Shooting Your Shot

Sign Posts: Gameweek 11

We’re still too early to place too much stock in underlying data, especially at the individual level, but in FPL we know we can’t wait until data is totally stable to look at it, because then the season is probably over! One factor that potentially makes this premature peek at the underlying data more palatable is comparing the per 90 rates for both total shots (which I think stabilize a little quicker than the noisier xG numbers) and the underlying numbers compared to prior season. Thus, we’re not looking for signs that Yoane Wissa is now a 0.7 npxG/90 player, but rather seeing his increased production in context to prior season to then sense check whether such an increase/decrease also aligns with our other priors (such as playing on a new team, with a new manager, getting more playing time).

There’s a lot going on in this chart, but we’ll go through some of the names to tease out some potential nuggets of interesting data below. The horizontal axis is the increase / (decrease) in total shots per 90 minutes for the current season compared to last, so the players in the right quadrants are taking more shots this season. The vertical axis is the increase / (decrease) in non-penalty expected goals (npxG) per 90 minutes for the current season compared to last, so the players in the top quadrants have accumulated more npxG per 90 so far this season.

Fewer, Lower Quality Shots (bottom left quadrant)

There’s a fairly large group of very good players who are currently getting not only fewer shots this season, but also lower quality efforts. Several of these names make intuitive sense as the likes of Marcus Rashford, Jarrod Bowen, Phil Foden, and Anthony Gordon are all currently performing at a level below last season (in Foden’s case this is because he was great last season but hasn’t yet found that form, whereas Rashford has sunk even lower from an already mediocre 2023-24 campaign). Chris Wood is an interesting name here, as this certainly runs counter to the narrative that the Kiwi is playing at a notably higher level this season. In reality, the Forrest man is posting a lower npxG/90 (0.38 vs 0.59) and registering fewer shots (2.02 and 2.38) compared to last season, all of which supports the notion that he is overperforming to date (7 non-penalty goals from just 3.7 npxG). Dominic Calvert-Lewin owners - and possibly Everton fans alike - won’t be overly surprised to see him appear here, and while his 0.37 npxG/90 isn’t a disaster (recall that “free-scoring” Wood is posting 0.38), his historic underperformance against his xG suggests a haircut against these moderate numbers is needed.

We also see Bryan Mbeumo in the bottom left quadrant, and it was the Brentford man’s numbers which first sparked a closer look at this data. Mbeumo is well fancied by both statistically minded and more general managers alike (99.7% Elite Ownership at FPLReview and 97.8% among LiveFPL’s Elite group) thanks to his solid historic underlying data and strong results this season (8 goals and an assist in 11 starts). However, we actually observe a decline in his nxpG from 0.34 to 0.21 and we see this is a product of fewer overall shots, which have declined from 2.30 to a troubling 1.83. Mbeumo has never been a prolific shooter (2.35, 2.17, and 2.30 shots per 90 before this season), and his ability to get high quality shots on counter attacks has been a product of his success. Yet, 1.83 shots per 90 is just really low - as in, less than Trent Alexander-Arnold and Miguel Almiron posted last season - and building consistent FPL returns on the back of those shot numbers while also posting good-not-great creation numbers (0.19 xAG) is a genuine concern. Brentford are about to walk into a tough fixture run from GW14 on, and Mbeumo will be a tough sell, with many seeing him a reliable long-term hold. However, these warning signs suggest there may well be scope to move on without being punished too severely down the road.

Underperformance doesn’t of course mean poor performance, and so we also see Mo Salah here who by any metric is having a very effective season and any slight decline highlighted here doesn’t seem like a cause for concern. More on Liverpool in a coming post.

Fewer, Higher Quality Shots (top left quadrant)

If you are looking for players who might be being overvalued by casual users of expected data they’d probably fall in this quadrant: high or at least improved npxG numbers but without the support of shot data which suggests they are well placed to continue (i.e. the xG is potentially coming from one or two very good chances, which are great, but perhaps subject to spiky production over a longer period).

