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?
A pretty meandering aimless article rounded off with a good dose of disrespect and ageism. Arrogance. Last time reader of this site here
lol, you’ve only ever been critical, for example, you didn’t like that I wasn’t harsh enough on Arteta in 2021, when most of the other folks here thought I was too harsh.
As for the ageism, I’m 52. I’m literally talking about my generation who like to live in the past and who need to think the past is modern. It’s not.
But I guess you won’t be reading any of this or anything else!
Thanks Tim for perspectives we did not even envisage.
Okay will say that LOUD
Ramsdale makes mistakes, but all keepers do. Even Allison. And De Gea, who can be an amazing shot stopper, simultaneously also makes a bunch of errors.
And I still think I’m happier having Ramsdale vs Leno or Emi. He’s definitely better at distribution, which has been very important to our improvement this year.
On the old stuff front, some of the football in the 90s was definitely pretty ugly. That said, you could go back to the Dutch Total Football in the 70s and say that it bears a lot of resemblence to some current trends, FBs becoming central midfielders and CBs comfortable carrying the ball and playing long, accurate passes to launch attacks. Seems a pretty traceable line from Michels to Cruyff to Pep to Arteta.
I’m not saying that history is irrelevant. And while there are definite trends (Pep does actually play a WM) you can’t possibly think that football – with all of the data and science that’s been injected the last 7 years, I mean just look at how shots from outside the box have gone down, or how almost no one plays Allardycian football anymore – is the same now as it was even 5 years ago? It’s wildly different, as I say in the article. Pressing and playing out the back are two very obvious areas.
You are obviously right, Tim. The game has evolved beyond recognition. I do however think that things like “sorry, Pops” is border contemptuous. Also assuming that people thinking that 90’s football is modern have an agenda (making their era relevant) is a step too far. It could just be lack of expertise, lack of analysis. Or the fact that, as all evolutions, this one was gradual and could go unnoticed by an average viewer.
I will, however, keep reading you avidly.
Well, the people I was talking to are
1. younger than me (I’m 52)
2. considered experts by every known metric
I was looking at fbref for evidence that Ramsdale is good with the ball at his feet, but it seems mixed at best.
Also, please stop using “kid”, “kiddo”, “whippersnapper” etc. It upsets the young’ns.
fuck them kids!
I thought Chapman used WM in the 1930’s Arsenal. Anyway I’ve been a gunner since 1946 and do not fit your profile of an older supporter. The game has evolved and is still evolving in leaps and bounds. The pace, skill and fitness levels are orders of magnitude better than ever before. Even the women are so much better than the men of my era.
Old is not stupid it is more often experienced.
I love that people tell you that they’re never gonna read again who f****** cares. Great article enjoyed it I think Ramsdale is better than middling but not top shelf yet, he may get there though.
He’s been coming here since 2009. He usually says something caustic about the article and then bogs off for a few years.
For me Ramsdale is better than average. He does make some amazing saves and gives the defence confidence. He does not pass through the middle as much as last season. I was really impressed with Allison in our game against liverpool. He was brilliant with his feet and countered Arsenal press like it was nothing.
Any metric that rates Leno at number 2 and Ramsdale at number 91 is as relevant as using shoe size to grade musicians. Having watched both men live many times, I would choose Ramsdale over Leno every time. His command of his box, his communication with and organisation of his defenders, his confidence with the ball at his feet and his contribution to the team’s morale far exceed Leno’s abilities in these areas. Every goalkeeper has let in a shot he could have stopped if he had made a different decision. With any player, you get a package rather than a selection of qualities that can be converted into a meaningful metric. Undoubtedly, Ramsdale has flaws but he is still a young goalkeeper and he is still improving.
Well, I guess I wasn’t clear: Leno is ranked #2 in psxg +/- meaning that he’s stopped more goals than expected, while Ramsdale is just average in that metric. That doesn’t make Leno the #2 overall keeper in the top five leagues.
Recently, I had a fun discussion with a fellow supporter that went as follows: Given (1) the popular narrative that it wasn’t Holding’s defensive play that cost us in the 3 points from 12 run, but rather missing Saliba made it so we couldn’t play out from the back, and (2) the notion that Matt Turner is the better shot stopper but worse at building up, should we have made a change at keeper and relied less on playing out from the back?* My friend rejected the premise that Ramsdale’s shot stopping could be improved on.
* Notes: The drop in form for #5 is probably a much bigger factor, and given that this team seems to fall apart when we made too many deviations from the top 12 swapping out the keeper doesn’t actually make sense.
Letting Martinez leave and signing Ramsdale is probably the most Arsenal move ever.
Ramsdale might be a keeper whose intangibles exceed his pure ability as a keeper. He’s good with his feet, good with distribution and shows leadership. He plays his best in big games – usually.
Is that the last we ever see of Kiwior?
There is something a bit soft, a bit Squillaci about Kiwior. Great interceptor and passer, but looks to me not a PL type defender yet. The handling and feet trodding in the box is standard PL fare, no CB goes down to check on his boot while the ball is in play. Absolutely weak defending.
Farooq: not Kiwior but Tierney. Had a terrible match. Something like just 7 touches? I think he is out.
Mitoma is an absolutely spectacular and well rounded winger. Comfortable both cutting in and swinging in crosses, and a great temperament- no wonder they were fine letting Trossard go.
Now that the season is over, hopefully we can start doing a feasibility check with ESR/Odegaard as twin 8s.
Great article. I am probably unduly harsh on Ramsdale because I would love to have an above average shot stopper, and I think it is an undervalued skill. I also think it is a particularly hard position to play. Nobody else has the gloves on, unlike scoring, where it gets shared around players. I think Ramsdale is probably the weakest member of the fully fit starting XI, but he is also very young for a keeper and there’s room to improve. I often cite the John Muller article for 538 where he looks at the shot stopping of Matt Turner and proclaims him the most valuable player in the MLS, and also says that possession value added for any goalkeeper will never outweigh shot stopping. Not to say Matt Turner should get the No1 spot, just that it is the primary job of a goalkeeper. Okay, thanks again.
Comments are closed.