And yet they remain the best keepers in the league, basically rendering the entire argument around De Gea redundant. Shot stopping alone isn’t enough.
1
Season Average Rating
- Appearances
- 46
- Clean sheets
- 10
- Goals
- 0
- Assists
- 0
- Yellow cards
- 0
- Status
Not quite. It reflects how they're performing in that specific regard this season, not how good a shot-stopper they are generally. For example Alisson was 3rd best in the league last season while De Gea 5th worst. So you could no more conclude that Alisson is a poor shot stopper from this season's stats than that he was the third best shot stopper last season.
They're the keepers for the best teams in the league, not necessarily the best keepers.
It’s one of those occasions where stats and “the eye test” align. De Gea’s shot-stopping was unusually poor last season but back to his usual quality this season.
Obviously tells us nothing about the other, more long-term, flaws in his game.
Yep. There was an argument that his reflexes had slowed which his current form has probably put to bed but other than that we're just seeing what you'd generally expect from De Gea: very good at shot-stopping, not particularly good at anything else. After 460-odd games for the club I think we have him fairly well sussed.
Cheers for the mention @roonster09!
@TrustInJanuzaj The only minor thing I'd add to @Classical Mechanic's wonderful post is that a crucial element of post-shot xG is that shots off-target aren't considered - whilst these are factored into standard xG as chances.
I also think you bring up an interesting point vis a vis Ederson and Alisson - one reason why Ederson and Alisson do poorly by this metric is that they are extremely proactive and quick off their line - this results in many chances that don't come to fruition, and also results in some finishes becoming far easier than they otherwise would have been as there is no one in the goal.
I'm not 100% sure how different models account for this, but perhaps an example makes the most sense - off the top of my head (and obviously unrelated to any personal allegiances), consider Kai Havertz's goal in the CL final. That goal, which was hit more or less right in the middle of the goal rather slowly on the floor. This would result in quite a low post-shot xG figure - since shots hit into that area of the goal at low speed are almost always saved (for the sake of argument let's call it 95% of the time it's saved). This therefore would result in a -0.95 PSxG figure for Ederson for this shot after the goal was scored.
Like I said, I am not certain whether this is accounted for - but I suspect this metric may underrate Ederson and Alisson given their proactive approaches (although at the same time I wouldn't argue that shot-stopping is a particular strength for either).
I understand that, I get how PSxG works after reading this thread and FBREs own definition.
That wasn't the point I was making though.
My point is simply - who or what is the baseline?
Let me explain why that is important...
If I was the baseline, every PL GK would look amazing, because they would all be saving loads of shots every game that I wouldn't save. At the end of each game, they would all have basically +PSxG equal to the number of shots they faced - because if I am the baseline, I would save nothing!
If prime Peter Schmeichel is the baseline, then the bar is going to be really high, and it's going to be hard to earn a positive PSxG
Now, why does that matter?
Lets say the "baseline" is what the average GK saves in all top flight, second and third division games across the major European leagues. Obviously the standard is going to be fairly high.
However, you would expect that ALL PL GKs would fall into the "elite" category i.e. they are all expensive, highly-paid, international GKs operating in arguably the best league in the World.
So, again, why is that important?
Well, if the base average is set fairly low (in the context as described above) then the PL GKs who faced the most shots would be at an advantage, because they would have more chance to earn PSxG
I strongly suspect this is the case and that De Gea has improved statistically this year purely because he has faced far more shots than he usually would. At one point under Ole we had faced the 2nd most shots in the league.
This would also explain why GKs like Allisson and Ederson feature lower down. They face hardly any shots, in comparison to the GKs at the top of the list.
De Gea probably has improved slightly this year, because we really were at rock bottom for 2/3 seasons. However, I find it ridiculous to suggest he is "saving us" because he pushed a header from the penalty spot away at a nice height and tipped an effort over the bar that he barely had to move to reach - yet supposedly he scored a +2 PSxG for that game...nonsense I am afraid. Give Allisson, Ederson or Karl Darlow those shots and they save them and earn the same +2 in the process.
