#001 Scatter plots: Using the per ninety metric, Goals per 90 and Assists per 90

Kiplagat Seroney
5 min readNov 11, 2021

--

INTRODUCTION

I am new to the sports analytics world. More on this on my first Medium Article.

I have been following many tutorials and learning a lot. However, I did not want to copy & paste data and information and publish it for the sake of it without even understanding and comprehending it. I will use this Medium page to talk about my understanding of Soccer analysis metrics and use publicly available data to spark conversations and try to derive insights.

With that out of the way. Lets Go!

What is Per 90 ?

The per ninety minutes metric (per 90) is important tool used in standardizing the number of minutes played with respect to 90 minutes (average length of a soccer match). Examples include, Goals per 90, Assists per 90, Expected Goals per 90 among other applicable metrics.

Scatter relationships

Scatter plots helps us under stand the correlation of two quantifiable variables and how they move in relation to each other.

Today we will look at the relationship between two variables namely, Goals per 90 and Assists per 90.

THE 2020/2021 EPL SEASON

The linear graph below attempts to show how the top 4 of the 2020/2021 converted of goals per 90 with respect to Assists per 90

ASSUMPTIONS

i. We are using a trend line to determine the strength and direction of the relationship of between Goals per 90 and Assists per 90.

ii. Assists are perfectly positively correlated to goals thus assists translate to goal in Football(soccer). This is a soccer fact strengthen mathematically with this plot.

iii. This Scatter plot attempts to show how effective a team converts its assists per 90 to goals per 90.

INSIGHTS

i. Chelsea (№4) vs Tottenham(№7)

From the plot above Tottenham outperformed Chelsea on both Goals and Assists per 90. However from last season’s table Chelsea finished Number 4 with 67 points and proceeded to win the Champions League. (#weknowwhoweareeee….), whereas Tottenham finished 7th with 62 points.

Tottenham had a higher Goal Difference than Chelsea, less draws than Chelsea but one important variable is losses. Chelsea had 9 losses while Tottenham had 12 losses.

ii. Chelsea (№4) vs Leicester(№5)

Chelsea had 19 wins while Leicester had 20 wins. Leicester had an impressive 6 draws whereas Chelsea had 10 draws. However, like Tottenham, Leicester had 12 losses.

iii. Tottenham (№7) vs Leicester(№5)

The above teams have an interesting relationship. Tottenham outperformed Leicester in Goals per 90 and Assists per 90. A detailed view of the data provided by Statsbomb via FBREF , indicated the goal difference variance is +5 for Tottenham. The difference between the two teams is that Leicester won two more games and drew two less games than Tottenham.

Therefore, focusing on the conversion of assists per 90 to goals does not translate to points. This is a demonstration to indicate this. More goals, more assists and more wins with the correct balance of draws and wins does it. (Example is my beloved Chelsea)

iv. Bottom 5 analysis

From the plot above, we look at the position of the bottom 5 teams along our scatter plot. We derive the following insights.

i. Five out of six teams that lie on the least conversion, were bottom five.

ii. The sixth team here was Wolves. What did they do different to be number 13 in the PL standings?

THE 2021/2022 SEASON (FIRST ELEVEN GAMES)

Using the previous demonstration, I would like to repeat the same using the 2021/2022 season’s data so far ( eleven matches played so far).

From the figure above, we can observe;

i. Liverpool has the highest correlation between Assists per 90 and Goals per 90 . From my POV, Liverpool is a dangerous team so far. Regardless of the last two games (a Draw and a Loss). Their relationship between Assists per 90 and Goals per 90 indicate how dangerous they are.

ii. Liverpool has already established that they are converting assists per 90 to goals per 90. This is a strong indication just like Manchester City in the 2020/21 Season who were leading in this metric.

iii. Despite Chelsea being the Table toppers as at Match 11 of the season, due to with more wins than Liverpool, Liverpool has more goals and assists than Chelsea but due to the high scoring draw Liverpool had and the 7–0 thrashing Chelsea had, indicates that the League is only getting started. (Lets get our popcorn ready for a blast)

CONCLUSION

This is a deliberate attempt in trying to relate two variable, goals per 90 and assists per 90 and how the relationship of the two can be used to see how various clubs perform using this metric.

I personally believe this metric can be used to indicate if the club is heading towards the right direction. (I still have a lot to learn from the community since this is my substantive article ever). I will endeavor to continue updating on this.

Feel free to hit my DM’s on Twitter and Instagram as well as my email, deltaanalyticske@gmail.com to further the discussion as well as input, addition and comments.

I would like to thank the Tiotal Football and Mckay Johns for the insipiration through their work and Fc Python for the club’s color code using their team colours tool . A special Thanks to datawrapper for providing the tools for analysis and visualizing.

Cheers.

--

--

Kiplagat Seroney
Kiplagat Seroney

No responses yet