What the Swede’s role last year means for San Jose in 2020
In 2019, Magnus Eriksson was one of the top performers on a surprising San Jose Earthquakes club. After a will-he-or-won’t-he offseason in which there were rumors the former Designated Player was heading back to his native Sweden, Magnus was moved from his favored in-cutting right wing to a central role. In this new spot, he contributed six goals and seven assists (plus an Open Cup goal) – a marked improvement over his 6g/3a 2018 campaign. Eriksson is no longer a DP, and he even wore the captain’s armband for stretches when Wondo was substituted out or not playing.
But it is tough to say he ended the campaign on a high note. He never looked terrifically comfortable in the central attacking role, and had mentioned last offseason a desire to play a more deep-lying distributor role. But when he was put back there, he was largely ineffective – whether at the end of games, or against Seattle where – it should be noted – San Jose was shut out and Eriksson’s weak header back to a centerback led to the Sounders’ game-winning goal. He seems well-loved in the locker room and has earned Matias Almeyda’s trust, but few of the Quakes faithful are truly excited about a player who has in his two Californian years appeared a square peg hammered in to a series of round holes.
Compared statistically to other central midfielders in MLS, Eriksson is an odd bird indeed. An examination of playmaking and buildup, using American Soccer Analysis (ASA) data, demonstrates how he may be better suited to other parts of the pitch for his third stint in black and blue in 2020.
The Methodology
Watching San Jose all year, it always struck me how Eriksson plays so differently than his peers in the central attacking position. At times he seems more like an advanced destroyer or a regista who got lost and ended up on top of the 18-yard box. But he was third in the league in Key Passes per 96 minutes, so clearly there was something there.
I took all the central midfielders (broken up by ASA into defensive midfielders (CDM), central midfielders (CM) and attacking midfielders (CAM)) who played more than 2,000 minutes in 2019. I wanted to compare Eriksson to the median of each type of midfielder so I could see how his usage compared – not necessarily to the elite, but to the workaday MLS types that provide a baseline to this league. In order to get a true median (an odd number of datapoints), I needed to drop one of each type of midfielder. So sayonara to those who garnered just a hair over 2,000 minutes: Brandt Bronico, Andy Rose, and Paxton Pomykal. The median of the 57-man dataset is noted as CxM.
For curiosity’s sake, and to see if there were any wrinkles due to San Jose’s spread-out style of play, I highlighted Eriksson’s midfield partners: Jackson Yueill and Judson. And even though he only played 316 minutes (often chasing points), I included Andy Rios as a reference.
The dataset is available on Google Sheets. This chart uses expected goals, pass chains, and other weird vocabulary. ASA’s glossary may prove helpful.
Magnus the Playmaker
In MLS, there are very few stars who can both create for others and themselves. These guys (basically: Vela, Ibrahimovic, Elis, Diomande, Maxi Moralez, Valeri) are, broadly speaking, out of San Jose’s price range.
Eriksson distributes like an expensive CAM and finishes like an inexpensive CAM. That, from watching him play all year, sounds about right. Here’s the chart of xG/96 by xA/96:
Looking at that chart, you can probably think “here is a somewhat-effective #10.” And you’d probably be right.
Going further into this: here’s a chart of average shot distance from goal by key passes per 96 minutes. I used this chart as a bit of a way to frame positioning: is the player hovering in the box trying to set himself up, or is he outside and trying to set up others?
Keep in mind that this isn’t a barometer of quality. A “Key Pass” can set up a shot with a low expected goal value, and a long-distance shot could still have a relatively high expected goal value.
Eriksson was second only to Diego Valeri in this dataset in terms of KP/96. He is great at delivering that final pass. But he’s doing it from far away than your standard midfielder, almost more like a defensive midfielder joining the buildup than guy keying it. Eriksson shoots from further on average than Yueill, and is positively on the moon compared to Rios who (in very small sample size) was connecting and testing the keeper from inside the box. (And note to Judson: maybe you should ease up on some of those 30-yard screamers. The Bay’s not big enough for two Steph Currys.)
Eriksson in Buildup
A look at how San Jose builds into their attack, and Eriksson’s role in that buildup, demonstrates two things: the Earthquakes play very strange soccer, and Eriksson has a passivity that may or may not be due to that system.
