An index of strength of rival’s faced in a league

The season is slowly – the international break made it a bit slower – progressing and we try to make some sense out of the various tables of the top European leagues. As tends to happen at the beginning of the season, we see some under- and overperformers in the competitions. At least a different performance to what our expectations were towards several teams. An often-heard and other hand sometimes forgotten flaw of the current league tables is that, with 11-13 games in, only just over half of the first competition round has been played. This obviously means that some teams have faced other rivals than their competitors. And more important, although some teams might share their league rank with another team, they might have faced tougher competition than the other team. In this post I introduce (or reapply as it has been used since 1882..) a way to distinguish the teams that gathered a high amount of points (partly) because they just faced weaker opponents so far.

Some examples of teams that are doing better than generally expected are Southampton, West Ham United and Swansea (2nd, 4th and 5th in the Premier League), Genoa (6th in the Serie A), Valencia (3rd in La Liga, which is not a very surprising position, but the fact that they stay close to Real and Barca is), Hannover 96 (4th in the Bundesliga), PEC Zwolle (5th in the Eredivisie) and Marseille (1st in the Ligue 1, again, no surprising position, but their high amount of points is). On the other hand we have some clear underachievers like Dortmund, Werder Bremen, Liverpool, Tottenham, Heracles, Athletic Club de Bilbao, Fiorentina, Lille and AS Monaco.

In some chess tournaments, where in the Swiss system you might only play only 9 to 11 games although the table contains over 40 competitors, there is a nice way to decide which player is ranked above the other in case of an even in total points. As it is hard to score goals and decide a tournament on goal difference in chess, the general way to do this, is via the ‘cumulative opponent’s score’; the sum of the points of all rivals that a certain team faced. It is a nice way to see which player, in the case they faced different opponents, faced more difficult competition. Moreover it is an easy and accurate way to solve the previously stated problem in our football league and I am surprised I have never seen this before.

I gathered data from 6 leagues: the English Premier League, La Liga, the Bundesliga, the Serie A, the Ligue 1 and the Eredivisie and made a table based on the previous ‘cumulative opponent’s score’-system. To be able to compare across the 6 leagues I than divided the sum by amount of opponents and games to come out at an average amount of points of the rivals faced (Rival’s average points per game or RPpG in the following tables):

RPpg table

In the table on the left we see the teams that so far have faced the toughest opponents. And here we recognize some teams that can be seen as teams that we would have expected to score more points so far. Best examples are Athletic Club de Bilbao (1st), Real Sociedad (2nd), Fiorentina (4th) and Heracles (5th).

On the other side you can see Real Madrid have relatively faced the ‘weakest’ opponents of all teams in the 6 included leagues (their opponents on average win 1.14 points per game). Inter Milan has the joint lowest RPpG (1.14) and makes the fact that they are not doing so well (9th in Serie A) even more worrying. Also in the list of teams with ‘weakest’ opponents we see league leaders Chelsea (1.19) and Juventus (1.19).

Some other teams of interest:

RPpG random

(Note: the most accurate system would be to take out the game the rival played against the team that is considered – considering the RPpG a rival has against all other teams – but for explanatory reasons and calculation ease I decided to leave it like this. Moreover, the tables turned out to be only slight different, but the top- and bottom-10 remained the same for 90%.)

To be complete you can find the league averages and standard deviations of the RPpG:

Average strength of rival per team

The very high standard deviation in Spain explains why there are so many (11) Spanish teams in the top- and bottom-11 of the RPpG.

While doing this I decided to add a lot more data to the table, which I will discuss in another post, but one of those fits in this post. It is the difference between the average WhoScored rating of the rivals faced by a team (Rival WS in table) and the league average. This gives the following first- and last 10:

WhoScored rating vs RPpG

Note: obviously when a round is completed and all teams have faced the same rivals, this point system is useless, as all teams will sum the same amount of points. In this case another system used in chess to decide who outperforms another player with the same total of points: the Sonneborn-Berger system. It is the same as the ‘cumulative opponent’s score’-system, but in this case a team only sums points if it won against an opponent.

(Note: the WhoScored rating of rival´s faced and the RPpG have a correlation of .53)

Here are the 6 leagues that were discussed in this post:

Leagues england span RPpG

Leagues germany netherlands RPpG

Leagues italy france RPpG

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