A couple of days ago Olympique Lyon chairman Jean-Michel Aulas announced that the club has filed two lawsuits against the LFP. The two lawsuits are a consequence of the decision by the LFP to cancel the 2019-2020 season in the Ligue 1, and base the final standings on an average of points earned per matches played. This resulted in Lyon missing out on qualification for European football, which is not well received by the chairman of Lyon.
► Uneven difficulty of clubs’ match schedules
Part of the reasoning behind the lawsuit is given by the following statement by Aulas: “When we look at the classification method used: Nice played 32 matches this year (OL: 44), played 15 times at home in the league (OL: 13 times) and had yet to play PSG, Marseille and Lille again. Nice end up 5th when it was 6th after the 28th matchday.”
The point made by Jean-Michel Aulas is very clear. Is it fair to base the final league standings on a points-per-game average, while the strength of the clubs’ schedules might differ significantly? The fact that there is only a one-point difference between Lyon and Nice provides even more reason to question ignoring the strength of the clubs’ schedules.
There is no doubt that the LFP had to use some method to rank the teams, and no decision would satisfy every stakeholder. Moreover, it appears that the LFP did try to account for the unfinished schedules to a certain extent by using head-to-head results to rank teams in case of a tied average. Although this might be better than the usual decider (goal difference) in case of teams ending with an equal (average) number of points, we would still completely ignore the other 26 games that the clubs have played. Ignoring almost 93% of all matches does not appear to be the fairest solution.
► The need for a more statistical approach
This difficult situation asks for a more statistical approach. Companies like Squawka and Opta have already developed statistical models to simulate the ending of the current Premier League season. It is possible to build a similar model for predicting the final league table of the Ligue 1 as well, but there is one issue with these statistical models: the clubs’ strength levels are based on historical results going up to four years back in time. From a statistical perspective it is certainly true that such models have great predictive value, but it is not fair to incorporate data from previous seasons when deciding upon a final league table in the current season.
What we need is a statistical method that accounts for difference in league schedules, but only uses data from the current 2019-2020 Ligue 1 season. One of the methods that is specifically designed for situations like this is Massey’s method, a rating system that is introduced by Kenneth Massey. The purpose of the Massey ratings was already described by Kenneth Massey himself back in 1997, and indeed fits the current situation in the Ligue 1 very well:
“In the course of developing a rating system, it is necessary to determine exactly what the ratings are intended to accomplish. Should ratings reward teams that deserve credit for their consistent season long performance, or should they determine which teams are actually stronger in the sense that they would be more likely to win some future matchup? The first objective is usually emphasized if the ratings are meant to determine a champion or those teams that should qualify for the playoffs. In this case, results should attempt to explain past performances. Popular interest in these ratings peaks when there is intense disagreement that cannot be resolved on the field.”
► Towards fair Ligue 1 league tables
Massey ratings are often based on goal differences, which results in relatively high ratings for teams that often win in by a large margin. In fact, this would rate Lyon as the second best team in the Ligue 1. However, given the current circumstances it does not seem correct to downgrade clubs that often win with a smaller margin. For that reason the Massey ratings in this article are based on points won (i.e. 3 points for a win, 1 for a draw, 0 for a loss). Now if we apply this statistical method to the 2019-2020 season in the Ligue 1, we get the following ratings.
*In case of equal ‘Fair points per game’ between two or more clubs we use the same method as used by the LFP:
1) Head-to-head matches; 2) Goal difference
From the above image we can immediately see that Aulas’ gut feelings are right. Nice has indeed benefited from a relatively easy schedule. Although the LFP placed them in 5th position, their actual ranking should be the 9th position. This would result in Lyon climbing one spot to the 6th position, which will probably be sufficient to qualify for European football next season.
It must be said that there still is a chance that the 6th position will not give access to the Europa League qualifiers, since we still have to wait and see whether the final of the Coupe de France and the Coupe de la Ligue (in which Lyon plays Paris SG) will be cancelled. We therefore cannot yet exactly estimate to what extend Lyon might feel disadvantaged by the decisions of the LFP.
Either way, the Massey ratings have shown that Lyon has a good point stating that Nice has benefited from a relatively easy schedule. It is very realistic that this decision by the LFP will even cost Olympique Lyon tickets for European football next season. Moreover, the statistical method proposed in this article would be a better solution for ranking the teams in case of a cancelled competition. The method used by the LFP does not account for strength of schedule, while Massey’s method is specifically designed for deciding unfinished competitions.
If we take a look at the rest of the league table, we can see that the first four teams in the Ligue 1 are ranked correctly. The same holds for almost every other team, including the two teams that have relegated from the Ligue 1 this season. The decision to relegate Amiens and Toulouse is a completely different discussion, but the fair table at least shows that Amiens and Toulouse are the worst teams in this season’s Ligue 1. The only clubs that actually feel the negative consequences of a relatively tough league schedule are Reims and Lyon, because they might not play in Europe next season due to Nice currently ranking 5th.
In the next couple of months we will certainly hear more about how the conflict between Lyon and the LFP will develop. It is also interesting to follow the affairs in other competitions. The Dutch Eredivisie has already cancelled the competition. Scotland and Belgium are likely to follow. And what will happen in England, or in Spain? Will they cancel the season as well, and if so, how will they construct the final rankings?