In search of a new midfielder

Introducing D.U.E.L.S., a meaningless new stat to help us sift through 80 defensive midfield targets and find the best fits for Arsenal

On the midfield front, the subtext is starting to turn into text. Asked about the health of Mohamed Elneny today, Arteta offered this:

"There are concerns, especially (because) Mo is a player who never complains about anything. He had some issues with his knee that we are assessing at the moment and we will have to see how he evolves."

Next, he turned to the topic of the transfer window:

"Well, we need some more cover in midfield ideally if we can. In this market it’s pretty complicated to do that. If there’s anything else available that can make us better we’ll look at it."

Arteta likes to keep his cards close to the vest as a general rule. We’ve all gotten accustomed to the two qualities that tend to characterize his updates on injuries and personnel: a vague non-update, and a knowing grin.

The specificity of these answers speaks to the real possibility that Elneny is out for a while and that Arsenal will seek to address the situation in the next week. There are headwinds—namely, ridiculous asking prices, a tighter January market, and a stubborn desire not to hamstring the long-term vision—but in the event they try to get something done now, I thought we’d waste some more time together on the question.

What route will Arsenal take?

Let’s rank the possible scenarios that Arsenal may pursue, starting with the most ideal (but least likely) and working our way down:

1. High preference, low likelihood: Find the perfect, high-level, long-term fit for the team (Rice, Enzo, Caicedo etc) at a cost you can live with — now

2. High preference, mid likelihood: Find a long-term Elneny upgrade with trajectory

3. Mid preference, mid likelihood: Find a savvy short-term solution via loan or expiring

4. Low preference, mid likelihood: Roll with current situation and 🤞

From there, here are the two scenarios I wouldn’t forecast Arsenal pursuing:

1. Dramatically overpay for players who fit in category 1

2. Make a punt on a castoff who doesn’t ultimately fit the “non-negotiables” of the team

What kind of players might be targeted?

The summer conjecture and reporting was largely around the possible additions of Youri Tielemans and Douglas Luiz. While they are different players, they both offer a flexible skillset as technical, “hybrid” playmakers that could potentially offer cover across the midfield, albeit to varying levels of aptitude (Luiz is probably a more natural fit as the deepest pivot). Basically: Arsenal was connected with talented, flexible midfield depth.

Much has changed since then, however. We’ve learned about the new responsibilities of Xhaka’s role in attack, making the gulf in responsibilities of a 6 and a left-8 (“fake 10”) wider than ever. This tweak, and the subsequent transfer business, means that this position is suddenly flush with talent: Xhaka, Vieira, ESR, Trossard, Ødegaard, and Sambi could all play there.

Meanwhile, things are even thinner on the 6 front. Behind Partey, Elneny has been injured for most of the year, and may be facing down a longer road to recovery. Sambi got a run to show his potential, and gradually revealed that he is not currently the answer for a team in a title run. The most pivot-al position on the squad is hanging by a thread.

For that reason, the ideal player must fit the overall modus operandi of finding true technical level-raisers wherever they are, while answering the pressing immediate question: who the hell is gonna play the 6 if Partey goes down?

Introducing D.U.E.L.S.

Those familiar around these parts know that we’re partial to inventing dubious stats to help make sense of transfer targets. Our last edition was Sakanelli Score, which ranked wingers amongst a host of factors and gradually placed Trossard at #12. He climbed up to #5 in a non-age-adjusted model.

Today, we introduce you to D.U.E.L.S., which stands for “Depth Upgrades for the Elusive Lone Six,” an abbreviation I came up with long before writing this current sentence, I swear.

“But duels are already a stat,” I hear you saying. Frankly, that's on you for expecting this all to start making sense.

Why focus on duels? It’s clear that Areta prizes them, as he made very clear in that famous All or Nothing scene. On the initial pull of 80 targets, two players rose to the top in terms of total percentage of duels won: Martin Zubimendi and Declan Rice. Both are rumored transfer targets.

Arsenal’s 96 recoveries against Manchester United

So with that, we set out on creating a quick catch-all stat to rank targets. Here’s loosely how it was done:

• Sample: I pulled a list of 80 midfield targets worldwide, largely focusing on target age range, positioning, and potential. The first thing we need to accept is this: there are few true “lone 6’s” playing at high levels of world football, and a player does not have to be an exact Partey clone to be successful on their own terms. Just as Arsenal did when they identified Partey, we’ll largely have to look at attributes and then make an estimate of potential. Saying “x player currently plays in a double-pivot” or “x player is more box-to-box right now” is insufficient to write them off. As an example: White didn’t play right-back before, but he was correctly judged on his underlying attributes.