Two names to look at here are Ollie Watkins and Dominic Solanke. Solanke’s nxpG/90 number looks great, rising to 0.70, up from last season’s very good 0.61, which supported a 19-goal season which won him that £55 million move to Spurs. However, his shot totals have actually ticked down to 2.7 - which is fine - but not necessarily reflective of a player making a step up in an offense that should on paper be better than Solanke’s 2023-24 Bournemouth. As we previously discussed, Solanke’s expected goals number have been really spiky, due to a handful of very good chances that have fallen his way this season. 0.23 npxg per shot this season is really high, and if sustained for a season would rank ahead of everyone from last season who played a high level of minutes other than Nicolas Jackson (who we recall had his own wacky xG numbers thanks to that ~3.0 xG effort against nine-man Spurs). On the other hand, if we (perhaps inadvisably) chop up the data further, we could note that in Premier League games with Son Heung-min Solanke has registered 3.2 shots per 90 versus 1.7 per 90 in those without the Korean star, so we might look back and conclude there was nothing to worry about. If nothing else, last time I wrote about my concern with Solanke’s prospects he registered 5 shots, 1.5 xG, and 2 goals against Vila, so congratulations to his owners on this week’s haul.

Watkins’ numbers are in a similar place to Solanke, but narratively we’re in a different place. Here, I think it’s well understood that Watkins overachieved last season, well exceeding his underlying data (19 goals from 16.8 npxG) despite historically being a slightly below average finisher (40 goals from 42.6 xG in 2020-2023). Villa’s total shots have ticked down this season from 13.4 to 12.4, similar in ratio to Watkins’ decline from 3.0 shots to 2.8 per 90. I haven’t looked at the data for this, but one hypothesis could be that Watkins was able to add late shots last season with Villa playing so many games with a lead, whereas now those minutes (and chances) are going to Jhon Duran. It’s hard to really conjure a narrative as to why Watkins’ shots would improve from an average of 0.16 npxG per shot to 0.20 and if we regress those back to the mean and the shots don’t also tick back up in number, then Watkins becomes a ~0.45 npxG/90 man who doesn’t offer up too much assist potential, doesn’t take penalties, rarely play 85+ minutes anymore, and costs £9.0m. In other words, not a great FPL asset despite his perceived consistency.

The other name jumping out here is Yoane Wissa, who hasn’t really lost any shot volume but has dramatically increased him expected goal numbers. Wissa’s npxG/shot has almost doubled from 0.16 to 0.30 which I suspect is why the models aren’t convinced by the Brentford man’s future prospects. This is where this additional slice of data can be useful to tease out a bit of context around an inflated xG number that can otherwise be tricky with these small sample sizes. Wissa is a good player and Brentford do seem to be favoring a more adventurous style this season, but it’s hard to see Wissa continuing to get so many really high quality chances, which will cut into his FPL production down the road unless he finds a way to get more shots to compensate.

More, Higher Quality Shots (top right quadrant)

This is where we’d expect to see players who are already attracting FPL attention as their xG numbers pop and they drift towards the top of the data tables. This is the case with players like Antoine Semenyo, Danny Welbeck, Brennan Johnson, and Raul Jimenez who are posting excellent shot numbers to go with their FPL returns. A few other names here deserve a quick mention. First Erling Haaland, has incredibly increased both his total shots per 90 (3.99 to 4.91) and npxG per 90 (0.81 to 0.93) so far this season. With all the recent talk of whether to sell the Norwegian - who casually notched his 25th career hattrick while on international duty - it’s sometimes easy to forget that he remains an absolutely insane goalscorer.

