The Argument for Phil Foden

 

Introduction

I've seen a lot of debate on twitter recently on whether Foden is a good FPL asset or not, or more precisely, whether his teammate KDB, who is the same price, is a better option. I have to credit @FPL_figures for the original tweet which caught my eye, but since this tweet I have seen others. In this article I'd like to make the case for Phil Foden, as I don't think he's as bad as the data that people are currently quoting suggests.

Underlying Numbers

If we take a look at last season, it's clear to see that Foden and KDB are very different players, and also very different in how they are likely to deliver FPL points. Looking at the below graph, we can see that KDB is far more likely than Foden to provide an assist, both in real terms and in xA terms, while Foden is much more likely to score (again in real terms and in xG). Their expected points/90 are virtually the same, with KDB slightly out ahead, although it must be said that his data is from a much smaller sample size than Foden's (1233 minutes v 2895 minutes), meaning that it is far less reliable.

Now let's take a look at the season before... (22/23)

Foden is once again more likely to score, with KDB more likely to assist, and their expected points/90 is virtually the same (although KDB played more than Foden this season: 2448 minutes v 1837). What about the season before that?

This is the last season where KDB was ahead of Foden on goals/90, and he was in fact still behind him on xG/90. Their expected points though... still the same, mainly because KDB scored 9 goals over his expected goals, something which he has been nowhere even remotely close to repeating since then.

What does this all mean?

What some FPL creators are trying to suggest, is that, because Foden's xGI was substantially lower than KDB's last season, he will score fewer points in the upcoming season, so it's a no brainer to pick KDB ahead of him. There are a number of problems with this suggestion however.

    Firstly, this fails to acknowledge the difference in points awarded when a player scores versus when a player assists. For midfielders, a goal is worth 5 points, whereas an assist is worth 3 points. So a stat like xGI, is very misleading. For example, if Foden's xGI of 0.58/90 was 0.58xG and 0.0xA, whereas KDB's xGI of 1.03/90 was 1.03xA/90 and 0.0xG, Foden's expected points per 90 just for goal involvements would be 2.9, whereas KDB's would be 3.09. In addition, goals for midfielders are worth 18 in the bonus points system, double the 9 that assists are worth, meaning that Foden is also more likely to be getting bonus points (10.44BPS/90 on xGI for Foden v 9.27BPS/90 on xGI for KDB). Of course, these are extreme examples, and are not the reality, but the first point is that xGI is not a great stat to use when comparing midfielders, particularly in this case where they offer completely different things. In his short career up to this point, Foden has scored 9 or more goals in the league in 4 separate seasons, 4/5 of the seasons in his career in which he has played 400 minutes or more. KDB on the other hand has achieved this total only twice in the last 10 years - they are different players.

    Secondly, as mentioned previously, KDB's minutes were limited last season due to injury, meaning that his data is less reliable. Across his career, last season was the first in his career where he was achieving an xA anywhere close to 0.87/90 (previous career best was 0.66), and the first time in his career where he has broken 1xGI/90. Maybe he just had a great season, or more likely, his limited game time coincided with a purple patch and/or he was able to exploit tired defenses when coming off the bench. When analysts look at expected data they tend to look at players who have played over 2000 minutes; KDB played almost half of that total - for me these numbers are not reliable, especially when it's the first time he has produced anything close to this in his entire career.

    Thirdly, this suggestion ignores the fact that Foden is a great finisher, over performing his xG by more than 2 in 3 out of the last 4 Prem seasons. For sure, he out performed his xG massively last season, and an over performance by almost 8 goals is not likely to be repeated. But, an over performance is not unlikely, and he has demonstrated that he is a great finisher, so just because his xG is 'only' 0.35/90, doesn't mean that he'll 'only' be scoring at a rate of 0.35/90. 

    Finally, KDB is a year older. FPL players cannot expect him to play 38 games; he will be rotated and he will get injured. Foden is much less injury prone (but of course is prone to rotation), and has been getting better season on season as he gets older - the goal and assist numbers back this up. KDB, in my opinion, is on a slow decline, and as much as Pep loves him, he's 33, he won't be able to continue at his exceptional level forever. Another icon of the Pep Guardiola era, David Silva, left City at 34, and with players like Alvarez, Bernardo Silva and Foden all able to play KDB's position, it won't be long before he plays less and less, regardless of whether his expected data is still good or not. My point is, Foden is on an upwards trajectory, whereas KDB is going in the other direction.

Conclusion

Look, I can see where people are coming from with the suggestion that KDB is a better option than Foden, and anything can happen, maybe he'll haul in GW1 and Foden will be benched, maybe he'll continue at the rate of his expected data from last season - but it's unlikely. Foden is a rare player who consistently out performs his xG, and this should not be ignored, he's a great finisher. Furthermore, he played more games than he ever has before last season, and his expected points are still pretty much the same as KDB's, despite KDB's exceptional underlying stats (which I am skeptical he will be able to replicate this season). I just don't think that the decision is as clear cut as some are suggesting, and for me Foden is a better option. I think, over the course of the season, he'll play more games than KDB, score more goals, have a higher xG/90 and get more points. That's my prediction :).

Dylan

@FPL_figures posts some really interesting stats on his twitter, so please do go and check him out. No hate intended, no offence intended, I just disagree with the general idea that KDB is clearly a better option, and I think the raw data can sometimes be misleading. Thanks!