Here’s a draft of the Nagbe article posted over at DirtySouthSoccer. Posting it here just so I’ve got it here, and also cuz it was just sitting in the draft folder.
By now you know the big news. That Atlanta has struck a deal with Portland Timbers for 27 year old USMNT wide/central midfielder Darlington Nagbe in exchange for $650K of GAM and $300K of TAM in 2018, $100K of GAM in 2019, and $600K of additional allocation money contingent upon Nagbe’s and ATLUTD’s performance during his contract term, and a 2018 international spot.
This is exciting and I have a bunch of questions. This post is going to explore these questions and attempt an objective evaluation of Nagbe as a player, as a fit for ATLUTD, and as a transaction in MLS. Put another way I’m going to try to answer the question: Why Nagbe?
I think the first question on everyone’s mind (especially as the roster continues in a state of flux between the inaugural season and the upcoming 2018 campaign) is: was Nagbe brought in to play from a wide attacking position or from a more central box-to-box role? According to whoscored, Nagbe started 22 matches out wide (16 left), and 7 times from a central midfield spot. From 2013 – 2017, 125 (77%) of his 162 starts have come from wide positions, and 27 (17%) have come from the central midfield role. Also 10 appearances in a #10 role.
Here’s Carlos Bocanegra from the official team site: “Darlington is one of the most dynamic players in the league, and what stands out to me about him is that he compliments our attacking options so well. He has the ability to dribble out of trouble, he has a very unique quality in his ability to skip past a few defenders and drag defenders out of position, which leaves his teammates in great positions one-on-one or with mismatches down the field.”
He continues: “There were a couple guys we considered that we thought could come from within the league and make a difference in our midfield. The others would’ve probably forced Miguel Almirón a little deeper in the middle, and we didn’t want to take him further away from goal. So Darlington fits our lineup perfectly. Whether he’s deployed centrally or on the wing – he will chip in with a few goals and assists but he’s going to be the connector. He had one of the highest possession retention rates in the league, which is perfect for our team and style building out of the back and playing through the lines. He’s the exact piece that we wanted, and we were honestly surprised that he was available. So we’re excited we were able to get him”.
While the “What Darlington Nagbe brings to the 5-Stripes” video is just a compilation of goals he’s scored, I agree with many others who have written about the signing in that I don’t think that’s why they signed him.
By all accounts Nagbe’s strengths are in passing, ball retention, and carrying the ball through lines of defenders in midfield.
I thought I would check some of these conventional wisdoms against the stats and pulled some individual player stats from whoscored.com and started sifting through them. This chart, which includes players with 800 minutes or more and who played primarily in midfield (as broadly defined, think center mids, wide mids and 10s) caught my eye:
Did you find Larentowicz and Carmona? Hint, fly Southwest.
While one might not expect for Larry and Carmona to be taking players on 1v1 from the central midfield areas, I think it’s telling that even compared to other midfielders like Ring and Bradley and Chara, they are on an island all alone called zero. It’s plausible that Tata identified this as an area of improvement for his midfield, the ability to eliminate defenders on the dribble to open up high quality passes into the final third or throughballs in on goal to Martinez and Villalba. There are three Nagbe dots above, his overall average, his average playing wide, and his average playing centrally, where I think we can expect to see him for Atlanta United, and they’re all quite good. While his dribble output is the highest in wide spaces (makes sense), even in a more central role, Nagbe is successfully dribbling by defenders more often than once per game and he’s completing dribbles right around 70% of the time, which is a better rate than player’s like Villalba, Almiron, Schweinsteiger, Blanco, Asad, and Lodeiro. This seems like it could be a positive addition to an Atlanta United midfield that frankly I didn’t see much in the way of weaknesses in 2017.
By the numbers, Nagbe is a better passer than maybe anyone else in MLS. By now you’ve certainly heard the soundbite about him leading MLS in passing at 92% accuracy. While this is an intuitive data point, a flat pass completion % is often misleading. What if the player only attempts easy passes? Luckily, there are other metrics we can use to take a stab at evaluating passing ability. AmericanSoccerAnalysis.com has developed one such metric, which they call xPassing. By comparing all of a player’s pass attempts against the average success rates for similar pass attempts across the entire league in previous seasons, they develop an “expected passing percentage” for each player based on the passes he’s attempted. If a player’s actual pass accuracy % is greater than the calculated expected passing percentage for the profile of passes he attempts, than he is an above average passer. I took ASA’s data, which is categorized into final third passes, middle third passes, and defensive third passes and dropped them in excel.
