Keep, Sell, Loan: Goalkeepers

As we head into the summer, and off the back of the worst Chelsea domestic season of the past 28 years, it’s fair to say there are quite a few key decisions to be made by the new owners. Todd Boehly and his Clearlake comrades have certainly not been shy in stamping their mark on the club, with upheaval and change brought on in just about every department. One of their main strategies upon sealing their $4.25bn takeover of the club has been refreshing the on-pitch personnel, having spent an eye-catching £600m on new signings. This does come with its own challenges, of course, and one of those is having to part with several players in the transfer window to allow for balancing of the books, as well as making space in a (literally!) overcrowded dressing room.

Let’s start with one of the positions that is likely to see a new face added to the ranks this summer, the Goalkeepers. (Ed.note: stats accurate as of the end of April.)

---

CURRENT SITUATION

Chelsea have struggled to find a true number one goalkeeper since Thibaut Courtois packed his bags, and left for Real Madrid in the summer of 2018. Kepa Arrizabalaga was immediately brought in as his replacement, for a world-record fee for his position of £72m. Kepa started rather brightly at Stamford Bridge, adhering to then-coach Maurizio Sarri’s insistence on his number one acting as a ball-playing goalkeeper. His shining moments came in the form of being a penalty shoot-out hero a couple of times, once versus Tottenham in the League Cup semi-final and then making two stops from Eintracht Frankfurt to help Chelsea on their way to Europa League success. He did have some questionable moments, perhaps the biggest one being his refusal to be substituted for Willy Caballero in extra-time of the League Cup final with Manchester City, which Chelsea would go on to lose.

Regardless of his first-season success, Kepa eventually fell out-of-favour with Lampard, in the legendary midfielder’s first managerial stint with the club, and only recently has been able to bring himself back to being the first-choice goalkeeper he was always envisioned to be. This season has been more hit than miss for Kepa, and we’ll get into his numbers further below, but he’s still been inconsistent and at times, causing self-inflicted problems for a team that doesn’t need any extra issues on the pitch. This is why it’s very likely the club will at least scope any potential upgrades. However, as we’ll see as well in a bit, that might be easier said than done.

Perhaps surprisingly, it’s been a season to forget for Édouard Mendy. A year ago, the Senegal international had just come off an incredible period winning the Champions League, UEFA Super Cup, Club World Cup, African Cup of Nations, and was named Champions League Goalkeeper of the Season. However, his lapses of concentration when receiving any type of press from the opposition became an issue too large to ignore. Mistakes against Karim Benzema, Jarrod Bowen, and Rodrigo in an eight-month span were schoolboy-level, and Graham Potter opted for the less-out-of-form Kepa as his starting goalkeeper.

Mendy didn’t exactly have a great World Cup either, and has struggled with a couple of minor injuries which have kept him on the sidelines until recently — making one last token start under Lampard. It’s unlikely the newly-assembled recruitment team view him as the long-term solution to Chelsea’s goalkeeping woes.

---

BUYING OPTIONS

The issue is that, while neither Kepa nor Mendy provide full confidence they can be good enough to be a starting goalkeeper in a Chelsea team desperate to rebound from a disaster year, there really aren’t any obvious candidates to step in their place. From those abroad, Borussia Dortmund’s Gregor Kobel seems like a decent option, but has reportedly committed to staying at Signal Iduna Park until next summer at least. Mike Maignan of AC Milan has been touted as another possible option, however he is contracted until 2026, and it’s unlikely Milan would be willing to sell. That leads to scouring the domestic options available, but again, none really stick out. Looking at the potential options from a shot-stopping perspective, here’s what the advanced numbers, via OptaAnalyst, tell us.

There’s a bit to get through in the above metrics, but basically Sánchez, Meslier and Bazunu have conceded more than we would have been expected based on the quality of shots faced. The latter two have been abysmal, and are either reaching the level Kepa found himself in the aforementioned Lampard season, or are performing worse (Bazunu…).

