler, who Jokic had a slightly higher RPM, but played about 10 minutes per game less. No. There needs to be a factor in the equation that values the more minutes a player plays per game in addition to per 100 possessions.
If I’m doing my top 10 list of best NBA players from last season it’s
- 1. LeBron James
- 2. Kevin Durant
- 3. Steph Curry
- 4. Kawhi Leonard
- 5. Chris Paul
- 6. James Harden
- 7. Russell Westbrook
- 8. Giannis Antetokounmpo
- 9. Anthony Davis
- 10. Draymond Green
- 90. Robert Covington