  • A note that this sample was manually run, so I inevitably forgot to include a few people (and they will inevitably be who is signed).

• Quantitative & Qualitative: From there, I grabbed a suite of metrics that should point towards aptitude as a defensive midfielder. This is primarily focused on the two things Partey does best: defending and passing. This includes stuff like total duels, defensive duels, aerial work, pressure-adjusted interceptions, passing volume, passing %, progressive passing, long passing, and dribbling. I added some other factors based on current positioning, potential, and schematic fit.

• Creativity & Technicality: From there, I flowed in attributes from SofaScore around creativity and technicality to help properly weight skilled technicians (think, Zubimendi) against total destroyers (think, Ndidi). Each of these attributes combines dozens of stats.

• Game Score Baseline: Next, I gathered FotMob scores for the last calendar year, and league-adjusted them to give a general baseline of performance.

• Age: Based on their percentiles within the sample, players get bumped up if they achieve production at a younger age, and vice versa.

• Value: Based on inflated Transfermarkt fees, a player’s total ranking is adjusted based on their production relative to potential cost. This is less weighted than usual, because after the Mudryk saga (and the rumored Rice pursuit for the summer), I’m not sure what the budget is. More than a set number, this search seems to be about finding the best player while not dramatically overpaying.

In the winger list, I removed some of the dream options (Leao, Kvara, etc) just because they didn’t seem like realistic targets at this juncture. For this list, I decided to keep some of those in, however unlikely, because I’m not sure anyone can say with authority how things are going to shake out over the next week. Let’s just try and keep our expectations in check, shall we? 🥴

To the rankings!

Top 30 Defensive Midfield Targets Based on D.U.E.L.S.

Next, here are my assorted ramblings on some of the players included here. I have varying degrees of visibility into those on this list, so read this as “assorted impressions” rather than “definitive scouting reports,” please. I probably will get a few things wrong:

• Declan Rice simply housed everyone, showing elite attributes across the board. He led almost all of the true destroyers in underlying defensive metrics while keeping pace with many of the best progressors as a passer. There may be questions about whether he’s been well-drilled enough in tight-control modern build-up in the defensive third, but there are few players with such overall potential as Rice. Not in this sample—period.

• Likewise, Eduardo Camavinga and Enzo Fernández understandably brought up the second and third spots. There are various factors that make them highly unlikely, but I don’t think I’m going out on a limb to say they’d both be dream signings for this current Arsenal squad. I recently said that “my toxic trait is thinking Enzo is worth his release clause,” and one look down his fb-ref page will show why I’d say something silly like that. He may not have the baseline defensive peak of others (he’s still good, IMO) but is so superlative elsewhere that it doesn’t matter. A few months ago I did a little amateur scouting run through his old tape to decide if he could project as a Rodri, a Kroos, or something different. My conclusion was basically: yes.

• From there, Moisés Caicedo headlines the zone of maybe/kinda/probably-not possible. His stats are all great for his age, but signing him is ultimately all about eye test and projection. He is similar to other young stars in that he’s good already, his floor is super high, and he just turned 21. He is different in his path: this time last year, he was coming from Beerschot. This time the year before, he was playing for Independiente del Valle in Ecuador. Depending on how you interpret all that, he could max out as a really good flexible midfielder, or an all-timer.

• Martín Zubimendi tops the sample in overall duels won with 59.88% over the last year. All his efficiency stats are through the roof, and combines all the close technicality and passing you’d expect from a potential Busquets regen, adding a strong defensive presence. He hasn’t been quite the hub of possession some others have, however.

• Ibrahim Sangaré is huge and skilled, and some of his defensive stats are hilarious (he has 3.39 successful tackles a game and wins 75% of his defensive duels). His stats are all enormous, which helped propel him to the #1 spot when I ran a search in the summer (marginally ahead of Caicedo). So it’s just two questions from here: how well does he project from the Eredivisie? With his bulkiness, would he get hassled on close-control stuff?

