I was listing to probably the Tifo podcast (which spent about 40 minutes of the intended 60 minute episode trying to be funny with American accents) or maybe it was another podcast, I don’t know and I don’t care because it’s not important. The important bit is that one of the presenters was asked who the best goalkeepers in Europe were this season and he essentially read the sorted list of psxg+/- from the FBREF site. Which looks like this:
They then spent 10 minutes trying to pronounce Reims – which Google tells me is RAANS. Now, obviously, as an Arsenal supporter, hearing Bernd Leno’s name at the top of the list piqued my interests and I went to FBREF and sorted the list to see where Aaron Ramsdale pops up and it’s… 91 with -0.3. That’s below Emi Martinez but well above the worst keepers in the league: Meslier and Bazunu. Bazunu is actually the worst keeper in the top five European leagues with a -16.6. That’s pretty impressive!
A quick word on what this measures and why it is and isn’t important: this measures goals allowed by a keeper and how close they were to expected goals. The idea here is that they are measuring how well a keeper saves the goals they are supposed to save. A high positive number suggests that they are saving a lot of goals that an average keeper would allow and a high negative number suggests that they are allowing a lot of goals that an average keeper would save.
Ramsdale has made 90 saves and allowed 39 goals in each of the last two seasons (weird, I know) and his psxg numbers are +0.2 and -0.3. Which means that over the last two seasons, he’s saved about exactly the number of goals you’d expect from an average keeper. Which I think passes the eye test. Some days he makes great saves, other days he lets in low chance shots. There is also some variance here: allowing a lot of goals from distance (by a lot I mean, say 10) could be a sign that the keeper’s legs are done (as we saw with Cech in that disastrous 15/16 season) or it could just be bad luck. Either way, goals from distance will hurt this number a lot more than any other stat because most goals from distance are a less than 3% chance.
Ramsdale is 5th worst in the league in goals allowed from outside the box. He’s conceded 7 on 27 shots. Some of those goals, however, were just absurd wonder strikes (Rashford’s opener, for example) and I think we’d need a much deeper dive into the data than I’m going to do to see if he has a problem here or if there’s just some variance. But I believe that this stat is what’s dragging his psxg +/- down.
Another stat they mentioned which is I think is controversial and which I believe isn’t an individual stat is clean sheets. For example, David de Gea has a league leading 15 clean sheets. Meanwhile Alisson is 2nd in clean sheets with 13 (tied with Pick Nope and Aaron Ramsdale). De Gea’s psxg is -1.7 and Alisson’s is +9.3 (he’s the best shot stopper in Europe this season). If clean sheets were a meaningful metric tied to keeper performance, I would expect some correlation, but there just simply isn’t one because while the keeper CAN help keep a clean sheet, you don’t actually need a keeper to get a clean sheet; since a really great defense can prevent the opponent from ever getting a shot on target or by limiting the shot quality.
The team matters more than the keeper, in my opinion, when it comes to clean sheets. Aaron Ramsdale’s clean sheets are a great example. He only has 3 at home but 10 away from home. Is he just better at saving goals in away games or does the team adopt a different approach in away games? I think you’ve been following Arsenal for the last few years, and I bet you know the answer.
Another odd keeper stat is errors (errors leading to shots and errors leading to goals). These have long bothered the anti-stats crowd because they are “subjective”. In other words someone has to decide that something was an error. Well, I hate to break it to you, love, but all football stats are subjective. Someone is constantly deciding whether or not some action was a pass or a shot, whether it was a bad pass or a miscontrol, whether there was a tackle, when did a player dribble, and so on. In fact, you watching the game is subjective!
But back to errors. Let’s just assume that we accept that the errors data is accurate, Ramsdale is again, just an average keeper. He’s allowed two goals off errors, which isn’t great but also isn’t horrifying. Obviously we want no goals off errors but it’s also part of the gamble that Arteta takes (and most managers these days) in playing out the back. Ramsdale actually has 6 total errors which have led to a shot or goal this season and most of them have been passing errors, passing out the back. You can quibble about whether or not it’s his fault but the one that most immediately springs to mind is the pass for Southampton’s first goal in the 3-3 draw. He doesn’t “do this all the time” but there are other prominent moments like this, when teams press Arsenal high. The funny thing is that 5 years ago we’d have slaughtered the keeper and manager for playing like this! The fact that we know why they do it and accept it shows how far football has come in the modern era.*
I’m not hugely critical of errors unless they become a prominent feature of the player or team’s performances. For example, the second error Ramsdale committed for a goal was the header by Martinez in the United match. I watched it again and ok, he flaps at the cross, and I guess that’s an error. One error like that and one bad kick that cost Arsenal a few goals (and maybe two points) are acceptable from a keeper. Especially one who is fairly solid in distribution, the way that Ramsdale is!
*The “modern era” for football is a constantly moving bracket of about 5-7 years. Anyone who thinks that the modern era of football started in the 90s is an old person who wants to believe that their younger days’ football experiences were modern and are still relevant. There is nothing modern at all about the way football was played 30 years ago, sorry, pops. What was the big invention back then? Banning Mars bars? Look at Premier League football in just the last seven years: passing it out the back, inverted fullbacks, Guardiola’s revolutionary WM? I’m not saying football didn’t matter (to you) in the 90s, I’m merely saying that what was modern then is absolutely old fashioned now. Imagine a team playing football the way United did for most of the 90s and 2000s? Why, that would be David Moyes’ 52 crosses United. Remember that?