Your team is down 1-2 with 15 minutes remaining. You desperately need a goal to draw, or perhaps, are going for the win. If you could have any player on your bench to bring on, who would it be?

A tall, strong, target man to get on the end of a cross for a cheeky header? A pacey winger who can dribble down the line and deliver the perfect cross? A creative midfielder to bring on fresh legs to thread the perfect through-ball? Lets take a look at the data.

Click on the image to go interactive


A few housekeeping notes:

  • “Super-sub” is defined as a player who is brought on to make a difference going forward (attacking). Thus, we will use goals and assists to judge the player’s impact.
  • We define “best super-sub” as the player with the highest impact rate. The “impact rate” is defined as total goals and assists divided by number of substitutions.
  • The data includes top 25 substitutes from each of the top five European leagues over the past 5 seasons.
  • Some players who have played for clubs in multiple leagues over the past five years may have more goals than stated in the model. This is because the player might have been a “top 25 substitute” (according to transfermarkt) in one league, but was not in the top 25 in the other league. This does not change anything drastically in the data represented as the goals in the other league are too small in number to make a difference, except for…
  • Two players were in the top 25 list in two different leagues: Álvaro Morata and Kévin Gameiro. Their goals and assists have been aggregated.
  • I highly recommend viewing the visualization on the largest screen possible (it may not be the best experience on mobile)

Filtering Outliers

Now, I know what you’re thinking…

“Lewandowski? Messi? Agüero?! They’re not substitutes! Fake news!”

Yes, you are correct. We need to filter the data to remove outliers in order to find the real super-subs. This is where we can use the “Number of Substitutes” slider to filter out players who have fewer than a certain number of substitutions.

Lewandowksi stands out with his absurd impact rate of 0.82. Yes, you guessed it, this is because of the day he decided to go super-saiyan and score 5 goals in 9 minutes against Wolfsburg as a substitute, thus inflating his impact rate. Messi is second with an impact rate of 0.75, but that is because he has a very low number of substitutions (12) and 9 goals. The story is the same with other players who are not “substitutes” but appear in the data.

Let’s see what happens when we filter them out using the slider:

  • 30+ substitutions: When we take out players who have less than 30 substitutions, the data makes a little more sense. The leaders based on impact rate now are Francesco Totti (0.53), Frank Ribéry (0.47), and Nils Petersen (0.45). While this makes more sense, we can go a step further.
  • 49+ substitutions: At this point, we find the real super-subs. The leaders here are Álvaro Morata, Kévin Gameiro, and Edin Dzeko. As mentioned before, Morata and Gameiro have been incredible substitutes in two separate top leagues. Morata (0.38) has an impressive 12 goals and 1 assist at Real Madrid and 8 goals and 6 assists from his time at Juventus as a substitute. Gameiro (0.28) had 6 goals and 1 assist at PSG and 11 goals and 3 assists at Sevilla and Atlético Madrid as a substitute.
  • Honorable Mentions: Since the data only ranges from the 2011-12 season to present, some players were – for the lack of a better word – screwed. Notably, this data range shows that Chicharito only had 9 goals and 1 assist. If we were to go back one more season, Chicharito would be the best super-sub in the EPL with 5 additional goals to his name. Furthermore, Jermain Defoe has only 7 goals in this data range. If we were to go back all the way to 2000-01, Defoe stands out with 23 goals and 1 assist. Although he is not higher up in this data range, it is worth mentioning that he holds the premier league record for most goals as a substitute.
Alvaro Morata scoring against his own team Real Madrid in the Champion’s League while on loan at Juventus (Source: Getty Images)

I recommend you play around filtering by different leagues and number of substitutions to see who’s the best substitute. We can see that in the premier league with 35+ substitutions, Olivier Giroud and Victor Anichebe are tied with an impact rate of 0.33, followed by Dzeko, Milner, and Chicharito. Milner is interesting in that he has more assists than goals, whereas most of these players have significantly more goals than assists.

TL;DR: Álvaro Morata and Kévin Gameiro?

The graph on the right shows the distribution between goals and assists.

Who would you sub on if you want a goal? Morata stands out with 20 goals, followed by Gameiro with 17 goals (Edin Dzeko, Chicharito, and veterans Souleymane Camara, Defoe wouldn’t be bad choices either). What if you want an assist? Dries Mertens, Romain Hamouma, and James Milner stand out.

I hope this data visualization helps you see which players have had the highest impact coming off the bench. We need to keep in mind that this data does have flaws: players’ who scored in multiple leagues but weren’t within the top 25 in each league have underrepresented goals, the data covers only the past five years (Jermain Defoe would be a force to reckon with if we went back further), and the data is only for Europe’s top five leagues. This visualization is not meant to be a perfect, super-detailed, mathematical model – my aim to help illustrate data in an attractive manner which viewers can interact with without getting overwhelmed.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s