NanoPress

Stats for Dummies: episode 7 – PER

Stats for Dummies: episode 7 – PER

In the seventh episode of Stats for Dummies we discover the PER.

Welcome back to Stats for Dummies. In the last episode, we introduced individual statistics. Today we continue our journey with another individual statistic, that is the PER, an acronym that stands for Player Efficiency Rating. Here we go!

PER, like Rating or Box Plus Minus, is a statistic that tries to convey all the player’s contributions into a single number. The human mind, although he knows that joining together more contributions is complicated and “dangerous” (because you can lose important points of view for an analysis), try to classify everything in the simplest way. This is why several “all in one” statistics have been created: thanks to them it is possible to classify players and teams easily. The PER is an example of this type of stat but if you do not understand it, you can run into some mistakes.

I like to consider the PER as the evolution of the classic efficiency, that is the sum of positive contributions (points, assists, rebounds, blocks, etc.) to which the sum of the negative contributions is taken away (missed shots, lost balls, fouls, etc). However, the efficiency does not take into account different factors, such as the Pace and the minutes played by the player. It is a simple sum and so can not take into account these factors.

The PER is born with the same idea of efficiency (sum of positive contributions minus sum negative contributions), but also considers some factors, so that it results more appropriate for an evaluation of the players’impact on the court; that’s why I like to call it the evolution of efficiency. Moreover, the real peculiarity of PER is that is standardized: averaging a PER of 15 is equivalent to being, within the League, always an averaged player; instead, a 25 is equal to being a potential MVP.
But let’s start to understand well how John Hollinger’s stat is born.

The PER takes into consideration the following positive contributions:

  • 2 and 3 shots made (you will give a different weight to the two types of baskets);
  • Free throws made;
  • Assist;
  • Defensive rebounds;
  • Offensive rebounds;
  • Steals;
  • Blocked shots;

And the following negative contributions:

  • Missed shots;
  • Missed free throws;
  • Lost balls;
  • Fouls;

As you can see, Hollinger has decided not to consider the drawn fouls. And also all intangibles are not counted in the PER because it is designed to be calculated from the box score.

The beauty of PER compared to the efficiency is that the various contributions are not simply summed: each of them is converted into a value that takes into account different game aspects; for example, the shots made are multiplied by a coefficient that takes into account how many assisted baskets are made within the entire League, in order to give a different weigh to the baskets scored with an ISO or after a mate’s pass. The points for possession of the entire League are taken into account for the missed shots, the turnovers, the steals, the blocks, the fouls, and the rebounds, in order to give a specific weight to all these contributions (in a league where the scores are high, the lost balls will have a heavier impact). Rebounds are also combined with their advanced stats (DR%, OR%) because a defensive rebound in a league with DR% at 70 will have less weight than one in a league with DR% at 55. So, each contribution is carefully evaluated with respect to the average level of the League, in order to give a different importance, which is not in any way carried out in the efficiency.

This particular sum is then adjusted with respect to the average team Pace: this is because a player who plays in a high-pace team will have more opportunities to generate contributions since he has more possessions. In this way, the contributions of all the players are compared with the same Pace.

There is one last step to find the true PER that also implies the characteristic of being standardized: the final statistic considers the contributions of all players and their average minutes played to define the average level, which is always set to 15. In fact, the final calculation gives a number that usually ranges between 0 and 45 and each number corresponds to specific performances:

More than MVP: 43 or more
MVP: 33 – 42.9
A strong candidate for the MVP award: 30 – 32.9
Second tier candidate for the MVP award: 25 – 29.9
An All-Star player: 22.5 – 24.9
A possible All-Star player: 20 – 22.4
Second offensive option: 18 – 19.9
Third offensive option: 16.5 – 17.9
Average player: 15 – 16.4
Bench player: 13 – 14.9
Bench addicted: 11 – 12.9
Deep bench lover: 9 – 10.9
How is it possible that he is on the team?: 0 – 8.9

Therefore, within the same League. it is possible to classify all the players with respect to their contributions, without having to control the usage and Pace and moreover, the value will declare us immediately how good the player’s contribution is. I often say “within the League” for a very specific reason: whether for my amateur championship or for the NBA, the “average” player of those competitions will always have a PER of 15. However, it is clear that an amateur player with PER of 15 will never be comparable to an NBA player with the same PER. The scale is the same, but the starting point is different.

The PER is a useful statistic to classify the players, but it is clear that it does not take into account different invisible contributions, which are very important with the team. Take for example OKC of the past season.

Andre Roberson, until his injury, had been a key factor for Donovan and his teammates because of his defense. Defensive abilities such as lateral speeds or the skill to anticipate the opponent’s movements are impossible to obtain from the box score as they are not tangible contributions. Here then, through the PER stand out more the shooting problems of Andre and not his defensive skills: this is the limit of PER. Although we can consider all the possible contributions, some remain hidden and therefore we do not get the right value for each player. Another example is Adams: despite having a PER of 20 due to the rebounds that grabs, the box out and screens are not counted. Most likely he would have a higher PER if these were counted too.

In conclusion, as for each statistic, some players’contributions cannot be contemplated do not result in the final value of the PER: the important thing is to remember this fact and not just say “Roberson is useless because he has a PER of 10”. It would be one of the wrongest statements to say.

Post your comment