There is a caveat that both have faced 100+ shots this season, Meslier faced the second most with 127, but if we’re looking at upgrades in the position, neither inspire much confidence. Sánchez has also been dropped in favour of Jason Steele, and with Potter gone from the helm, it’s likely any interest that may have existed, has faded away.

The shots per goal can be considered harsh to use against a keeper, given that it’s not necessarily their responsibility to prevent the opposition from gaining efforts on goal, and these could be in the form of one-on-one’s, tap-ins — i.e chances as a result of poor defending. But again, it is still hard to ignore that Bazunu, Meslier and Sánchez need only to face about 3 shots on goal before they are likely to concede. Perhaps these numbers improve as a result of being on a better team, but looking at Kepa’s numbers, it’s unlikely they would provide an upgrade worthy of their cost.

Alisson is included to represent the gold standard, while Raya and Martínez provide a positive impact to their teams.

It’s hard to ignore Kepa’s superb numbers here, as well. If the idea is to find a suitable replacement, then judging by this season’s stats, there might not be one available. Of course, there is a caveat which we’ll get to later, and that is this isn’t an expected season from Kepa, but he has posted the 2nd best score of the goals prevented metric, ranks 3rd in goals prevented rate and is 1st when it comes to the shots per goal counting. Kepa has had a really impressive season stats-wise, and performances away to Aston Villa & at home to Crystal Palace come to mind as those where he played best.

---

KEPA VS MENDY

If the question is ‘Who should be Chelsea’s starting goalkeeper next season’, and the answer is ‘Any of the above’, there might be a case for sticking with the current crop, and revisiting the position next summer. If that does happen, then either Kepa or Mendy will take the reins, but the choice is a difficult one as neither have been convincing enough over long periods to retain the status. There’s also a choice to be made between the two in terms of if a new goalkeeper is brought in, which of the two are likely to be made back-up? This likely comes down to whoever registers more interest from other clubs will be sold first, but assuming Laurence Stewart and Paul Winstanley (and others?) are given the responsibility, who should they keep amongst the ranks?

Kepa has had a stellar statistical year, but the intangible parts of his game are still concerning. His constant flapping at crosses and general inability to command his own area are major issues, and teams tend to exploit this. Chelsea are in the top half for conceding the highest amount of Set-piece xG in the league, and are tied with Crystal Palace for the proportion of Expected Goals Against which come from set-pieces, with 0.29 (only Man United are worse with 0.30). This means in general play Chelsea are generally solid in keeping the opposition from creating high-quality chances (6th in the league), but when it comes to set-pieces, we struggle, and have conceded the fifth-highest amount from these dead-ball situations. These numbers aren’t solely on one position, even the goalkeeper, but it does point to a clear weakness which Chelsea have, and one that Kepa himself has shown he’s not been able to master. For further reference, he’s tied-last with David de Gea when it comes to the amount of high-claims from a cross for the season with 12.

Kepa has made a name for himself for being at least adept at playing the part of a possession goalkeeper. This is largely true, as he does provide a calmer head than Mendy with the ball at his feet at times. According to stats from the Premier League on completed passes, Kepa is right around on-par with that of Alisson for completed passes per 90 minutes with 30 (Alisson has 32.3).

However, it’s not always smooth when he’s asked to participate in build-up play, which was evident from the recent outing against Wolves where he completed just 63.9% of his passes (3rd lowest of the season), and went 4/13 on long-balls. It’s fair to say that Wolves’ aggressive press, and Chelsea’s inability to deal with this led to Kepa not having much choice but to boot it long a few times, although it does show that his biggest strength might not be a good enough reason to keep him around.

Now let’s take a closer look at passing stats for Kepa, and the GK’s mentioned above (all stats from FBRef).

When it comes to completion percentage for launched passes (>40 yds), Alisson leads the way with Kepa ranking below all but two other goalkeepers (Sánchez & Meslier). What stands out here is that Kepa launches the ball the least of any goalkeeper excluding Alisson, but still only completes at a below average level for this group. Mendy has slightly better numbers here, but his low sample size makes it difficult to compare.

