By Alfie Cairns Culshaw (Chief Editor)
(All data taken from FBREF.com)
"Alfie is the analytical, statsy nerd guy on the pod, whereas myself and Bert just run on pure emotion," said Daniel Finton, on the 178th instalment of the Arsenal Cannon Podcast. I may well be somewhat paraphrasing, but I'm too lazy to find the exact quote within a 90 minute audio extravaganza.
While I refute Dani's dismissal of me as some cold-hearted, emotionless and completely sterile analyst (as those who watch football with me can attest), I do indeed love the numbers, and might as well play into this narrative within this piece. Given we are at the first interlul of the season, it feels very appropriate to have a generic metric based article, and this feels like the perfect topic of discussion to break my writing hiatus.
Of course, there are plenty of people out there who reject any use of data to analyse football and stress that the eye test is paramount. Personally, I feel like the data can contribute to the narrative, not define it, and can reveal interesting snippets of information that you may not have picked up with the eye test.
I think the general consensus amongst Arsenal fans with the eye test thus far this season is that we've been very good in the first two thirds of the pitch with our build-up, have pressed well, suffocated teams and limited our opponents to few clear-cut chances, but have perhaps not started firing in the final third as of yet. The automatisms in our front five seem slightly off, which has meant that we've not created the quality and quantity of chances we'd have liked. In a nutshell, the numbers support all this perfectly.
I must caveat this entire article with the fact that we are indeed operating with a very SMALL SAMPLE SIZE, so definitive judgements on what we are and could be this season have to be taken with a pinch of salt. Having said that, taking last season into account we can begin to look at trends. Although we've undergone a few personnel and systematic changes, we are still a largely similar side that are not entirely in a transition of identity (like a Spurs or Chelsea), so we can expect to be at a similar level to last campaign.
As aforementioned regarding the eye test and our ability to suppress the opposition's threat, the data suggests we do indeed have one of the two best defences in the league at this point, alongside Manchester City. While the discourse has been verging on doom and gloom on the offensive side, we still possess the fourth best attack in the division with regards to xG created. We've already been awarded two penalties in our first four games, which certainly distorts these numbers slightly, but nevertheless, it's a good sign that it doesn't feel all that free flowing in the final third but the numbers say it's still a top four attack. Imagine how good we could be when it does still clicking in the final third.
This also demonstrates how expectations have risen. It does feel as though if we'd produced these exact four performances a year or so ago, we'd be a lot more pleased about how our attack has been functioning. But given how excellent we were last season, even having an attack that is statistically the fourth best in the league feels like it's underwhelming.
We currently have an xG difference per 90 of +1.15, which ranks us third in the league behind City and Brentford, and which also betters the +0.79 we finished with last season. To caution this metric from last season, when we were good (until the Saliba injury), we were around the +1 mark consistently.
An offensive metric in which we are currently certainly underwhelming in is the shots we are producing per match. We're currently only taking 15.75 shots per 90, which ranks us 8th in the league. While there are several teams that are drastically over-performing in their shot production (last season's league high was Brighton with 15.97, and there are seven teams currently with 16 or more) and who will likely drop off, it does feel like we aren't creating the volume of shooting opportunities that we could do once those attacking players hit form.
There is certainly a game state element that comes into this metric, with such a limited sample size. We played around half an hour with ten men at Selhurst Park, a period in which we produced not a single effort on goal, while in the 50 minute period where we held a 2-0 lead over Forest, we went into second gear and produced a total of four shots. There's something to be said about the way we approach games in certain game states and how we can maintain potency in these scenarios, but they certainly impact our overall shot output in a way that may not reflect how good we actually are offensively.
Defensively, teams aren't managing shots against us, and that's partly due to how much we control games. Teams generally aren't getting anywhere near our goal. We have conceded the second fewest shots per 90 (behind City of course) and had the third fewest opposition touches in our own penalty area per 90 (16.8 - United have had 35.3 lol). In two and a half games, this has been down to our ability to dominate possession (78% against Forest, 71% against Fulham and 69% in the first half against Palace before the red card changed things).
Keeping the ball and winning it back quickly on the occasions we do lose it means we've been able to pen teams in for large periods and thus prevent them getting at our goal. Against United, there was a clear directive to allow Ten Hag's team to have the ball in their defensive third and wait for them to give it us in our mid-block, because they're so poor at building through the lines.
Defensively, we're on course to be better than we were last season at this early stage. The goals we've conceded have either been in transition (Awoniyi and Rashford), which is unfortunately a symptom of playing an aggressive high line, or through lapses in concentration (Pereira and Palhinha). There aren't signs of systemic or structural defensive issues, which means if we can iron out conceding these stupid early goals and tighten up slightly on the counter, we could become impenetrable on the defensive front.
Player Specific Data
These are the only real takeaways from a holistic team standpoint at this early stage, from the limited dataset we're working with. Perhaps I'll do one of these at each interlul, and see how drastically they change.
As for player specific data, this is certainly not an area I'm going to go big on because the sample size makes conclusions even more fanciful. If a player has created a single big chance then their expected assist numbers will be hugely inflated, while the same can be said with their xG numbers. For example, Gabriel Jesus is currently at 5.06 xG per 90 because of his single huge chance one-on-one with Onana, while Fabio Vieira is currently producing 2.0 expected assists per 90. Of course these are two players with minimal minutes so represent the extremities, but it still stands for those who have 3-4 90s already.
However, I do think there were some interesting little bits worth mentioning. There were doubts surrounding Declan Rice's progressive passing prior to his move to North London, but the number 41 is currently completing 8.5 passes into the final third per 90 (a career high) and 6.25 progressive passes per 90 (a bordering on elite number for a central midfielder). Most would agree that Bukayo Saka hasn't hit the ground running in accordance to the expectations surrounding him, yet he's producing a career high 5.78 shot-creating actions per 90. Even not at his best, he's incredibly influential.
Kai Havertz has endured a difficult start to the season in his new role, but is putting up career highs in certain criteria you'd want from a midfielder, including pass accuracy, passes into the penalty area and passes into the final third. He's doing the things we doubted he could do as an eight, in ball retention, ball progression and the defensive work. He's not yet provided the value in the final third, which is the stuff we believed he'd be an upgrade on what Granit Xhaka could produce. Let's give him a bit more time.
That wraps up this data round-up after four Premier League games. It was almost futile, but there were some interesting points. I'll check back in on this front in five weeks time, when we'll have our next international break, and when there should be some slightly more conclusive statistics.
Comments