Revisiting How to Use xG and Alienate People (/Your Competitors)

Back in early October, after the opening stanza of the 2021-22 Premier League poem, we saw how early season overperforming (as indicated by expected numbers) has often, historically, been related to considerable decreases in fantasy points per 90 (PP90) for the remainder of the season. Based on this, we then identified some candidates to be wary of during the current season. Now, 22 gameweeks later, we’re going to revisit those names and see whether our hypotheses have borne fruit.


The Names: Overperforming

Twelve players were overperforming their XG+XA by at least 1.5 goals after seven games of the season this year. The midfield powerhouses Paul Pogba and Abdoulaye Doucoure were the biggest culprits, doing so by 3.94 and 3.52 goals, respectively. Goal poachers Jamie Vardy and Danny Ings were next up on the list, at 2.68 and 2.65. Then we had an array of talent ranging from Mohamed Salah (1.88) to Javier Manquillo (1.72), zig-zagging to some of fantasy’s biggest names in the process.


The Names: Underperforming

Fifteen players made up the underperforming group, with four fantasy forwards leading the way in Romelu Lukaku, Harry Kane, and the then-Wolves pair Trincao and Adama Traore. Not quite frustrating managers to the same extent were another quartet; Diogo Jota, Mason Mount, Nicolas Pepe, and Bryan Mbeumo, who were all underperforming by between 1.92 and 2.05 goals. Then finally, a mixed bag of names also satisfied our minimum 1.5 goal underperformance criteria, most notably Liverpool’s Sadio Mane (1.63) and Manchester City’s Ilkay Gundogan (1.64).


The Analysis

As before, we are going to use PP90 as our fantasy metric of choice. In order to examine whether our semi-predictions came true we are going to calculate the change in PP90 from gameweeks 1-7 to gameweeks 8-29. Positive numbers reflect an improvement in fantasy performance (which we should expect from the underperformers) whilst negative numbers reflect a decline in fantasy performance (which we should expect from the overperformers). We are also going to calculate the gameweek 8-29 PP90 as a percentage of the gameweek 1-7 PP90, in order to allow for fairer comparison between players. Here, numbers above 100 would indicate increases in fantasy production and numbers below 100 would indicate decreases in fantasy production.


The Results: Overperformers

The above figure shows the gameweek 8-29 PP90 as a percentage of the gameweek 1-7 PP90 for each of the overperforming players. For instance, Mohamed Salah has seen his PP90 go from 23.29 to 18.91, this latter number being 81% of the former.


Success! Well, success for the hypothesis that is, not the players (or fantasy managers) concerned. Ten of the twelve overperforming players have seen their PP90 decrease, with only Tottenham’s Heung-Min Son and Aston Villa’s Danny Ings managing to increase theirs (and even then, not by much). Notably, nine of the players have seen a considerable decline in fantasy performance – that is, their gameweek 8-29 PP90 is less than 85% of what their gameweek 1-7 PP90 was. Our two biggest overperformers – Pogba and Doucoure – have seen huge declines (now performing at 50% and 45%, respectively), but the autumn superstar that was Andros Townsend takes the gold medal, as his gameweek 8-29 PP90 of 8.08 is just 44% of the 18.33 that he was sporting early on in the season.


The Results: Underperformers

As before, the above figure shows the gameweek 8-29 PP90 as a percentage of the gameweek 1-7 PP90 for each of the underperforming players. For instance, Sadio Mane has seen his PP90 go from 13.21 to 13.28, this latter number being 101% of the former.


Whilst early season overperformance seems to be associated with declines in fantasy performance later in the season, the historical data for underperformance was far more ambiguous, and so it appears again for the current season. Harry Kane’s improvement has been astronomical (239% of his early season production!), whilst Mason Mount, Ilkay Gundogan, and (for whatever it is worth) Nicolas Pepe have also been kinder to their fantasy managers since mid-October. But for every increase, there’s almost an equal decrease. Curtis Jones, Trincao, and the now-departed Adama Traore may not be of too much relevance, but Romelu Lukaku, Dwight McNeil, and Bryan Mbeumo have not only failed to halt their underperforming fantasy start, they’ve also somehow got worse.


Just tell me what it means!

Well, not a lot now. With 29 games gone, fluctuations in fortune (which, to some extent, is what XG and XA reflect) have likely levelled off*. But next season, if in October and November, you’re not using expected numbers to project fantasy value, then you’re putting yourself behind the 8-ball in a major way.


The relationship between early season overperformance and later season changes in PP90 seems to be a robust (and logical) one, so having the conviction to jump off that soon-to-be slowing bandwagon should be a priority for fantasy managers aiming for glory. Obviously knowing who to target in such a trade will be critical, and unfortunately, it’s not as easy as simply looking for those underperforming players – we’ve seen that the relationship with changes in PP90 here is very much mixed. But given that 10 of the 12 biggest overperformers this season (and 19 of the 20 last season!) saw their PP90 decline, even if you can simply trade out a player for equivalent (at the time) value, then down the line you should be at a net advantage.



*That said, it may be worth noting that Kevin De Bruyne (5.05), Emile Smith Rowe (5.02) Jamie Vardy (3.46), and Maxwell Cornet (3.43) are all considerably overperforming in terms of XG, whilst Paul Pogba (5.16), Alexandre Lacazette (4.20), and Andrew Robertson (4.09) are all considerably overperforming in terms of XA.




For all the latest from The Inner Geek, follow @the_innergeek on Twitter!

And for more in-depth and exclusive resources, become a member of The Inner Circle.








234 views0 comments