The first graph is percentage of team touches belong to a player by that player’s passes per 96 minutes. In other words, how much involvement a player has. Here you can see that San Jose will pass, pass, and pass:
Fourth-most passes for (behind Toronto, LAFC, and NYCFC) and the fewest passes against (about 500 fewer than NYCFC). Though Eriksson may seem to distribute more than most midfielders, its more a product of San Jose trying to maintain possession through moving the ball. Yueill was one of ten players to complete more than 2,000 passes in 2019: he basically got the reins to a ballclub in his third professional season, without much on his resume beforehand. He ought to tithe his USMNT appearance bonuses to whatever charity Almeyda recommends.
Looking a passing by distance and verticality, you can see how CAMs generally have shorter and far less-vertical passes because they operate in the final third. For Eriksson, it’s even more pronounced. This spread shows some of the more interesting bits of how the three midfielders work on a string in San Jose’s system: Judson is flipping it between fullbacks, Yueill is getting it to wingers, and the CAM (whether it’s Magnus or Rios) is responsible for getting that final ball in tight spaces. It is a far more spread-out system than the median MLS setup, and it shows why Eriksson has so many key passes compared to the average guy in his position: he’s operating much closer to the forwards. The issue is that he is not, as we have seen, operating inside the box.
Buildup Versus Final Product
The next couple of graphs indicate what may be the biggest hole in Eriksson’s game: lethality with the final ball.
To explain how ASA determines passing score, it is a player’s pass completion percentage minus their expected pass completion percentage. As you can see, San Jose midfielders are on the more accurate side of most MLS midfielders. This, again, has to do with the system: spread the defense out and keep moving it side-to-side until you find a weakness.
At first glance, you could look at this graph and think: wow, Eriksson is a much better passer than any other CAM. But what’s really happening is: Eriksson is a much more conservative passer than any other CAM (well, except for Philadelphia’s Jamiro Monteiro and RSL’s Albert Rusnak). As you can see, the median CAM has a negative Passing Score. I read this as “a CAM’s role is to get some really tricky balls into a good space for the shooter.” It’s not really a negative trait but more of a “is this guy an assassin?”
The five lowest scores in the dataset are: Pity Martinez, Nico Gaitan, Quintero, Kaku, and Diego Rubio. Five guys who absolutely want to kill the opponent with a final ball. Lodiero is 9th, Valeri is 12th. True, Leonardo Bertone is 8th and Artur is 10th, so it’s not totally descriptive. But there’s something here. Any statistician would yell at me for showing the following chart as a positive, what with an R2 of .099, but there is a slight inverse correlation between passing score and xA/96 within the dataset:
Eriksson is much better than most midfielders! That said, he also has the non-lethality (I don’t want to necessarily say passivity) characteristic of most midfielders, albeit with a much higher passing percentage. This could be who he is (not a risk taker, a #6 playing #10, happy to redistribute) or it could be the system (San Jose wants to dominate possession, they’re happy to play safer passes).
Here is why I think it’s the player, not the system:
xB/96 is team-adjusted goals on possessions in which the player participated but did not make that final shot/pass (basically: quality of buildup); playerKP% is the percentage of possessions in which the player does make the final pass.
In 2019, Eriksson was fourth in Key Passes. It goes: Valeri, Gaitan, Jack Price (oh the joy of delivering set pieces to Kei Kamara’s head), Magnus, Pity, Quintero, Rusnak, Carles Gil, Lodiero, Kaku. These are, generally speaking, the dudes you want on your team with the ball in the final third.
The thing is, when Magnus is on the ball these are not necessarily great passes. He is (on average) further from goal than a CAM, making shorter passes than a CAM, and making less-risky passes than a CAM. This would, I think, lead to fewer golden chances than you’d want out of a CAM. This makes sense if you look at his great xB: he is distributing between wingers or between the strikers and the winger, which is good. But, Yueill’s doing the same thing and better.
To get more anecdotal: passing triangles are larger with him than with most CAMs. In San Jose’s attacking system, it’s the role of the wingers to accrete gravity and draw folks to them. Both Vako and Espinoza excel at drawing defenders. The fullbacks provide interplay to the endline (which is why Tommy is a fullback in this system: he is phenomenal in tight spaces, which gives more freedom to Espinoza). The CAM has the responsibility to quickly tilt the field away from the winger once they draw defenders, switching the attack and drawing the defense off balance.