• On Borussia M'gladbach, Kouadio Koné is probably my favorite young candidate as a Partey successor who hasn’t already been mentioned. He pops off the screen: tackling and running like a top enforcer, and then dribbling and sprinting like a top winger. His passing range is full of talent but not a finished product, and he’s still figuring out consistency as he’s playing a little bit more of a box-to-box DM role. But it’s basically my ideal profile of a younger DM: a player with the ball skills of an 8/10 who loves running into every challenge and shoving people around, and is good at it. Here’s a comp for you to become a believer like me.

• I would put Roméo Lavia and Khéphren Thuram in a similar tier, though what Lavia is currently doing in the Premier League at 19 has been a sight to see—making up for any of his in-progress attributes with pure effort. City has a buy-back on him. Some of Thuram’s underlying analytics don’t pop yet, but the tape shows a really nice player.

• I don’t see much wrong with Rúben Neves other than A) his likely price and B) that he can rely too heavily on longer, lower-efficiency passes and shots.

• If you were to pointedly ask me who may have the best blend of readiness, potential, and hopefully non-ridiculous valuation, it would be Enzo’s batterymate Florentino Luís. He shows the importance of situation on a player. While with Getafe last year, for instance, he had 18 short passes a game for 81.6% completion in a really negative setup. Now with Benfica, he’s completing 40 short passes a game at 94.9%, and 83.7 passes/90 overall. He projects as well as anyone in the world as a defensive 6, his passing now looks really tidy, and he’s occasionally outshined Enzo in terms of consistency and overall impact.

• I was reminded that Manuel Locatelli was the point guard for those De Zerbi-run Sassuolo clubs, and it all started to make more sense.

• Amadou Onana strikes me as a good investment in the €25-€30m range, but after that, he’s a really difficult player to project in possession. If he’s viewed as an Elneny replacement with an interesting trajectory, there’s plenty to like. If he’s seen (and valued) as more than that, I have some reservations.

• There are five DM’s from South America who made the top 30, and they’re all interesting for their own reasons: André (Fluminese), Aníbal Moreno (Racing Club), Alan Varela (Boca Juniors), Allan (Atlético Mineiro), and João Gomes (Flamengo). I suspect they may all be readier to roll than one may expect (lack of experience in a top 5 league does not immediately mean “raw”). Off-hand, I’m particularly intrigued by Alan Varela.

• Anton Stach (Mainz 05) is the only Anton Stach in the world. He’s a huge, two-footed deeper midfielder who was recently a striker and still plays on the wings. He’s nowhere near the necessary level as a passer right now, but the talent looks there, he’s finding plenty of ways to be productive, and he’s a fun watch. His later start at the position may portend rapid development, I don’t know.

• I’m also fascinated by Salvatore Esposito (Spezia), who looks like he spent too long dominating Serie B. His quick turns, immediate passes, and defensive work rate all project really well. Very early days, however.

• Ibrahima Bamba is somebody I had my eye on this summer. He’s big and talented and loves challenges. I haven’t watched much since, but he seems to have settled in as the middle CB in back 3/5 instead of DM. He’s still getting forward a bit, but I’m not sure the full story here, or how it impacts the long-term projection.

And finally… 🚨 Speed round! 🚨

• Camara looks to be having a nice first year in Monaco… Caqueret offers a lot to like, but his duel % is low in France … Agoumé is raw but highly talented at the two biggies: tackling and passing … Tielemans is an underrated defender and fares pretty well in this ranking geared towards 6’s, all considered. That said, he’s best used further up … In terms of an overall, jack-of-all-trades level-raiser, Koopmeiners offers many of the same things that made Douglas Luiz so attractive. He might seek more of a locked-in role than Arsenal could offer, however. Merino is in this category, as well … Álvarez does not strike me as an Arteta player. Not technical or progressive enough … Ugarte is one of the ones who surprised me for being so low. He tackles a lot and he passes a lot. He didn’t score well on creativity … Santi Comesaña would be a nice depth signing. He’s big and smart, but probably not at the same trajectory of others … Kenneth Taylor probably doesn’t project as a 6, but is doing so well in his first year at Ajax that he showed up at #31 anyway.

Other players:

• To answer your next question: Rodri would place first, Partey would place second, and Casemiro would be in the twenties because of age and valuation. Elneny was in the 60’s.

• I didn’t include him in this sample because he probably projects further up, but Nicolò Rovella is looking really good at Monza (on loan from Juventus), is pretty big and secure technically, and can probably find a place almost anywhere in a midfield.

