Stats for Dummies: episode 5 – Assist

Stats for Dummies: episode 5 – Assist

Nel quinto episodio di Stats for Dummies andremo a scoprire le statistiche avanzate relative agli assist.

Welcome back to Stats for Dummies! I have to apologize for the missed article last Friday, but there have been technical problems. Said that we continue our journey to discover the differences between traditional and advanced stats. Today we will talk about assist.

The assist is an interesting statistical tool because it reflects the willingness of the teams to play together. In fact, generally speaking, higher is the value of the assist, better will be the offensive plays. Obviously, this is not always the case, but it is an assumption that can be accepted more or less.

However, the number of assists is fallacious: tendentially the assists range in a game is between 15 and 20, with little variations among the various leagues. Having numbers so close it difficult to understand which teams distribute more assists. Moreover, there is another problem: tendentially 50-60% of the shots made are assisted. This happens in every team, with little differences that make the just mentioned percentages oscillate in that range. This fact translates into an influence of shots made on assists: in other words, the assists are dependent on the shots made. Is it trivial to say? Absolutely, but this aspect is not actually always remembered. This correlation can be observed intersecting shots made and assist per game:

As you can see, by distributing the shots made on 100 possessions and intersecting them with the assists (also redistributed on 100 possessions) you can see a clear trend: higher are the shots made, higher will be the number of assists. We obviously have some special cases, but the trend curve confirms this thought.

Taking into account this correlation, when making a comparison among several teams it is easy to run into some misjudgment. In fact, if you remember what has been said in the previous episodes, we are not interested in the number of assists distributed (the absolute value), but rather the percentage of assisted shots (the relative value).

That’s why there is the Assist Percentage, which is the ratio between the number of assists and the goals made: a simple relationship, which greatly improves the analysis.
Performing this simple division frees us from the just demonstrated dependence:

The above chart shows a very different trend: this is because now the dependence between goals made and assists no longer affects the chart and so we can immediately understand that between Khimki and Baskonia is the Russian team the one who has distributed more assists. Another example is that CSKA was strongly influenced by the many goals made, but actually distributes fewer assists than the league average. Yes, in some cases they are small differences, but it is always better to analyze data not influenced by other aspects of the game.

Therefore, the Assist Percentage is a useful value to understand how a team distributes the assists within the game. This statistic is also available for the players: through the individual Assist Percentage we can find out how many teammates’ goals have been assisted by the player.
And when do you want to compare players with different minutes played? In fact, a player who remains on the court 30 minutes will certainly have more chance to perform some assists than one who plays 10 minutes. The Assist Percentage makes it clear how many teammates’ baskets are assisted by the player, but not how he uses his possessions.
To understand this we have to use to the Assist Ratio: this statistic takes into account assists, turnovers and goals attempted of the player and redistributes them on 100 possessions. Consequently, a player with a high value of Ast Ratio will certainly be more inclined to pass than to shoot. This comes back very useful to understand who are the most “playmaker” players.

In conclusion, it is clear that observing the single value of the assists per game is not enough to have a clear vision of how they are distributed, both by the teams and by the players. Having tools like Assist Percentage and Assist Ratio is a shame not to take advantage of it.
A similar discussion can be also made for turnovers: through the Turnover Percentage we have a relative data, that is the relationship between the lost balls and the possessions played. With this statistic, analogous to the Assist Percentage, we can understand who loses more ball and it is an useful information and certainly more precise than the number of lost balls. I will not repeat the charts and the explanations also for turnovers, they would highlight the same defects of assists. Turnovers and assists are very similar and, therefore, it’s better to use advanced stats in order to avoid some mistakes during an analysis.

Post your comment