To start, I filtered on only players with greater than 700 “middle third passes” and charted their success rates above or below the expected rates on the y axis (at the top of the graph are better passers and at the bottom the worse passers), and on the x axis I’ve plotted the average xPass % (players who on average attempt easier passes are on the left, and more daring passers are on the right):
You can see that while Carmona and Larentowicz are above average passers, Nagbe towers above everyone, a true outlier. Also, Carmona is a bit of an outlier in terms of attempting almost exclusively high percentage passes. Nagbe is closer to Carmona than to Larentowicz in terms of passing profile (passes that are historically completed at a higher rate). For grins, check out Asad (lower left) who as we might expect takes huge chances with riskier passes. The models have him coming in lower than one might expect, and I won’t argue with it, except to say you make your own luck. Watching Asad in 2017 was so thrilling because he would try literally anything and when surrounded by goal scorers, this has its benefits – and he saw those benefits with monster assist numbers.
Nagbe pops even more if we look only at passes in the final third. The chart below includes all MLS players with 300+ final third passes.
Again, Nagbe is in a league of his own in terms of passing accuracy in excess of expected pass accuracy based on historical averages for similar passes in the final third. He’s also in the upper tier in terms of choosing less risky passes. So in terms of passing, it is true that whether he’s in the middle third or final third, Nagbe chooses to play exceedingly high percentage passes and he completes those passes even more routinely than they are typically completed. He appears to be a good passer, and probably a better overall passer than a chance creator.
One thing that not taking on defenders in the central midfield areas gets for you is somewhat higher ball retention rates (Carmona and Jeff high-five each other). If we’re going to have old Darlington dribbling around in midfield, let’s make sure he’s not a turnover machine like the beloved Asad (who fortunately for ATLUTD, turned the ball over further up the field). Charting the sum of “unsuccessful touches” and dispossessions” per 90 minutes from whoscored on the y axis and average number of passes per 90 on the x axis, we get the following for “midfielders”:
Upper left are players who turn the ball over at a higher rate per touch, and players lower right care for the ball like a baby bird. So the usual suspects show up on the high turnover end of things, the attacking midfielders and wide players. Of note for our purposes are Asad and Almiron in the 4+ p90 range. Poor Johan Vinegas isn’t even pictured above because I had to cut off the y axis (10.2 turnovers p90!!!). And also of note, Nagbe shows up in the 3-4 range when playing wide, but only the 1-2 p90 range on 56 average passes, when playing centrally. He profiles similar to a Carmona or an Alexander Ring when he plays there, and that’s good. For reference, Larentowicz is at the absolute midfielders’ league floor of 1 per 90. Good old Jeff. Michael Bradley shows up impressively giving the ball away under 2 per 90 despite averaging 77 passes p90. These figures make me feel a little better about having a fancy dribbler guy in midfield.
Much of the talk of Nagbe’s qualities has focused not on his final third product but more on buildup figures and his unique abilities to move and retain the ball. There’s a reason for this, he only registered 3 goals and 3 assists in 2017 and his underlying xG and xA numbers were even lower at 1.95 and 1.23 respectively (according to @analysisevolved). Most everything points towards his value being in buildup and not in chance creation. Here’s a chart of MLS midfielders’ key passes and shots per 90 minutes (minimum 800 minutes):
To me, this is informative because while Nagbe’s 2017 minutes at central midfielder were limited, he put up key pass / shot volume profiles that were more attacking than the household central midfield names like Bradley, Dax, Alonso, Ring, and also Atlanta’s Carmona and Larentowicz. He basically put up key pass numbers that looked like a decent attacking midfielder but shot volumes that looked like a #6. Schweinsteiger was a good comparable in 2017 in terms of chance creation. It strikes me that this “link up” / “connecting” player we keep hearing about that Nagbe will be in Atlanta… there just aren’t many examples of this type of player in MLS, partly because its such a unique profile. That’s somewhat worrying if only because it’s hard for me to completely visualize how 2018 will look in an average match. But it’s also a reminder that Nagbe is a rare talent — I just can’t think of many “connecting” midfielders in MLS who dribble and pass with ease and play behind the #10. Nagbe’s unique mix of skills is very rare, and so it make sense that you’d have to pay for something this rare. We’ll get to that later.
