Fans and analysts who have followed United’s issues in the past few years and are also aware of the principles of play Ten Hag likes to implement, would have come across one concern immediately – David de Gea.
The Spanish goalkeeper has proven most of what any keen observer would have told you when Ten Hag was announced as Manchester United manager – De Gea is not fit for the Dutchman’s goalkeeper needs. The difficulties in build up is one key issue, while the inability to proactively sweep and claim crosses only further multiplies the adverse effect the Spaniard has on team tactics. After two games of seeing de Gea fumble, stay on his line, not attack crosses, hesitate to sweep and make mistakes in distribution, Ten Hag employed a kick-it-long system which bypassed the need to build up, and relied on his defence dropping deep to limit the cons of de Gea’s poor sweeping and claiming.
But in recent games, especially post the World Cup, the team has been slowly building towards a more high line high press game with focus on retaining the ball and de Gea has just not looked like a good cog in that system. The game against Arsenal was the most clear example, when both teams were implementing a similar 3-man high press, which meant a fullback was always free during build up. Ramsdale always nailed the cross ball to his fullback and helped Arsenal beat the United press, but on the other end, de Gea routinely failed to find his fullbacks and put them under pressure when he did.
There are many more examples of de Gea not being it, but I’m going to move past that and focus on today’s exercise – finding a starter level goalkeeper for Manchester United who suits Ten Hag’s tactics.
The Stats considered
Before diving into the process, let me explain the stats that have been considered and the logic to why they are good or bad for our assessment.
1) Save %
A high save % is good, since we want our goalkeeper to be able to make as many saves as he can, with respect to the shots he faces. A caveat here is that the save % for a keeper conceding too many shots usually drops, while it’s easier to maintain a higher save % when you face fewer shots. One example would be Illan Meslier, who faced the most shots in the league in 21/22, thanks to the chaos in defence before Marcelo Bielsa’s sacking.
2) PSxG Differential:
PSxG or Post-Shot Expected Goals is a measure of the quality of shots faced (xG faced by GK one can say). The difference of PSxG and Goals allowed is called PSxG differential or PSxG +/-. A high positive value for this denotes high quality saves. David de Gea is a prime example of (usually) scoring well on this stat, thanks to his dramatic saves from positions most keepers usually aren’t able to. I’m giving this the highest weightage, since it’s critical for a GK.
3) Cross Stop %:
This is calculated by dividing crosses stopped by crosses faced. A high value denotes good cross claiming ability. Again, the caveat here is that a keeper bombarded with crosses might find it tougher to maintain a high Cross Stop % compared to one that isn’t.
4) OPA or Defensive Actions outside Penalty Area:
A measure of sweeping, a high value on this denotes the ability of the keeper to perform actions outside his area.
5) OPA Avg Distance:
The average distance from goal of all defensive actions outside the penalty area. Similar to the above stat, a higher value is good, for the keeper that we need.
6) Launched Passes Completion %:
When a keeper decides to launch a ball from open play, does he complete the pass accurately or is it a random boot-out under pressure that goes to an opposition? This stat answers that question. It separates purposeful long kicks to a teammate from random hoofs under pressure. One caveat here is that having a great target man could result in many ambitious long kicks getting a completion status as well. One example is Samir Handanovic topping this chart two years ago, thanks to Edin Dzeko being the target of most of his long passes.
7) Passes Attempted:
For the GK we are looking for, we want someone who isn’t afraid to get on the ball, during open play. A higher value for this is better for us.
8) Pass Launch %:
This denotes the percentage of passes a keeper decides to launch, instead of playing on the ground. For our ideal keeper needs, we want a lower value for this – a keeper who has the technique, press-resistance and composure to avoid booting the ball long and instead tries to find a player close by.
9) Pass Avg Length:
Quite self-explanatory, the preference is for a keeper who has a lower average pass length, which indicates his priority to find a player closer via a shorter pass, rather than kicking it long.
