Leicester City started the season in sensational form. Their win against Aston Villa on December 8th was their eight in a row and gave them 38 points through 16 matches. They’ve leveled off since, registering 16 points in next 14 matches. So who are the real Leicester City?
Astute observers may have seen that Leicester were significantly overperforming on their underlying numbers during those first 16 games. So, is this just standard regression to the mean, or have Leicester been materially worse? Well, yes.
Leicester’s expected goals, a metric that estimates chance quality using historical data, through the first 16 were good with an expected goal difference of +13, according to fbref.com. Looking at their actual number and they were unbelievable with a goal difference of +29. Looking at that enormous discrepancy, the Foxes were unlikely to keep that level of performance up. What they were doing appeared to be pretty unsustainable.
Looking at their 14 matches since and there has been a mix of regression to the mean and regression in performances. Their expected goals and actual goals have pretty much matched up in that time, with a +1 goal difference and a -1.5 expected goal difference, according to fbref.com. Their performances have also been worse, with their expected goal difference per game dropping from +0.84 in the first 16 to -0.11 in the last 14, according to fbref.com.
Now, time to apply some context to those samples. The first 16 game sample likely skews positive due to some outlier results, specifically the 9-0 win over Southampton and the 4-1 win against Aston Villa. Conversely, it can be argued the next 14 skews slightly negatively due to both matches against Manchester City falling in this sample.
So which Leicester is the real one, the first 16 or the last 14? It’s probably somewhere between those two extremes. Their total expected goal difference comes out to 0.39, which is essentially the same as it was in their 10 games under Brendan Rodgers last season (+0.4, according to fbref.com).
Through 40 games of the Rodgers era, his expected goal difference per match is +0.39. Extrapolate that across a full season and it comes out to a +15 expected goal difference. For context, that would have made them the fourth-best team last season and the sixth-best tram the season before, according to fbref.com.
Considering we have a 40 game sample, it’s reasonable to assume the “real Leicester City” is something in the vicinity of that +15 goal difference team. A high Europa League, borderline Champions League team. Considering where Leicester are on the football hierarchy, that is extremely impressive. But, it is crucial to calibrate the expectations somewhere closer to that than the earth-shattering pace they were on for the first 16 matches.