Quickly on Shot Selection
If there is one area of Nagbe’s game I’m confident he needs to improve, it’s his shot selection. Here are the shots per 90 (“inside the box” on the x axis and “outside the box” on the y) for “midfielders” in 2017:
This isn’t completely worrying, but Nagbe doesn’t shoot often compared to other midfielders and when he does it’s from range. These shots are very very rarely high percentage shots. We won’t expect him to contribute very often as a shooter, and it’s not uncommon for a holding midfielder to take a higher percentage of his shots from range, but even glancing below Nagbe in the graphic we can see the other low volume shooters contributing some from inside the box. With so many goal scorers and chance creators playing in front of him in the team, Nagbe will need to pass up the opportunity to have a bite from range and instead feed the ball to the expensive South Americans who create goals.
So if Nagbe isn’t a “find the final ball” guy and he’s not a “get on the end of the final ball” guy and he’s not a “shield the back line” guy, I guess he’s the “everything else” guy. AmericanSoccerAnalysis has a metric with which to view a player’s contributions to possessions which end in the creation of a chance. It’s called “xG chain” and it basically assigns value evenly to every action in a sequence leading up to a shot based on the final xG value of the shot. So if Nagbe’s good at linking up plays that ultimately end in good chances, he should still show up with decent xG Chain numbers, even if he’s not involved in the final ball. Similarly “xG chain – buildup” measures the same thing but without considering assigning any values to the final passes or shots. We would expect Nagbe to shine better in this second metric. Anyhow, here are both 2017 xG chain (Y) and xG chain buildups (X) per 96 minutes according to ASA:
Nagbe’s buildup play is on the higher end, in good company at +0.32 in the same ballpark as Kljestan, Victor Vazquez, Mauro Diaz, and Almiron. And unsurprisingly, his overall xG chain isn’t isn’t much higher than his xG chain buildup, suggesting most of his contribution does in fact come in linkup play and not in the final ball. If we’re going to ask him not to creat chances but just to link up midfield and attack, than we’d want him to have a good buildup number, and he does. So that’s good.
Assessing a player’s defensive value or even defensive contribution using on-ball event data is definitely questionable. Counting-type stats like interceptions and tackles and clearances are mostly descriptive and are almost entirely detached from “value.” A player might tackle more because he’s facing more take-ons, not because he’s a good tackler. That being said, because Nagbe will likely be shifting into a role that he hasn’t played as often in the past, I thought we could take a look at interceptions and tackles, and as long as we keep our “descriptive” goggles on, and don’t get too carried away, maybe it’s worth looking at. Here are tackles and interceptions per 90 minutes for all “midfielders” in 2017:
Remember, the frequency a player tackles or intercepts is often just as much a reflection of how often he is put in the position to make such a tackle as it is a reflection of his ability to break up play. Adjusting these figures for possession would be better, but I don’t know how to do that cleanly. So if we just caveat this to hell and back, what do we see? I’m slightly encouraged by the fact that when Nagbe plays CM he either does (or is asked to do) a similar amount of on-ball defending as Carmona and Asad. None of these guys sniff Larentowicz who appears to be an on-ball defending demi-god of sorts, but at the very least Nagbe is either just as capable of, or has been put in similar situations (albeit in limited appearances) as Carmona and Asad (who is a very high defending wide player). If what many expect to see comes true, and Carmona does indeed drop back into a more 2017-Jeff role, and Nagbe drops into the 2017 Carmona role, it would appear as though he’s not entirely foreign to it (I was surprised to see this honestly).
Also, I should add that Nagbe tackles with efficiency. Pictured below (tackles p90 on the X and tackle success rate on the Y):
He’s… wow.. he’s very good by this metric. Not sure how much stock to put in this, but it pops visually. Cool.
What I’m most concerned about defensively is Nagbe’s ability to find his spot in the high pressure scheme that Tata employs. From watching the games last year, Carmona very often led the high press by launching forward from central areas to squeeze opponents in possession. He may not have collected all of the tackles and interception stats, but he likely caused many of them by initiating the press and funneling the ball towards pressing traps. I haven’t seen enough of Nagbe to feel confident that this is in his game, nor the awareness in a transition moment to jump start the press but I don’t know for sure. Matt Doyle commented on ET Radio that this was not a strength of Nagbe’s game.
Finally, the tricky question: Was the trade worth it?
There are many ways to answer this and lots of factors to consider. First, let’s just conclude quickly that based on the above charts and what not that Nagbe does in fact carry a unique and perhaps even rare combination of highly valued skills. Those skills do not or perhaps have not translated into final third “end product” in the way that most highly valued soccer skills do.
Second, let’s assume a “highest and best use” proposition that Nagbe will in fact play as a “linking player” in a “#8-type” “box to box” role and that Darren Eales and company identified him to fit into this role. Quotes from the team certainly indicate that all of this is true and his lack of chance creation numbers suggest that perhaps he shouldn’t have been playing out wide. Who knows.