Interestingly, FBRE also themselves say that "luck" has a role to play in GKs scoring good PSxG too. Until I could actually interview a member of their team and understand how this is actually calculated in detail, I can't take a great deal from it.
Last edited:
It's a valid point you raise but I believe StatsBomb uses competition-specific data - so the probabilities are all based on top-flight situations.
They play for the best team because they are the best keepers. Would you hand on heart take De Gea over either?
No, both bring more to the table overall. It's still notable that De Gea has returned to being among the best in the world at this specific aspect of goalkeeping, though.
I think it's entirely fair to suggest that De Gea's skillset would be better optimised at a lower-level team - given the extent to which goalkeepers have become essential to elite possession sides, having one who is a huge negative there is problematic.
I'd take him over Ederson, 100%.
Edit: City's and Liverpool's strikers must be better than Ronnie as well, good to know.
Fair enough. I think Ederson is top class. Again he isn’t as good of a pure shot stopper but he beats De Gea by a mile at everything else.
Definitely been our best player this season ..unbelievable performance against Norwich.
I'm not really seeing why the baseline is important as all the keepers in the PL are measured relative to it thus creating a legitimate hierarchy. If the data was taken from all top 4 professional English divisions then the top keepers would simply outperform the baseline more
On the 'luck' element. It is mitigated for to a degree through PSxG SOT, which takes into account how hard the shots were to save on average. It's an average too so it doesn't matter so much if a keeper has faced more shots. I'm not saying there aren't blindspots but I don' think they're as pronounced as you seem to.
The scatter plots used to illustrate these stats express this.
Tweet
— Twitter API (@user) date
A couple of weeks ago Ramsdale was slightly ahead of De Gea on PSxG but had faced easier shots, for example. That point also shows the volatility of these statistics when looked at with such a small sample. DDG has raced into the lead on the back of a couple of games. A few bad games could render him average again. His shot stopping has been excellent so far, however.
That DDG and Ramsdale are ahead of Allison and Ederson on this metric currently is neither here nor there. In Allison's time in the PL he's finished as one of the best shot stoppers by these metrics and in the average range. Ederson is invariably in the average range. The point is that these keepers bring a lot more to the table overall than DDG. Personally I think Allison is the best all round keeper about but this doesn't mean that his shot stopping is always the best about. I feel Ederson is very much a system keeper and perfect for Pep with his ability on the ball but his 'traditional' fundamentals are not elite. As for Ramsdale, he's a very good all round keeper at his best. Not elite, at least so far but a lot better than a lot of people give him credit for.
There is no excuse for that calamity of a transfer - Kepa.
What a disastrous signing.
Tweet
— Twitter API (@user) date
I've always been a massive De Gea fan ever since I watched a Athletico vs Barcelona match while terribly hungover. I was very excited when we signed him and seemed to be in a small minority of his defenders during his early struggles.
Thought wow its been a long time he's been with us now, he must be on track to be one of the top appearances but it seems he's actually way off. What am I missing here? He's now 31, signed as what an 18 year old? Has been first choice most of the time other than a few short periods. There should be a good few years left in him anyway so hopefully he'll make his way up there.
I think it would fitting for him to go down as one of our highest capped players and I do think he deserves the status of legend at the club even if they have largely been barren years he's been here.
Thought wow its been a long time he's been with us now, he must be on track to be one of the top appearances but it seems he's actually way off. What am I missing here? He's now 31, signed as what an 18 year old? Has been first choice most of the time other than a few short periods. There should be a good few years left in him anyway so hopefully he'll make his way up there.
I think it would fitting for him to go down as one of our highest capped players and I do think he deserves the status of legend at the club even if they have largely been barren years he's been here.
Bring back Eric Steele.
So what is the difference between PSxG and PxSG SOT? Because I was led to believe PSxG took into account the difficulty of the save?
The baseline is important because if it's set too low, it would be biased in favour of PL GKs who face more shots.