10) Goal Kick Launch %:
The goal kick version of ‘Pass Launch %’. Again, we want a keeper who executes a short pass from a goal kick more often than not, which is the ideal Ten Hag goal kick routine.
11) Goal Kick Avg Length:
The goal kick version of ‘Pass Avg Length’. We want someone who has a low value on this.
NOTE: Before we move on, I must state one important caveat. A few of these stats are highly dependent on team tactics. A goalkeeper may not be playing short, because the manager has instructed him to kick it long (like ETH did with de Gea after the first 2 losses) or a GK may not be able to score well on sweeping actions, because his team doesn’t play a line high enough for him to get the opportunity to sweep. So, for the sake of clarity, it might be best to take the results of this ranking as to not reflect goalkeeper ability, but rather reflect goalkeeper suitability to United’s needs. In short, the score won’t measure absolute talent, but will identify goalkeepers who are doing well at the the role that we want this season. There could be other talented keepers who just aren’t playing enough or are being shackled by team tactics. With that clear, let’s move on.
The Ranking process
The rest of the process is largely mechanical. A quick run-down of the steps:
1. I downloaded all goalkeeping stats from fbref.com for Europe’s top 7 leagues for the 22/23 season and picked out the above 11 stats of interest to this assessment
2. I filtered out all goalkeepers who didn’t even manage to play 5 90s in 22/23. This left us with 160 goalkeepers, who make up the data set on which we will be doing further calculations
3. I converted the 11 stats into per 90 versions and then percentile versions. This means that each stat was divided by the number of 90s played by that goalkeeper, and then converted into a percentile stat. E.g. David de Gea with a 40 percentile on ‘PSxG Differential’ indicates that he’s better than 40% of the keepers in our data set for this particular stat. I’ve also accounted for the reversed percentiles on the stats for which a lower value is desirable. E.g. Low values of ‘Pass Avg Length’ were ranked with a higher percentile. In summary, all high percentiles now mean ‘better’, for this assessment
4. Finally, I created a weighted formula using these 11 stats using the above weights. I then factored in league difficulty using UEFA league coefficients for 22/23 giving it a 15% weightage. The last step was to create a percentile version of this weighted score, to ensure our top ranked goalkeeper scores a 100% and the worst one scores a 1%
The top goalkeepers, based on this, are as follows:
Manuel Neuer scores a majestic 100% on our metric, closely followed by Kepa, who’s having a resurgent season and Alisson, who still stands tall as Liverpool’s best player this year. This could be considered as a fair assessment of sweeper keeper performances in Europe’s top 7 leagues in 22/23 and Neuer is exactly the kind of keeper we need.
The reason we’re here is to find a suitable young/peak keeper who is worthy of starting for United in the short and long term. I use age to filter out all goalkeepers above the age of 29 and add a few important contract details and transfer rumors. I also eliminated some GKs who barely crossed the 5 90s played threshold and some unrealistic options like Donnarumma, who surely won’t move from PSG.
The results are as follows:
Gregor Kobel has been having a great season at Dortmund for the second year running. Even last year, his stats were pretty good. United should really be enquiring if he’s available. There are no rumors of him wanting to leave or Dortmund entertaining such a thought, so maybe he’s well settled there.
There are similarly no rumors for Diouf and Vaessen in the market, but it might be wise to keep a track of them given their great stats while having a low profile.
The real targets who have been rumored to move and seem to be available for the right price are Bijlow, Raya, Costa and Meret. I’ll be doing mini reports on the first 3 below.
Mini Scout Reports
Before we dive into the reports of the targets, I want to highlight the holy grail – the profile that we are going for. Here’s Manuel Neuer’s pizza chart using the same 11 stats and assessment detailed above. Remember, the percentiles are adjusted based on what I mentioned (Eg. High percentile for Avg Pass Length means he’s passing short and so on)
What’s immediately noticeable is that Neuer scores well on almost every metric except cross stopping. He’s stopped only 4 crosses from 103 this year, which seems to be surprisingly low. But keeping that aside, it’s as good as it gets. Playing for a league leader team that maintains a high line and engages in short build up, does help the stats, of course, but Neuer has shown what a modern sweeper keeper is all about.