One name here who didn’t immediately feel intuitive was Mohammed Kudus, whose West Ham side have often felt impotent and a pretty miserable watch this season. First, it’s interesting to note that West Ham’s xG total (15.1) isn’t much worse than Aston Villa (16.2), Newcastle (16.3), and Brighton (16.4), all of whom are fancied as solid attacking teams and boast at least a couple of potential FPL assets each. Second, it’s worth looking at just how many shots Kudus is taking:

The names at the top of the list here aren’t typically the most clinical bunch and wouldn’t be confused with someone like Mo Salah who is ostensibly a winger but gets quality shots (0.16 npxG/shot) like a forward. However, Kudus has averaged 0.10 npxG/shot which isn’t bad and is driving that uptick in xG90 we’re seeing in the top chart. This is particularly notable given that West Ham seem to have appointed a manager whose system doesn’t really seem to favor Kudus or Bowen (arguably West Ham’s best two players). Unfortunately, Kudus will be suspended during the next two difficult games (NEW, ARS) which would have been a good opportunity to see him in action before the favourable fixtures start in GW14 (LEI, WOL, BOU, BHA, SOU). The Mbeumo to Kudus swtitch in GW14 would have been perfect timing but without seeing improvements beforehand that does seem like it’s a bit of a risk given the plethora of other midfield in that price range. The talented Ghanaian is one to keep in the back of our minds though.

More, Lower Quality Shots (bottom right quadrant)

This was the quadrant I was hoping to find some juicier names in, as potentially these could be players who are getting chances, taking their shots, but perhaps not quite getting great shots, which could improve over time. While the two highlighted names aren’t exactly under the radar options, Bukayo Saka and Alexander Isak do represent something of a potential opportunity as we head into GW12. Saka is widely considered to be a “template” player but I don’t think many realise how much he’s been sold recently, dropping to 20% ownership in FPLReview’s Elite 1000 and just 15% in LiveFPL’s elite group. Saka has been good this season but seems to have found himself as the odd-man-out in the midfield template which seems to now favour Salah, Palmer, Mbeumo, and then Fernandes or even Son as that extra option (plus those still holding Haaland who presumably can’t easily then fit Saka in).

If the Arsenal man can maintain his current 3.4 shots per 90 level, it would mark another step forward for him on his way to becoming a truly world class superstar, building on the growth of the past two seasons (2.43 in 2022-23 and 3.14 in 2023-24). The reason his FPL points haven’t necessarily stepped up with those shots, if of course the decline in quality of the efforts on goal, with his npxG/90 dropping to a career low 0.08 compared to 0.11 in both of the past seasons. It’s hard to say whether that is just noise and something that could regress back to average, or whether the only reason the shots are up are because, for example, Saka has been forced into taking worse efforts with Ødegaard missing, or if he’s been forcing his hand as Arsenal have enjoyed less time controlling games this season due to some unfortunate breaks on the defensive side. Without a change in manager or really any significant personnel changes (except via injury), it seems unlikely Arsenal would intentionally be setting up the team to have Saka take worse shots, and this isn’t obviously akin to a situation we’ve seen with someone like Salah where the shot volume gets worse because a player is asked to hold their wide position for the good a team’s attacking structure. Saka seems to be playing a relatively similar role as before, albeit slightly less involved across the board with his possession numbers. Arsenal’s fixture run seems impossibly good for the foreseeable future, having already played difficult away trips to Man City, Chelsea, Newcastle, Spurs, and Villa. We also know that Saka is likely to get a lot of minutes despite a crowded holiday calendar and so there’s a chance he’s a little undervalued right now, albeit tough to fit into most teams.

Isak is in a relatively similar position as Saka, as while he’s recognized as performing well enough, he’s struggled to break into the FPL template past the preference for the budget forwards, Haaland, or Solanke. However, a couple of points can pique one’s interest here. First, we see Isak’s shot volume is up from 2.87 to 3.06 per 90 which compares favorably to any other forward in the league who are likely to get high quality chances other than Haaland. Second, we should note that Newcastle have played an under-the-radar difficult schedule so far. Indeed, in Michael Caley’s recent rankings at Expected Goals, Newcastle had the hardest rated schedule so far this season, beyond even the much discussed brutal start Wolves have faced. The upcoming fixtures switch this week with the visit of West Ham, but stay good through to GW24 so Isak will be there to jump on as the opportunity arises, even if not the number one priority for this week if he doesn’t work for your team structure.