There are no satisfying answers to the Arsenal slump

This is in place of the debrief today because I was going to write about this topic there and it is probably a more interesting to talk about than that match.

“Football is the only sport where you can lose by playing better, that's why it's so attractive to people” - Pep Guardiola

Arsenal are in a funny spot right now. They are playing well, perhaps the best they have played all season, but the results are not going their way.

Like the quote from Guardiola above this is partly why we love this sport and partly why we also hate it. I find it frustrating to write about because it makes trying to do helpful analysis hard. I badly want to be able to point to something like such and such is the problem, or this is going wrong, and this move is causing problems, but right now because I think generally the team is playing well it just boils down to the players should just kick it into the net more often.

Where we in the present situation reminds me of a classic comic from xkcd about sports and it feels like it rings true, especially right now.

https://xkcd.com/904/

Just take a look at some of the headlines about Arsenal right now:

Arsenal lack cutting edge – does their transfer strategy have to change this month? - The Athletic

Arsenal's FA Cup exit proves they must sign a striker soon - ESPN

Mikel Arteta has no choice but to boost flawed forward line - Sky Sports

Will a new striker solve problems for faltering Arsenal? – Football Weekly

You could probably go to just about any major publication and find their Arsenal section and see similar types of stories (not exclusively, here is a good example) right now about how the finishing means something must or should change to fix this problem.

The only issue is that this assumes that this is a structural problem and I just don’t buy that premise at all.

It is probably too dismissive to say that this is simply bad luck because it is more than just that with the team pretty clearly struggling with their shots (and to a certain extent the passes that set them up) but it is also exactly that bad luck, especially the confluence of multiple players going through it at the same time.

Looking at the finishing for Arsenal under Arteta going back to 2021/22 season you can see that the team’s number track really well their expected numbers.

There was a similar type of slump at the end of 2021/22 but the numbers rebounded towards the expected numbers afterwards. There was a couple periods of overperformance last season at the start and end which helped propel Arsenal to a title challenge, sandwiching a period of underperformance. This year the team started a bit cold, and rebounded to match the expected goals and are now in this current slump.

Try as hard as you want, but it just doesn’t seem like there is a systemic problem of under-performance compared to expected goals here.

Breaking down the players for Arsenal I think suggests that this is simply players that have generally been good (or even bad) finishers going through a rough patch all at the same time.

Here are what the main players in the front five have looked with their goals compared to expected goals like with their time at Arsenal and with a moving average of their last ten shots.

Saka after a slow start with Arsenal (playing further away from goal), has rounded into a positive finisher according to the chances that he has had. There have been some runs where has fallen below but these have been rare recently. We just happen to be in the middle of one the more prolonged down portions for some time.

For Martinelli, the story is similar to Saka. He started off his career with below-average finishing but turned that around with a very hot run last year. This season he is down overall, with him going through a rough spot currently as well.

Odegaard’s chart here looks really nice and I think points to a player that is a good overall and pretty consistent finisher. He has had spells where he is under his xG but they have been mostly short periods. Like Saka and Martinelli, we however find ourselves currently in one of those periods.

Jesus came to Arsenal as a player with a reputation as a bad finisher and he hasn’t done much to dispel that conclusion with his play. He does have a goal among his last 10 shots so he is a bit above expected, but that is just leaving the sample and as much as I like him as a player, he’s not one that I would expect to go on a prolonged hot run.

Havertz is another player who came to Arsenal with a reputation as a bad finisher and it hasn’t been proven wrong yet. He started below with some big and notable misses but has turned it around with a few goals pushing him close to even overall for the season.

Nketiah had built himself a bit of a reputation as a good finisher but his track record now with over 100 Premier League shots is pointing the other way. It is still probably too small to say with a lot of confidence that he is below average but he is certainly not on a hot run currently either.

Last is Leandro Trossard, and it is another story of a recent downturn in finishing.

Looking at this, it does look like we have players we expect to finish better going cold and players that we didn’t think were positive finishers not changing that opinion with their performance. It is a tough time for it all to come together, but that is how it goes sometimes.

Variance gonna vary

The start of this slump in front of goal really started with the Aston Villa match. Since then Arsenal have scored just 4 goals in the matches against Aston Villa, Brighton, Liverpool, West Ham, Fulham, and then Liverpool again.

In that time Arsenal produced 113 shots in those matches worth 12.4 xG on my model. Plugging each of those chances into a simulator you get a goal probability distribution that looks like this.

There is a tiny green bar that shows the likelihood of scoring just 4 goals from these chances

Scoring just 4 goals from this is highly improbable and really bad luck to have it turn out that way.

One of the things I have heard and seen is that one of the reasons for this is that of the 113 shots Arsenal have taken too many of them are low quality, and that is a fair thing to wonder. Taking lots of low-quality shots will juice the xG totals but doesn’t lead to lots of actual goals.

Unfortunately, it just doesn’t seem to match up with the data as an explanation right now (like I said, I wish there was an easy answer).

In the Premier League before the Aston Villa match, Arsenal had an average shot quality of 10.5%, with an average of 2.1 big chances per match. Since the start of this slump, Arsenal’s average shot quality has been 11%, with an average of 3.5 big chances per match.

Of the shots taken I would classify 42% as low quality (less than 5%), 37% as medium quality (5% to 15%), and 20% as high quality (greater than 15%). That’s compared to 47% low quality, 35% medium quality, and 18% high quality for the Premier League average this season.

You might be wondering to yourself, “What about shots being blocked? Maybe that is up, and could be the reason.”

Except it also isn’t the case that there is a meaningful difference between the number of shots being blocked. Pre-slump Arsenal had 33.3% of their shots blocked before they could hit the target, in the slump that has increased marginally to 33.6%.

The one item that has also dropped and makes sense as a problem for finishing is the number of shots that Arsenal are putting on target as a percentage of total shots.

Pre-slump Arsenal averaged 32.9% of their shots finding the target, which was already below the League average rate of 33.6%, and that has fallen to 25.7% during this bad run.

Like most things, the percentage of shots on target fluctuates but has generally tended to come back toward being around League average.

The lack of shots on target has made the post-shot xG for Arsenal quite a bit lower than the overall expected goals in this run, going from 12.4 to 7.1. But even this still doesn’t fully explain why Arsenal have converted so few of the chances into goals with the team still in the outlier region of the simulated distribution of outcomes.

Using post-shot xG instead of just regular xG

This is also still an unappetizing item for trying to say that there is something bigger going on.

For the season Arsenal are still scoring slightly MORE goals than their post-shot xG would indicate, it is just over this run of matches that it has gone negative.

TL;DR

Arsenal need to kick the ball better, and if we wait it will almost certainly start happening.