On the other end of the spectrum, let’s take a look at Manchester United’s current starter – David de Gea.
David de Gea scored 16.4% in the above exercise. Only 26 goalkeepers in Europe’s top 7 leagues scored lower than him and only 1 Premier League keeper (Mark Travers, Bournemouth’s backup GK) scored lower than him. It’s safe to say that de Gea is among the worst keepers in the league. What’s even more shocking is that, even keeping aside stats that might be dependent on team tactics, he’s below average for both shot-stopping metrics. He’s 33 percentile on PSxG +/-, which is apparently his strongest stat. The less said about the claiming, the better.
David de Gea is probably the weakest starting member in the United XI currently. Our aim should be to buy someone who represents a massive upgrade on de Gea in the short-term and has the potential to get close to that Neuer pizza chart in the long-term.
Let’s dive into the 3 targets that I recommend the most for this.
Justin Bijlow, 25, Feyenoord
Career History: Bijlow’s rise has been a fairytale one. Representing Feyenoord at all youth levels, he got his break at the end of the 17/18 season at the young age of 19. In the following season, he played a huge part, starting 20 times for the club. With 97 starts across 5 years since his debut, Bijlow is one of Feyenoord’s key players and has already become a household name in the Eredivisie. In 2021, he was called up to the national team by van Gaal and was doing well in World Cup qualification. An unfortunate injury that kept him out for the games in the run-up to the World Cup, ended up with him not being selected for the Finals in Qatar.
Pros: Bijlow is a very well-rounded profile. His eye-catching and strong saves are immediately noticeable, while spending more time on his games gives a clear indication of a good distributor who likes to be involved, and a good claimer who is authoritative in the air. In general, the two words that come to mind when seeing Bijlow are – aggressive and athletic. He almost feels like an energetic midfielder, with the physical ability and determination in most of his actions being a large driver to his good stats.
Cons: The major con with Bijlow is – his fitness. He’s missed almost 50 games for Feyenoord over the years with various injuries related to foot, knee and toe. It cost him a World Cup Finals spot, and has left Feyenoord missing their key starter in important games as well. 22/23 has been good so far, with no injuries yet, and one can hope that his fitness improves in his peak. Also, his aggression leads to the odd error in passing or claiming, thanks to an overzealous or over-proactive action, but I would say that in the long run, it’s still much better than having a keeper who’s rooted to the line or scared of picking the right pass.
Verdict: I think Bijlow has everything in his locker for a modern GK, and at the age of 25, is already well-rounded and performing at a good level. These pros far outweigh the slight cons of fitness and rashness, both of which should be fixable as he matures. Should be a strong consideration.
Devil’s DNA Score 8/10
David Raya, 28, Brentford
Career History: Raya started out as a youth player for Blackburn Rovers and after a few years of bit-part appearances for the side, at the age of 21, he broke into the first XI and enjoyed 2 years as an undisputed starter for Rovers. His performances earned him a pick up from the smart scouting team at Brentford in 2019. It’s only been up and up since that move, with great performances in the Championship, only followed by great performances in the Premier League. In 2022, his rise earned him 2 caps to Spain ahead of De Gea and that change should probably happen at club level too.