I suspect it is set too low because I do not believe for one second that De Gea saved two goals against Norwich that the average GK wouldn't save. (I checked - he scored +2 PSxG for the game)
I also suspect this is the reason De Gea has supposedly improved. He hasn't...he has just faced significantly more shots this year and had more chance to earn positive PSxG points, unlike better GKs like Ederson and Allisson who never have much to do
Look at the GKs who are always supposedly ranked as the best shot-stoppers. It's always GKs who face loads of shots! Henderson topped the metrics by a mile two seasons running when on loan at Sheff Utd.
De Gea gets to face his old club in the Champ league!
PSxG SoT is the average PSxG that the keeper is facing so indicates if the volume of shots faced is distorting the PSxG. In this particular case it's showing that DDG has been facing the hardest shots on average in the league which strongly indicates that his high PSxG value isn't some sort of distortion and that he is performing at a very high level shot stopping wise this season so far.
Allison was third is PSxG +/-/90 last season so it isn't always keepers that are facing lots of shots that are top of the charts. That said, nearly all keepers in the Premier League will be excellent shot stoppers as it's the fundamental goalkeeping skill and the PL is currently the top league in the world whilst being the richest to boot. Top clubs want something more these days, they want complete goalkeepers which are much rarer so it's more of a balance. Ederson isn't a world class shot stopper but there are probably no better goalkeepers than him on the ball which is what Pep wants. I don't think there's a top club in the world today that would sign DDG, maybe Atletico under Simeone if they needed a new keeper but not one with a modern style which is indicative of the change. If you look at the PSxG - GA per 90 for last season then there's not actually that much between the keepers in the PL when it comes to shot stopping.
The baseline is the average drawn from the players in the data group. In this case The Premier League. It doesn't matter if the you had the top 4 professional leagues in the data group because the best will always be at the top so it will strongly indicate what you're looking for.
edit: a data group comparing multiple leagues would become and issue if the players were compared together because of the varying qualities of the leagues. I think it could work if just the goalkeepers from one league were compared against that average although it would be better to compare within a single league. FBref scouting reports do compare players against the players from the top 5 European Leagues which if it's just a straight statistical comparison advantages top players in weaker leagues.........at least I think.
Last edited:
Tweet
— Twitter API (@user) date
Meslier is interesting. A media report recently that we're watching him. Going by the notorious eye test he does seem to make quite a few bloopers though.
In a similar vein:
Tweet
— Twitter API (@user) date
Also confirms that his shot-stopping really has been excellent tbf but his weaknesses are his weaknesses.
And Sanchez is someone I think better clubs than Brighton should be looking at.
Sanchez is a proactive goalkeeper, and at 24 years of age is someone that I'm hoping we're watching closely due to his potential.
I thought he used to be brilliant with the ball at his feet years ago and then all of sudden turned average at best even before pep came in and changed the game for keepers In PL
Just focusing on the shot-stopping stats (PSxG+/-/90 in this case), it really does tell a story. Starting with 17/18 (the 2nd place under Mourinho season and the last one in which De Gea was roundly praised):
17/18: +0.23
18/19: -0.04
19/20: +0.05
20/21: -0.01
21/22: +0.41
Puts this season's shot-stopping improvement thus far into context.
17/18: +0.23
18/19: -0.04
19/20: +0.05
20/21: -0.01
21/22: +0.41
Puts this season's shot-stopping improvement thus far into context.
His distribution looks good when I watch him play too. Seems very well rounded.
He was. When he first arrived one of the stand out things about him was that his distribution was so good. It's weird it's gotten so bad.
Yea, I mean its not just that other keepers have gotten better and surpassed him, he's actually gotten worse. So odd
He does look very rounded and also shares his birthday with Peter Schmeichel, which is a sign..
Those stats would have been even more damning prior to the Norwich game given I think one of those sweeps and at least two of the claims came in that game.
And we're supposed to pretend he's having a good season.
