Stats for Dummies - Episode 4: Ratings

Welcome back to Stats for Dummies: in this and in the following episodes we will go into some advanced statistics, as we did in the last episode with the rebounds. We will understand with the help of tables and graphs why the advanced statistics are more suitable for an analysis.

Question: when you want to understand how a team plays, what do you look at? At first, points made and allowed. In fact, points are used to determinate winners and losers, but they are also the reference value for the evaluation of a team.

However, we have found that comparing terms "per game" hides some pitfalls: even the statistics related to the points are not excluded. The points per game (made and allowed) do not take into account a basic element which, if not considered, leads to inaccurate evaluations: the game speed.

Many situations occur during the 40 minutes of a game: shots made and missed, lost balls, drawn fouls and free throws. The amount of these events within the playing time is determined by the game speed. For example, if two teams play 16" per possessions, more or less we will have watched 150 possessions, 75 for each team. If instead, the same teams play every second of a possession (24"), we will have watched about 100 possessions, 50 for each team.

Therefore, scoring 80 points in a game where you have 75 possessions is very different than making the same amount with 50 possessions.

This is the pitfall behind the points (made or allowed) per game: this statistic does not take into account the game speed and consequently leads to incorrect evaluations. To overcome this, there are the Ratings, already mentioned in recent episodes and in some analysis. What are the Ratings? Simply a redistribution on 100 possessions of the points made or allowed: they can be obtained by dividing the points by the possessions and then multiplying this ratio by 100 to get the potential points that a team would score or allow if it had 100 possessions available. Therefore, using the ratings we have the opportunity to evaluate more teams, even if they play at different speed.

When I speak of possessions, I am referring, of course, to statistical possession, that is the sum of made and missed shots and lost balls, minus the offensive rebounds.

So, let's take an example taking as a reference the past NBA season regular:

The above table shows the points scored per game on the left side, while the Offensive Rating on the right side.

I have highlighted some examples that make perfectly the idea of how points per game can lead to inaccurate evaluations. For example, look the Phoenix Suns or the San Antonio Spurs: they are respectively at 21st and 27th position. Actually, with a statistic suitable for this type of analysis, we find out that the Suns are last while the Spurs are 17th. They switched their positions. Another example is the LA Lakers, that drop down from 11th to 22nd.

What the Rating allows (in this case, Offensive Rating because it refers to the attack) is to take into account the average pace, in order to standardize the points of each team on a reference pace (the famous 100 possessions). In this way, teams that play at high or low pace can be compared without mistakes. This analysis is not possible with the points per game alone.

In fact, if you look at the table below you can find out that Phoenix and San Antonio have opposite pace. Comparing points per game would seem that the Suns score more than SA, but actually is the opposite.

Pay attention to Rating: we are not talking about beautiful of game, shots distribution or other elements: Ratings are statistics that simply communicate the effectiveness of a game system. In other words, they give us information about the skill of the attack (or defense in the case of the Defensive Rating) in making a basket (or not to allow). To understand how they play, there are other tools.

There is another peculiarity that the Ratings allow to take into account, unlike the points per game. Look at the positions of Portland and Indiana in the two previous tables. Both scored an equal amount of points during the season, around 105 points, and averaged a similar Pace (about 98 possessions per game). So, why do we find out that Indiana is at 12th position and Portland is only 15th in the Rating rank?
This is the second peculiarity of the ratings: the two franchises have averaged the same score and pace, but the difference lies in the possessions used to divide the points made.
The possessions, as we know, are the sum of made and missed shots and turnovers. If we look at the number of lost balls per 100 possessions we discover that Indiana has lost fewer balls than the Blazers and, at the same time, has shot better. These details seem trivial, but they influence the Pace and consequently the Ratings too. In other words, the total number of possessions is the same, but the composition of the two paces has been different and this is highlighted in the Offensive Rating.

We could do other examples, but the general argument is that thanks to the Ratings we can compare more teams without having to consider other statistics.
Beware of those who say that team X is the best attack because it scores more points: if you do not take into account the pace and its composition, you only make incorrect evaluations. That's why ratings are the best option for this type of evaluation.

Article by Luca Cappelletti