Three years ago, former HSAC Co-President Brendan Kent wrote a piece on aggression in Premier League derbies. In the article, Brendan looked at 8 derbies over a 5 year span and concluded that there is significantly more aggression in these rivalry games than in other Premier League matches. This article got me thinking, does this trend hold over a longer time horizon? Is there a similar effect in other major European Leagues? And finally, are there any specific rivalries that are extra aggressive?
To start, I picked the five biggest rivalries in the English, Spanish, German and Italian top flights. These rivalries were chosen somewhat arbitrarily through Google searching, with preference given to rivalries that have featured quite frequently in the top flight. The French League was not chosen because it was hard to pinpoint five true rivalries (PSG vs Marseille and St. Etienne vs Lyon were obvious choices, but finding three others was incredibly difficult). Other leagues were not chosen because the data was not available. I relaxed my definition from derby to rivalry, so major matchups in each country (Manchester United vs Liverpool, Borussia Dortmund vs Bayern Munich) could be included that don’t fit the traditional definition of “derby.” The following rivalries were chosen:
Premier League: Manchester United vs Manchester City, Liverpool vs Everton, Arsenal vs Tottenham Hotspur, Newcastle United vs Sunderland, Manchester United vs Liverpool
La Liga: Real Madrid vs Atlético Madrid, FC Barcelona vs Espanyol, Sevilla vs Real Betis, Athletic Bilbao vs Real Sociedad, Real Madrid vs FC Barcelona
Serie A: AC Milan vs Internazionale, AS Roma vs Lazio, Genoa vs Sampdoria, Catania vs Palermo, Juventus vs Internazionale
Bundesliga: Hamburg vs Werder Bremen, Borussia Dortmund vs Schalke 04, Bayern Munich vs Nurnberg, Borussia Mönchengladbach vs FC Koln, Borussia Dortmund vs Bayern Munich
I collected data from Football Data, calculating the number of fouls in each match in each of these four leagues from the 2005/6 to the 2016/17 season. For each league, I separated the data into rivalry matches and non-rivalry matches (note: for Serie A, there were a few matches that did not include foul statistics, and those matches have been omitted from this study), and calculated the mean and standard deviation of the number of total fouls awarded in these two datasets. I calculated the difference between the number of fouls between rivalry and non-rivalry matches and the ratio between the two (in case one league had fewer fouls awarded and throwing off the scale). Finally, I conducted a t test between the fouls in rivalry and non-rivalry matches for each of the four leagues, and recorded the T Stat. My results were as follows:
Although not the scope of this article, it is interesting to note how few fouls are called in the Premier League compared to the other three leagues, perhaps that is something that needs to be considered further in a separate blog post.
The first thing to note is that we confirm Brendan’s research from 2015. We find that there are significantly more fouls called when Premier League rivals meet than in the average Premier League match, and the t statistic is incredibly high.
We can also note that in La Liga and Serie A, there is strong evidence to suggest that there is extra aggression when rivals meet, although not on the same scale as in the Premier League.
Finally, we can note that with the low T statistic of .7, we cannot conclude that there is extra aggression in Bundesliga rivalry matches. This is an interesting result, and another result that can likely be delved into further.
Next, we will try to determine what the fiercest rivalry is in each country. First, the Premier League:
In this article, we will focus on the average number of fouls rather than t stats. Since some of the derbies have been played different amounts of times, we get t stats that may not be better interpretations of the data. The reason why the Merseyside Derby had a lower t stat than Manchester United vs Liverpool and the Tyne-Wear Derby is because the standard deviation of the 24 Merseyside Derbies was much higher (7 vs 5), leading to a lower T Statistic (the higher the standard deviation, the lower the t statistic will be). A deeper dive could find that outliers drive up the average number of fouls and standard deviation in these games.
Before moving onto the other three leagues, it is important to note that all five rivalries are (almost) significant at the 5% level. This adds credence to our findings from above, and shows that it is not one rivalry that is skewing results but a nationwide phenomenon (as Brendan also found in 2015).
Here, we find that the Madrid derby is the most fiercely contested by a considerable amount. This could be due to Atletico Madrid’s style of play, but that is merely conjecture. All five rivalries featured significance at the 10% level, but only the Madrid, El Clasico and Basque derbies featured significance at the 5% level.
Here, we see that the two lesser known Serie A rivalries feature a lot more aggression than the other 3, particularly Genoa vs Sampdoria. It is also important to note that despite Serie A having a high T stat above (similar to La Liga), we find that it’s one rivalry whose matches are driving the extra aggression. In fact, the Derby D’Italia features fewer fouls than the Serie A average, and the Milan derby only marginally more.
Finally, we look at the Bundesliga:
These are the most interesting results. We find that in four out of the five rivalries, there are fewer fouls than in the average Bundesliga match. In fact, in both of Dortmund’s rivalries, we find that this result is significant at the 5% level. However, it is super interesting to note how much of an outlier the Hamburg vs Werder Bremen rivalry is. Despite the other four rivalries having an average that is below the Bundesliga average, this rivalry compensates for all four of them to have the overall average for the five be greater than the Bundesliga average (albeit not at the 5% level of significance).
Finally, we look at all 20 rivalries merged into one table ordered by foul ratio. The rationale behind choosing foul ratio as our metric is that T statistics are biased towards rivalries that had a full set of 24 matches, and foul difference negatively impacts the Premier League due to the smaller amount of fouls in Premier League matches. Our results are as follows
Our results are pretty similar to what we found above. The Premier League and La Liga rivalries settle towards the top, followed by the Serie A rivalries and then the four Bundesliga rivalries end up towards the bottom. The major exception to this is the Genoa and Hamburg vs Werder Bremen, which we noted above.
In conclusion, we confirmed Brendan’s work from 2015 using a larger dataset, Premier League rivalries do feature more aggression than the average Premier League match. We found that this effect is more pronounced in the Premier League than the other three major European leagues, but that this effect is significant at the 5% level in both La Liga and Serie A, but not the German Bundesliga. We then considered the specific rivalries on a closer level, and found one Italian and one German rivalry that are outliers from the other four within their country, and we also found that the rivalries involving Borussia Dortmund feature statistically significantly fewer fouls than the average Bundesliga match.
However, there are a couple of things to note. This study was also conducted with yellow cards instead of fouls, the results were similar but not identical (certain rivalries were ranked much higher or lower than in the fouls analysis). Also, the results of the study might have been different if other rivalries were selected (especially in Germany, where it was difficult to find a solid fifth rivalry that is regularly contested in the Bundesliga).
Finally, I would be remiss to not mention that the results of this study might be a result of different refereeing protocols in each of the four leagues as opposed to how the two teams play. It might be the case the Premier League referees try to get control of derby matches and as a result call more fouls for challenges that might not have been called fouls in other matches, whereas that might not be the case for rivalries in other countries.
Let us know your thoughts in the comments below.
If you have any questions for Andrew, please feel free to email him at firstname.lastname@example.org or reach out to him on Twitter @andrew_puopolo.