Depending on the playing style new manager Mauricio Pochettino wants to implement, perhaps even allowing for an attacker to come deeper to allow Kepa an easier outlet option, these numbers might improve but it still points to a key feature that Chelsea need to be aware of when approaching the position.

As mentioned, we don’t have any recent data on which to judge Mendy, and including last season’s numbers feels like a stretch for any legitimate comparison given the change in playing squad, style and manager. If we were to include it, his numbers of 43.0% completion on 18.9% of launched passes would have pushed him ahead of Alisson on efficiency. Mendy’s pass success of 82% last year only pips Kepa’s 81.4% for the year so far (accounting for Premier League only), while in the Champions League his passing is 10 percentage points better. Taking numbers from Understat for xGBuildup (which measures the xG of every possession the player is involved in) Mendy (5.09 for 21-22) was 33% more effective than Kepa (3.43).

Kepa’s passing completion splits for short, medium and long balls are 99.1%, 99.3%, 54.7%, respectively. That latter number is good enough for third in the league, behind Alisson (59.7%) & Ederson (62.1%). Meanwhile, last season Mendy also ranked third in the same category, with a score of 54.5%. Alisson (60.9%) and Ederson (67.1%) were slightly better a year ago with their long-ball passes as well.

The argument of Kepa being a better ball-playing goalkeeper than Mendy has always felt like a weak one, and while some stats may point to one above the other, there’s not that much difference that a perceived strength for the former should be enough to keep him in the team ahead of the latter. The main concern with Mendy are the massive blunders and failures to launch when under severe pressure from the opposition, as they tend to directly lead to a goal the majority of the time.

If the difference between both in possession is minimal, then what about their shot-stopping? There’s really not much of a difference maker here, either. Taking Kepa’s score of PSxG(post-shot expected goal) net difference score in the most recent 4 seasons where he was counted as a full-fledged starter (2017-18, 18-19, 19-20, 22-23), his average comes at -1.47. Given that the 19-20 score is about as bad as a goalkeeper can have (-11.3), even halving this for a “normal ‘’ bad season still points to a goalkeeper who will struggle to reach the heights of an elite shot-stopper that Chelsea need.

As for Mendy, his numbers are based on the 2018-19, 19-20, 21-22, 22-23 seasons. There’s less variance in his performances, and his average lies at +0.97. Here’s a comparison breakdown of both players’ seasons.

The issue with both is that you’re really not sure what to expect from them in a given run of appearances, and for a team receiving vast funding in multiple positions, it’s quite a risk to leave the goalkeeper performance up to chance.

Mendy provides more stability and guarantee of performance. He’s either going to be just about good enough to make a positive impact or just about bad enough to cause harm to the team. With Kepa, his yearly performances so far have been very extreme. There was an argument for him to be Chelsea’s Player of the Season given the poor performances and injuries to others (ed.note: Thiago Silva rightly won that award), but it’s hard to look past his previous starting stint, when he was comfortably the worst goalkeeper in the league if not across the top five leagues (and at the same level as the starting goalkeeper for 20th-place Southampton this year). It’s far too volatile and unpredictable.

---

VERDICT

It’s a toss-up between keeping either Kepa or Mendy based on the assumption that a goalkeeper will be brought in to, at the very least, challenge for the number one position. It’s safe to assume the other will leave, likely in a loan deal or if some team spots value in the Kepa revival and/or sees potential in Mendy’s two-year run at the start of the decade, then there’s a chance of a healthy amount of transfer revenue being brought in.

Rounding off the goalkeepers squad list will be Marcus Bettinelli, who signed a contract extension to continue playing the part of third-choice goalkeeper, while a loan deal might be the preferable option for American sensation Gabriel Slonina, signed for £12m from Chicago Fire last summer.

Keep: One of Kepa/Mendy, Bettinelli
Sell: One of Kepa/Mendy
Loan: Slonina
Buy?: Yes