Is David de Gea really back to his best?
David de Gea has started every game for Manchester United in the Premier League this season. He's saved a lot of shots. But does that mean he's back to his best? Unfortunately, it depends whose statistics you believe.
These days, post-shot expected goals is one of the more popular advanced metrics to measure shotstopping ability. Post-shot xG measures the chance of scoring after a shot has been struck, once its trajectory and speed are known (as well as, in some cases, the position of the GK). We don't use it, though, and our FAQ includes a lengthy explanation of our decision. In a nutshell, we think that post-shot xG makes a GK look good when he or she takes the wrong approach to a shot and then miraculously saves it. Do this enough times, and post-shot xG will make you look like a phenomenon. De Gea's 2021-22 campaign may be a case in point.
We don't calculate post-shot xG around here, so we'll use the handy stats from Football Reference to prove the point. FBRef says De Gea has faced 39.1 post-shot xG, including four penalty shots, of which he saved three.* But let's put the penalties aside, since most differences there between regular xG and post-shot xG are probably down to luck. We'll give each penalty 0.75 xG, for a total of 3 xG, and say De Gea faced 36.1 post-shot xG in normal play.
What about regular xG? FBRef has Manchester United conceding 34.6 xG. If we subtract 3 xG again for the penalties, FBRef would say De Gea has faced 31.6 xG this season. That's a difference versus post-shot xG of 4.5, or 14% of the total regular xG.
Now, this is already unusual. Post-shot xG is telling us that the shots De Gea faced were on average 14% more dangerous than generic Premier League shots from the same locations and situations. Somehow, over close to 300 shots, De Gea faced considerably tougher attempts at goal, even adjusting for everything we knew about before the ball left the boot (or the head). And that included the positions of De Gea's defenders and other attacking players.
Yet somehow, De Gea saved a lot of these shots. Not including the penalty he conceded and Manchester United's two own goals, De Gea has allowed 29 goals this season. That's a lot fewer than 36.1 post-shot xG. So here's De Gea, consistently facing more dangerous shots and consistently making amazing saves. He sounds... like a phenomenon.
But wait. What if he failed to come off his line to cut off the strikers' angles? What if he never challenged an oncoming player?** Then he'd probably face higher-quality scoring chances than the average GK, and post-shot xG would be higher than regular xG. And what if De Gea, with his catlike reflexes and acrobatic ability, still managed to save a lot of those shots? Then yes, he'd seem like a phenomenon, but only because he made his job more difficult than it really was. He'd be a great shotstopper, perhaps, but a poor shot-approacher.
And that is exactly what De Gea looks like in the FBRef statistics. If you compare his 29 GA against 36.1 post-shot xG, he looks brilliant. If you compare his 29 GA to 31.6 xG, he's fine but no worldbeater.
We tend to believe the latter is true. To rate GKs' shotstopping, we handicap shots by xG and then adjust for the finishing skill of the strikers, which we rate simultaneously using an Elo-like framework. Essentially, we roll shotstopping and shot-approaching together into one overall shotstopping rating. Here's how we tracked the evolution of De Gea's shotstopping skill for non-headers in open play, the biggest category of shots:
Our data start on 1 January 2016, at which point our assumption is that all GKs are average. Our algorithm kept adjusting De Gea's rating upwards through the first few seasons, because he kept saving more shots than we expected. But the rating peaked in 2017-18 (at an extraordinarily high level!) and has come down gradually since then.
De Gea's other shotstopping ratings are nothing to write home about, except perhaps for his performance against headers from dead balls. And he's one of the weakest defenders in the top five European leagues:
You can see from De Gea's smartermap that he rarely leaves his area to challenge oncoming strikers. Just compare his touches with those of notorious sweeper-keeper and ball-player Manuel Neuer at peak aggressiveness:
It used to be that if post-shot xG was much higher than xG, you might conclude that a club’s defenders were poor. The shots faced by the GK were more dangerous than average, because there wasn’t enough pressure on the strikers. These days, though, a huge amount of information about defenders’ actions and positioning is already included in regular xG figures. So if post-shot xG is much higher than xG, it can’t really be the defenders’ fault anymore – it has to be either the skill of the strikers or the shortcomings of the shotstopper. And when a GK has already played more than half of a Premier League season, it’s unlikely that there would be such a big discrepancy in the quality of strikers he’s faced relative to the whole pool.
So here's what it comes down to. If you believe that De Gea really has faced much more dangerous shots for the entire campaign through no fault of his own – or of his defenders – then yes, after three seasons of so-so performances he's suddenly having his best shotstopping spell since 2017-18. If you don't believe this, then you can leave the post-shot xG stats in the circular file.
To be sure, the evidence presented here is circumstantial. You might have to put your eyes in front of some video and watch De Gea for a few hours to make up your mind. If you've already done that, then what do you think?
* Only three of these penalty shots actually counted, since one by Ismaila Sarr had to be retaken. Kiko Femenia scored the rebound from the retaken shot, but that didn't count either.
** An earlier version of this article suggested that post-shot xG might be higher for De Gea if the underlying model explicitly took his positioning into account. Though some models do so, the one used by FBRef apparently does not. This doesn't change the overall argument that De Gea might face higher-quality chances because of his actions or lack thereof.
Would you like to have your player featured in the playerfocus on our homepage, for thousands of users and football professionals to see? Just click the Contact link at the bottom of this page to email us your request.
[Photo: Анна Нэсси]
