I always think comparing goalkeepers is tough. There are so many variables and neither a data led analysis or the eye-test are really sufficient
On Sunday, WhoScored tweeted out what they think are the key stats to comparing David Raya and Aaron Ramsdale. But there are so many flaws in what they put out there.
Saves: This really does mean nothing when comparing goalkeepers.
The only thing it shows is a goalkeeper having more work to do, and how many saves a keeper makes is probably more more indicative to how poor a defence is.
Ideally, you want to be stopping an attacker before they get their shot in on goal.
David Raya saving 154 shots shows that Brentford allowed too many shots at goal last season. No other keeper made more saves. In second place is Bernd Leno and 3rd is Jordan Pickford.
Fourth is Alisson, which really highlights this point.
In 2021/22, Alisson was 15th in most saves. 2022/23 he rose to 4th. This did not mean that he was a better keeper last year than the year before, just that Liverpool’s defence was worse. Highlighted further with the fact that Liverpool went from conceeding 26 goals to 47.
Ederson made the least saves of any regular keeper last year – just 46. But like Alisson’s high figure showed how poor Liverpool’s defence was, Ederson’s low figure was more to do with how great Manchester City’s defence was.
Raya made 60 more saves than Ramsdale, but faced 61 more shots.
Save success rate: Some will argue that the way around the flaws in ‘saves’ data is to look at save success rate.
Whilst this is a much better way to compare goalkeepers, it is still flawed. It does not factor in the difficulty of those saves.
We have all watched a games where a keeper has made 10-12 saves and thought “what a great game they had”. But then you watch the highlights and see every shot was straight at him, and that he should have (and did) save all of them.
Say you have two keepers in a game. One saves 10 out of 10 shots (100% success rate), the other saves 8 out of 10 (80% success). Save success rate would have you think the one with the clean sheet outperformed the other.
But what if those first 10 shots were straight at the keeper. He did not pull off any saves he had not right to make. Then with the second keeper, the two he conceded were the old school sweaty FIFA goals where 2 players are clean through and one squares it to the other. Keeper stood no chance.
And then in the same game, the keeper that conceded twice pulled off two David Seaman v Sheffield United-esque saves.
Would you say the keeper who made 10 easy saves outperformed the keeper who had no chance with the goals he conceeded, and pulled off a couple of worldies?
Whilst ‘save success rate’ is not a way to establish who is a better keeper, it is a good way to see who is a poor keeper. Last season, Leeds United’s Illan Meslier saved less than 60% against him.
When you are looking at Ramsdale v Raya, 70% v 77% is probably not a large enough gap to be relevant.
Clean sheets: Like saves, clean sheets is fairly irrelevant. It usually just highlights who played in a better team. Or for mid-table teams, those that play more defensive football.
Manchester United’s David de Gea kept the most clean sheets – 17. But Man U scored just 58 league goals. 30 less than Arsenal. That would show that Man U tended to play more defensively, whilst Arsenal were more attacking. And the more you attack, the more open you can be at the back.
Ramsdale kept 14 clean sheets to Raya’s 12. This is not a surprise considering Arsenal finished 2nd and Brentford 9th, with us scoring 30 more goals.
Accurate long balls: It is interesting that in an era where most keepers pass it out from the back, WhoScored has used ‘accurate long balls’ as a comparator
Raya dwarfs Ramsdale with 410 accurate long balls against just 138 for Ramsdale.
The Brentford keeper attempted 973 long balls last season, meaning 42.1% hit their target. Ramsdale attempted 472, completing just 29.2%. And this is where too many variables comes into play.
Ramsdale was clearly instructed to play it short a lot more, with Arsenal looking to move the ball forward along the ground. Brentford were clearly a long ball team.
This meant Brentford would set up every time Raya had the ball like it was a set play. They would overload one side and have targets for Raya to aim for.
In Ramsdale’s case, he only really went long as a last resort or if he saw a quick break opportunity; this would mean Arsenal would be less likely to be set up for a long ball from the keeper set-play.
You also have to factor in who the target of the long balls is.
