Here’s our fantasy premier league underlying stats review where we look at the xG and xA data and seek out a few players who under and over performed. We also look at what that might mean for their performance and pricing next season.
fantasy premier league underlying stats review – a few over performers and under performers
Background and methodology
So here I’m going to look at the xG and xA data from the excellent free site understat.com and identify a few notable over and under performers. It won’t be comprehensive and won’t be the biggest over and under performers but I’ll pick a few players of interest. I’ll then look at historically if that performance is consistent with their history or an aberration. Then I’ll take a quick view on pricing.
Just as background, in simple terms, xG or expected goals tries to calculate the chance/probability that a certain shot will turn into a goal using an average premier league player analysing thousands of shots as historical data. Similar for assists although it’s worth noting that FPL has a wider definition of assists counting things such as earning a penalty as an assist.
It’s worth noting that different organisations will have collected data slightly differently meaning different xG/xA numbers depending on where you go for them.
A few over performers
If you look at my FFGeek contributors review article you’ll see that Sadio Mane is in most contributors top 5 points per game list.
This season he scored 22 goals against an expected goals total of 16.74. That’s a 31% over performance.
If you look at his last 2 seasons at Liverpool overall he scored 23 goals against an expected goal total of 21 so he makes the case of an over performer. However that wasn’t to the same degree as this year as it is around 10% for those 2 seasons. Also it’s not that predicable as in 17/18 he under performed by 13% and in 16/17 he over performed by 37% .
If you include his Southampton seasons and include 18/19 he scored 66 goals against expected goals of 58.83. However it was 3 seasons of over performance vs 2 seasons of under performance. That makes it potentially a bit of a lottery for what you’re going to get.
In terms of pricing his Points Per 90 in 17/18 was 6.04 which gave him a GW1 price of £9.5 compared to a GW38 price of £10.3. Expect a price hike on his GW1 price with a PP90 of 6.76.
Is his performance of repeatable? His unpredictable xG performance attaches risk to that although you may believe that Klopp’s ability to improve players and the players adjustment to his system means that his returns have improved because of an improvement of ability rather than just unsustainable great finishing. Although, it’s worth noting that his xG and xA wasn’t vastly different than in 17/18 it was about the xG over performance and an additional 900 minutes played.
Interestingly he under performed his xA substantially with only De Bruyne having worse stats. His great finishing wasn’t replicated by his team mates. However this wasn’t the case the last 2 seasons at Liverpool so you’d think there would be a boost there.
As another aside I imagine (hard to trace who assisted who) one of the beneficiaries of Mane’s great finishing was Alexander Arnold who finished with 12 opta assists vs an expected assist total of 7.22. A big change from last season where he under performed his xA.
Son Heung Min
Son scored 12 goals against an expected goals total of 8.36. An over performance of 44%.
Son is a more consistent over performer than I would say Mane is in that his last 3 seasons with Spurs have seen an over performance although far less than this season. Overall for Spurs he’s scored 42 goals against an xG total of 31.33. An over performance of 34%. His xG 90 has been relatively consistent as has his xA over the 3 years although last season was definitely the best in terms of underlying stats he just didn’t over perform to this seasons ridiculous amount.
His PP90 this season and last is fairly similar in FPL so you’d think his £8.5m price might be similar depending on their view of gametime security. A decent pre-season as well could also make a difference.
Again one of the beneficiaries of his good finishing could have been Eriksen who significantly over performed his xA numbers. Some thing that’s been replicated to a lesser degree historically for Spurs although in a bit of a hit or miss fashion.
A few other notable over performers
Ayoze Perez scored 12 goals against an xG of 8.29. He has consistently over performed his xG in each of his 4 seasons although not to this extent
Zaha scored 10 goals against an xG of 6.49. In 3 of the 4 seasons before this season he under performed although they were fairly small margins of under performance.
James Ward Prowse scored 7 goals against an xG of 3.4. 4 of the 6 non penalty goals were from outside the box. He’s under performed his xG in 3 of the last 5 seasons although has improved each year and last season marginally over performed them.
Hazard over performed his xG as has every season with Chelsea. I’ve not focused on him as I expect he’ll leave for Real Madrid next season.
A few under performers
Here I’m going to avoid players who have been relegated, while they may come back through some other club it’s doesn’t seem to be enough of a reason to devote time to it.
Despite consistently posting good underlying stats which are also consistent with last season David Silva never made my FPL team. Obviously the Sterling, Aguero and Laporte threesome were the tempting combination but there was more to it.
David Silva for 4 out of the last 5 seasons has under performed his xG. The last 4 seasons have seen him score 21 goals against an xG of 26.21. You would think his xA would compensate but even that has seen him with 44 assists against a xA of 42.18.
At age 33 his minutes maybe even more limited and a drop his in PP90 from 6.25 in 17/18 to 5.37 in 18/19 could see a reasonable price drop. That’s if he’s still there after rumours of a summer departure have started.
Jesus is a very interesting case. Aguero is now 30 and the worry for FPL managers is that at some stage surely Jesus will eat into his minutes. That certainly wasn’t the case this season when his minutes in the Premier League fell from 1660 in 17/18 to 1017 in 18/19. Similarly his all competition minutes fell from 2594 minutes to 2253 minutes.
However his xG90 was an insane 1.14. For players who played over 900 minutes the distant second and third were Aguero and Aubameyang in the 70’s. His 2 previous seasons aren’t as good but they’re still pretty mind blowing. Now there’s no doubt that either coming on late as a sub or starting some of the less competitive fixtures to rest Aguero helps those stats but still.
However the flip side of this is his consistent under performance of xG. This season he scored 7 goals against an xG of 12.62. He also under performed in the last 2 seasons and what’s more the gap is widening rather than improving.
Factor in a decrease to normal size numbers if he becomes a starter at some stage and that consistent under performance becomes a real issue. It’s hard to guess his price next season. His actual returns were similar to the previous season so it maybe whether gametime is factored into the price difference.
It’s interesting that De Bruyne despite posting xA90 stats way above anything he’s done previously just had no luck with people finishing them off. For the first time in his Man City career he massively under performed his xA stats. He’s never had much of a goal threat but his ability to hit screamers from distance tends to mean he has seasons that he over performs those numbers. This season wasn’t one of them. Despite an excellent underlying stats season his PP90 dropped from 6.12 to 5.49. If his new season price drops from the GW1 price of £10.0m he could be worth considering as a City option.
Ings posted excellent xG90 numbers this season. In fact in xG90 he was 8th with Callum Wilson the only striker within anything like his price. The only problem with Ings (apart from being constantly injured) is that he just can’t finish. In his 4 seasons in the Premier League he’s under performed his xG by a reasonable margin and the season he bettered it, it was only by a whisker.
His price at £5.4m doesn’t have much scope to go lower. A PP90 of 4.91 is still ok but whether you believe his natural level is 8th in the League for xG90 is a big question.
A couple of others
Benteke – 3 seasons at Palace 3 seasons of massive under performance of his xG
Harvey Barnes – hard to gauge much off 1 season but 1 goal vs an xG of 3.62 isn’t great especially when your xA 90 numbers are terrible.
Digne – 4 assists against an xA of 7.71 is disappointing but surely couldn’t be as bad next season. The reverse though is his ability to score from outside the box surely can’t be replicated with 4 goals from an xG of 1.2.
That’s it hope you found it interesting