• Update (26 Jan): Ran some calculations with more players: Milinković-Savić is after Locatelli as a 6 (he ranks around 12-15). His dangerous passing, creativity, and technicality is about the highest in the sample. He's older, doesn't have the highest duel % (except aerially), and probably expensive. He'd fare even better in a ranking for 8's ... Tyler Adams is around Álvarez in 29th. He's done really well in his first PL and is at the right level of club for him to be a locked-in starter, as he has some passing limitations. On an Arsenal side, he'd be more of a max effort, crowd favorite, potential Elneny upgrade ... Cristante is in 22nd, bouyed by his defensive solidity but held back by his age and passing % ... Junior Mwanga for Bordeaux looks really interesting. He's moving up from CB to DM and doing well at 19 on initial impression. Small sample size, but he's around 30th ... Enzo Le Fée is also around 30th. He's an incredibly active, likable player who basically does everything a lot: touches, shoots, passes, tackles, dribbles, loses the ball, gets dribbled around, steals it back, everything. On first blush, I'd like him more as an 8 … As far as in-house options, Sambi/Zinchenko/White are all in the 30-50 range, but there's a lot of statistical noise there. Sambi is still helped by high passing % and high duel % over the last year. I can't project Kiwior/Tomi with much confidence (nor do I necessarily want to 😊). I have some #thoughts, and write something up about them in the event that a signing isn't made ... Patino is the bottom 10 as it stands. He gets hit by the league adjustment and the passing efficiency. Not ready yet.

Remaining rankings

After deleting some of the top (unavailable) DM’s, and cleaning up some others, here’s the list of remaining players. There’s probably some more noise and issues with the bottom half as my familiarity with them wanes (and impacts my ability to spot-check). Nonetheless, maybe it’s helpful to get the whole list in groups:

Players in the 31-50 range

K. Taylor, Ajax / M. Hjulmand, Lecce / M. Dahoud, Borussia Dortmund / N. Dorsch, Augsburg / L. Paredes, Juventus / R. Zerrouki, Twente / D. Sow, Eintracht Frankfurt / I. Gueye, Everton / F. Beltrán, Celta de Vigo / A. Bernabé, Parma / A. Haidara, RB Leipzig / N. Domínguez, Bologna / V. Janelt, Brentford / G. Fernandes, Beşiktaş / B. Soumaré, Leicester City / E. Palacios, Bayer Leverkusen / V. Souza, Espanyol / J. Garner, Everton / S. Berge, Sheffield United / N. Fagioli, Juventus

Players in the 51-72 range

S. Eustáquio, Porto / J. Lepenant, Olympique Lyonnais / S. Ricci, Torino / U. Račić, Sporting Braga / R. Tapia, Celta de Vigo / S. Serdar, Hertha BSC / P. Gueye, Olympique Marseille / M. Perrone, Vélez Sarsfield / S. Özcan, Borussia Dortmund / D. Frattesi, Sassuolo / R. Onyedika, Club Brugge / H. Guillamón, Valencia / S. Abdul Samed, Lens / Y. Musah, Valencia / L. Gourna-Douath, Salzburg / J. Chotard, Montpellier / P. Rosario, Nice / I. Baba, Mallorca / W. McKennie, Juventus / L. Tousart, Hertha BSC / Du Queiroz, Corinthians / E. Matazo, Monaco

Anybody I forgot, let me know.

🔥 Wrapping it up

Looking through this list, I think there are plenty of interesting options. This exercise confirmed the reasoning behind a few possible pursuits (Rice, Caicedo, Zubimendi), while floating some less-obvious options that could nonetheless add quality down the stretch and beyond.

The trouble, as always, is going to be sticking to the valuation strategy and balancing short-term needs with the long-term project; I believe the team did a superb job of this with Trossard. Because the long-term project would be so bolstered by trophies and a Champions League place, I don’t think the team can afford to approach the second half of the season with such a dearth of options in the defensive midfield.

As always with data like this, my hope is that it just gives you helpful context to what your eyes tell you. Ultimately, this is less about making the case for certain players, and more about being supported by data instead of “data-driven,” which I think can lead toward some unfortunate tendencies. Read, watch, then make thy own mind up.

In conclusion: leeeeet’s move.

Have a great one.

And happy grilling.

🔥