I mean against that you do have to weigh in the fact that he has been the best shot-stopper in the league this season, statisitically quite a distance ahead of everyone bar Sa.
He's a very one dimensional goalkeeper. But he's a one dimensional goalkeeper having a very good season in terms of personal form.
If a striker is very limited in general play, can't press and can only function as a poacher then it's fair to argue that he isn't a top class striker, that he doesn't suit top sides, that the team would improve with a different type of striker instead, etc. But if that poacher scores 20+ goals then you can still say he's had a good season, for what he is.
Last edited:
It depends what we're classing as personal form really. I appreciate he is making good saves at times but the sheer number of goals conceded this season that he could have done more on for various reasons is ridiculous.
For example, the last couple of months alone - please note I'm not saying other players aren't at fault as well because there is some horrific defending but I'd expect better from the goalkeeper in these situations.
Leicester: First goal, terrible pass out to Maguire. Second goal, literally kicks the ball at a Leicester player in the 6 yard box. Fourth goal, allows ball to go all the way across the goal 3 yards out.
Atalanta: First goal, ball into the 6 yard box with contact made 4 yards out centre of the goal. Second goal, ball into the 6 yard box with no real pace on it that is simply headed in.
Liverpool: First goal, gives Keita the entire goal to aim at by not rushing out and falling backwards. Second goal, ball along ground into 6 yard box. Third goal, some of the worst positioning you will ever see. Fifth goal, not quite as bad as the first goal but not far off. For reference in this game see the difference when Ronaldo & Fernandes went clean through on goal against Allison.
Atalanta: First goal, straight through him. Second goal, it's his ball all day to come and collect but he waits about half an hour for the striker to get there and poke it in.
City: Second goal, has to come and collect the cross and when he hasn't done that he cannot get beat at his near post.
Watford: First goal, more proactive goalkeeper comes and takes the original cross but the shot from King ends up almost straight at him & I think with his shotstopping form he'd be disappointed not to stop it. Second goal, positioning is poor and gives Sarr the whole goal to aim at. Third goal, straight through him.
Arsenal: First goal, playing dead facing away from the play.
Tweet
— Twitter API (@user) date
Best of the best! De Gea is agile and has fast reflexes. He has good feet for distribution.
I am one of DDGs biggest critics and I think you are being a tad harsh but I agree with the gist of what you are saying.
Off the top of my head, I can think of the two goals away at Atalanta, one of the goals against Watford, the Arsenal opener (you play to the whistle) and the 2nd goal against City that should have been saved. I fail to see how that represents a good season, especially given we have only played about 20 games and there are probably more errors I am forgetting or goals he could have done more to prevent by being proactive.
I am a big believer in the importance of Data and statistics, however, you have to be incredibly careful that you aren't blinded by it in the face of all other evidence, especially metrics such as PSxG and PSxG SOT, which are in their relative infancy (and nobody seems exactly clear on what Data is used).
The PSxG for the Norwich game awards De Gea +2. I don't care what the Data says about it, they were not difficult saves, especially the Pukki one. If that goes in, we're calling it a mistake. So why then is it a +1 PSxG if it's saved? The Kabak one...again, it looks spectacular because he throws himself for the cameras but that's never going in. Again, if a PL GK concedes a header from that distance that doesn't go anywhere near the corner and is at a nice height you are asking serious questions.
I am not dismissing the Data, I just think we need to understand it better and what it can/can't tell us.
Another poster made a good point earlier about how it accounted for a shot like Havertz's in the CL final...it trickled into the centre of the goal because Ederson had rushed out. Does that mean Ederson gets a big negative PSxG? I think the honest answer is, none of us really know.
Another example of a ball that he should be out claiming and not having to make a save. Similar to the Bernardo Silva goal, that ball shouldn't even have to be contested by the defender. All they should have to worry about is making sure his path to the ball isn't being blocked.
I made a point before about the Roma game last year, a lot of the saves he made that night were entirely preventable if he'd just come out and taken care of the initial ball into the box.
- Status