Ivan Toney is a much better target man than Gabriel Jesus. Kevin Schade, at 6 1, is a better option than 5 10 Martinelli. Brentford have better players in the air than Arsenal, so it makes sense more long balls hit their target.
I would not be surprised if this season Arsenal’s accurate long ball percentage increases. It will not be because of Ramsdale improving (or Raya joining) but because Kai Havertz is better in the air than anyone we had last season. He will be the target in that front line.
Goals conceded: See clean sheets.
Errors leading to a goal: This can useful at times, highlighting a keeper that might be error prone. But can we read too much into Raya making one error that led to a goal against Ramsdale making two? The data is too small.
Eye test: Finally we come onto the eye test.
The eye test is what is now used by hipsters to discuss traditional scouting where a fella with a flat cap goes and watches a player a few times and writes his report. But even this can be flawed. And the fans eye-test is even worse.
The fans eye-test is the most unreliable of them all when it comes to keepers.
Most of us only get to see the likes of Brentford play on Match of the Day or whatever highlights package you watch.
Keepers lower down the leagues will have their good performances against top teams amplified by producers. Plenty of praise and analysis. There poor games tend not to have the same coverage.
It is a 180 for keepers playing for top teams.
A keeper at Arsenal, Manchester City or Man U will have a poor game analysed in depth. whilst when they have a great game against lower opposition, it is glossed over.
This can lead viewers to believe that a lesser keeper is better. I call it the Jussi Jaaskelainen affect.
Jaaskelainen was a top keeper for Bolton, and always seemed to raise his game against Arsenal. MOTD would go out of their way to highlight his performances 4-5 times a season. But he would also have 4-5 games where he would have a stinker. Yet thse would not get the same airtime.
This would lead fans at the time to think Jaaskelainen was one of the best keepers in the league. It was not true. It just meant that, playing for a poor team, he had more chance of being a hero.
As for the eye test from scouts, again this can be tough as it is all dependant on what games you go see them in. Just because a keeper has a poor game, does not make him a poor keeper. Nor does one good game make him world class.
So what then?: Expected saves is a good metric that I like. It shows when keepers make saves that they should not, whilst also factoring the saves that they should make.
Post-shot expected goals (PSxG ) is expected goals based on how likely the goalkeeper is to save the shot.
If a shot is direct at him, you would not expect him to concede. Whilst that Samean-esque save would carry a high weight for conceding a goal. A keeper that has a higher than 0 PSxG has outperformed one that has a lower than 0.
Last season, Raya’s was +5, whilst Ramsdale was -2. That means Raya conceded 5 less goals than he should have, whilst Ramsdale conceded 2 more.
Whilst this is a good metric for seeing how good a keeper is at saving, there are still plenty more variables that have not been factored in such as: handling at crosses, coming off the line for intereceptions, short passes, command of box and more!
Often, it is the work on the training ground which shows who the better keeper is.
The jump test, reaction test, speed test, handling test. These are what teams use to determine who is the better physical and technical keeper. But that does not count for what is between the ears.
Manuel Almunia was one of the best in the league when it came to those tests in training. But when he came to being on the pitch, he did not have the composure. And that is a big factor.
So who is Raya an upgrade on Ramsdale? Personally I think it is close to call and we might only find out when both are an Arsenal player.
Enjoy your Tuesday.
Keenos
She Wore A Yellow Ribbon 
Hello ken.
At first i was excited we want to sign raya, but after reading your thought i have become worried. Do we arteta and co really do their scouting on raya well? They should pop in here and clear my doubt o, befor i break down this wall in front of me.
LikeLike
I think David Raya might be attractive for Arteta for the reasons not mentioned above. Let’s analyse the reason why Ramsdale was bought by Arteta 2 seasons ago. It was his character, leadership qualities and confidence that Arteta most drawn upon. That constant energy, outrageous flair and admirable presence. Aaron Ramsdale is very hot blooded. Maybe Arteta needs to mix it up with someone cold blooded. A mixture of humility, dependability and intimidation is what David Raya is. The age profile of Raya also can help Ramsdale to grow as a professional.
LikeLike