Pros: At age 27, Raya is probably the closest to peak in terms of development, among our shortlist options. And that shows in his game. He is a very consistent shot-stopper, excellent cross claimer and strong sweeper. In terms of overall goalkeeping ability, he has no weakness. Even his distribution is a strong trait. The main reason why his pizza chart scores low on 4 distribution stats is largely due to Brentford adopting a long ball strategy this year. They often go direct to Ivan Toney in the final third thanks to the striker’s excellent hold up play (United should look to him for CF too). The 2 distribution stats which are more in his control score well. Back when he was in the Championship and Brentford dominated possession, Raya would play very short and often boss the build up. So, if you disregard the 4 stats he ranks low on, as issues that won’t matter in a top team playing a short build up style, Raya’s overall ability probably ends up being the best on the shortlist.
Cons: It’s very hard to pick any cons for Raya. As explained before, his distribution is simply not a negative. He’s excellent at it. If I had to nitpick, the only con I could pick from watching a lot of Raya is that he sometimes tries too many ambitious counter-starters with low-percentage throws after claiming, which lead to the opponent getting the ball, but they also help start some great counters. It’s a minor thing.
Verdict: No matter how hard I look, I cannot see any reason why Raya shouldn’t be a top target. He has all the traits, is already showing them at a high level in the league and will only look better in a top team. His price also seems affordable since he has just 1.5 years left in his contract. At the rumored price, this is a no-brainer. If United don’t seal this, I’m positive that Spurs or Chelsea will pick Raya up this summer. For the first time since I’ve been doing scout reports for United, I’m giving someone a 10.
Devil’s DNA Score: 10/10
Diogo Costa, 23, Porto
Career History: Having just turned 23 last September, it’s fascinating to realize that Costa has already started 72 times for Porto in 2 years as a key starter and got 12 national call-ups for Portugal which includes starting every game in World Cup 2022. It speaks of his monumental talent as a player. 4 years with Porto B leading to 2 years at Porto and reaching the point of every top team wanting you, is as good a career goes till this age.
Pros: Aiming to be a well-rounded sweeper keeper, Costa is an excellent distributor and claimer and a decent shot-stopper. Standing tall at 6’4”, he has excellent aerial reach and the agility to go with it. He is really good at beating the opponent press by playing smart passes to the fullbacks or midfielders, effortlessly starting attacks for his team, no matter what the situation. He is also ever-ready to come off his line and sweep or claim with good control, without being too overzealous or too rooted on his line either.
Cons: There are no obvious weaknesses to Costa, except the fact that due to his young age, he still feels raw. There is the odd mistake in him, especially when it comes to saving. His positioning might need some extra maturing to bump up his save %. He has also made some high profile errors in Champions League and World Cup games, which could suggest a slight requirement of nerves to handle big games.
Verdict: In terms of profile, Costa is a legitimate wonderkid who has all the traits, but his rawness means that he isn’t the most starter-ready on the list. The ideal top team move probably comes 1-2 years early for the young GK and with a huge Release Clause of £65m, does a summer 2023 move become a case of “too expensive for too raw a player”? I’m inclined to think so. Don’t get me wrong, I think he’s a fabulous talent and would still take him if there isn’t anyone in the top 5 shortlist names willing to come. But the price and rawness make me feel that there are a few, if not many, better options out there that would help us save some money for other targets like a striker or midfielder.
Devil’s DNA Score: 7/10
In summary of this exercise, for the summer transfer window of 2023, I feel Manchester United should pull out all the stops to sign David Raya from Brentford. He seems ready to move with an appealing price and his profile only suggests that he can become one of Europe’s best. If that doesn’t work out or someone else snaps him up, United should enquire about Gregor Kobel, who has also been having a fabulous 1.5 seasons at Dortmund. After those 2, a move for Justin Bijlow of Feyenoord has to be entertained, given the Dutch GK’s skillset and potential. As a final throw of the dice, an expensive move for Diogo Costa that seals the GK role for many years, is a good option as well.
Thank you for reading this far. I hope you enjoyed this one. Let me know your thoughts on the options and if I’ve missed any, on my Twitter handle. I will be doing more scout reports in the near future, with a Casemiro partner playmaker carrier style midfielder next on my to-do list. Cheers.
(All stats from fbref via Opta)