Eriksson, generally, doesn’t do this. Part of this may be his style of play, but part may also be skillset. To be even more anecdotal and a bit theoretical: if I’m defending him in the box, I’m just keeping him on his right foot until he turns around and passes back to Espinoza or recycles to Yueill. There’s no reason to foul him in the box as long as you can keep your body between his left foot and the goal. Magnus has a great spatially-oriented mind, but he just can’t operate in tight spaces because the defender can restrict his already-limited passing range.
What does this mean for San Jose in 2020?
The Earthquakes were third to only LAFC and Chicago in Expected Goals (the Quakes’ 54.7 was just behind the Fire’s 54.9 but quite under LAFC’s obscene 68.1). But they were 10th in actual goals due to their league-worst gap between Goals and Expected Goals (-4.6). Sure, a solid chunk of that is luck. Another bit is Danny Hoesen misplacing his finishing shoes over the winter: his -3.9 G-XG was second only to the truly cursed Daniel Salloi (-5.9!!) in MLS.
But this G-xG gap is also a failure to create great opportunities. It’s letting luck (in the way of a deflection, or a crossbar) get in the way and snowball because those high-xG chances just aren’t there. Even in San Jose’s possession-dominated system, several low-xG shots don’t equal a single high-xG shot because shots so often lead to loss of possession. Eriksson creates many opportunities, but he doesn’t create many great opportunities.
This seems to be the missing link in San Jose’s offense: somebody who can work with the immensely talented wingers to deliver a lethal final ball.
Is Rios that person? It is honestly, too soon to tell. One would hope so, with his wages and from his limited looks. Rios likes to lurk in the box and has the profile (if not the statistics) of an Argentine enganche. It is extremely unlikely San Jose can find the money to both keep Espinoza on the roster and fill this role from outside the system, so Earthquakes fans are likely best served to hope for Andy Rios.
If Rios does take that role, Magnus Eriksson could end up as a rotation midfielder. This is, simply put, a good thing. If San Jose essentially keeps Eriksson as Anibal Godoy’s replacement, he is a not-too-expensive and versatile midfielder who can play anywhere between the centerbacks and the striker without giving up too much. Eriksson seems well-suited to both CONCACAF-ing a game to sleep or providing an extra attacking spark, depending on the need. And if Yueill is going to miss large chunks of the season on international duty – or even be sold to Europe if his stock continues to rise – then it will help to have a quality passer who knows the system and is very useful in set pieces. As excited as I am to see more of Eric Calvillo, I am more comfortable in seeing 1,500 minutes of Eriksson if San Jose is going to be competing in the playoffs.
Eriksson fits the profile of a deep-lying distributor forced into a role he wasn’t ideally suited for, but he did well enough in that position and appears to have Almeyda’s trust. He was a strange bird in San Jose’s strange system, to be sure, but San Jose’s possession-heavy system relied on his spatial awareness. As the team develops, they may find a more lethal attacking midfielder but that is no reason to drop Eriksson entirely: having him in the midfield rotation is crucial for depth and maintaining the system in the next iteration of San Jose soccer.
Great article!
Love this post. Can’t ever get enough of Magnus analysis. (I love him almost as much as I love how divisive he is to the Quakes fanbase, lol.)
> The Earthquakes were third to only LAFC and Chicago in Expected Goals (the Quakes’ 54.7 was just behind the Fire’s 54.9 but quite under LAFC’s obscene 68.1). But they were 10th in actual goals due to their league-worst gap between Goals and Expected Goals (-4.6). Sure, a solid chunk of that is luck. Another bit is Danny Hoesen misplacing his finishing shoes over the winter: his -3.9 G-XG was second only to the truly cursed Daniel Salloi (-5.9!!) in MLS.
This is like, the craziest paragraph in the whole damn thing.
Would love to get more of a deeper look on Rios and Fierro as well. I know they haven’t had a whole lot of minutes yet, which in itself could be cause for concern, but they still just seem like such an unfamiliar part of the team for me and likely a number of other fans, too.