Third, I just want to contextualize this some more before we start talking dollars and cents. When I watched Atlanta United in 2017, I don’t think that at any point in any of the games I said to myself or those around me “I wish Jeff Larentowicz played more like Carlos Carmona in the deepest midfielder role and I wish Carlos Carmona played more like Darlington Nagbe in the box2box role.” We can agree that Jeff and Carlos did not complete 1v1 take-ons the way Nagbe can. We can even agree that the extent to which they did not attempt these duels is perhaps low by central midfield standards. And we can agree that Nagbe has another level to him in terms of ball retention in traffic. But it seemed to “work” last year without him, at least in my eyes. And that’s worth pondering.
That all being said, as long as Larry and Carmona are coming back, and we we agree that we needed to add a central midfielder anyways just for depth reasons, Nagbe isn’t a bad type of player to add (Jeff to the bench). On balance, he would likely improve every single team in MLS right now, so that should count for something. Are Atlanta that special?
Further, the advantage that Atlanta United may have had in 2017 in terms of being able to efficiently create a team’s wage budget from scratch without having to navigate existing chunks of budget already allocated to players (and yes this has its downsides) may be coming to an end as a) LAFC rolls into the league with a presumably huge war chest of expansion allocation money to play with and b) all teams were just handed the option to dip into $2.8M in additional TAM per year. So, to the extent that teams were handicapped in 2017 by their past mistakes (bad contracts), 2018 is going to be a different story in which every. single. team. will have the ability to rebuild and improve their squads. It is in this very context that dealing for Nagbe makes sense — the context that says every single team in the league (except DC United and Columbus) will theoretically improve in terms of roster strength and depth, and so because Atlanta will have to as well, you may as well grab a domestic player who is almost entirely unique in his skill set and in the prime of his career.
What about the price though? $650K in GAM now. $300K in TAM now. $100K in GAM in 2018, $600K in contingent fees if Nagbe and ATLUTD hit certain milestones. First of all, when we saw the news of the $650K in 2018, many of us were worried because this is the exact amount of GAM that the team would generate (by rule) if it sold a player out of MLS at decent a profit – so for a moment there, we thought the Nagbe deal meant either Carmona was being sold, or Almiron or Martinez was being sold. And I’m not going to lie, this initial reaction has colored the way I’ve thought about the trade the whole time. BUT, Darren Eales on ET Radio came out and said that actually the GAM that went northwest in this deal was dry powder from the existing expansion year GAM allocation that the team had been saving just in case a huge opportunity like this one showed up.
And by all accounts, this did *show up* — Portland decided to deal Nagbe and started shopping him, and Atlanta was like “what now? I mean, I’ll have to ask my boss, but like… is this a good number to reach you at? And and can you say which player it is again that you’re looking to move? Yea, let me get a pen, it’s Darlington (how do you spell that) Nagbe and not Liam Ridgewell, right. Yea, ok got it. Well, I’ll see what I can do, money is tight these days but I understand where you’re coming from and we look out for each other right”
If you’ve kept up with some of my other posts on the finances of building an MLS roster, you know by now that I DO-NOT-LIKE-TRADING-ALLOCATION-MONEY-FOR-PLAYERS. Simply put, when a team does this, it is trading its ability to increase its wage bill above other teams…to another team. It is saying “here, we will pay our team less and you can pay your team more, and in exchange we’ll take your player and we’ll have to pay him with what is now less.” So my immediate reaction was that this was a bad trade. A mistake. I’m now fairly neutral on this (though I am definitely excited to see him play), having been persuaded by a few arguments in favor of the deal:
- While sending allocation money to your competition isn’t ideal, in this case, the alternative of finding a player of Nagbe’s quality for $1M + add-ons in the international market is difficult (transfermarkt shows a bunch of “comparable players” in the 2M euro range). And finding a player of Nagbe’s quality for less than that who is only making $565K a year in wages is probably impossible. I should add that Nagbe will get a raise when his contract is up – he wanted $1M from Portland, thus the trade shopping. Because of this, his contract is currently a “favorable” one, an “asset” in accounting talk. If by chance his contract has more than one year left on it (and it appears that this might be the case), than this might in fact be a slam dunk of a trade. A player worthy of $1M making half of that for 2 years running starts to pay for the $1M fee VERY quickly. Now, who knows what Atlanta will do and when in terms of extending his deal and giving him the money he wants, but this seems pretty favorable. Further, paying GAM for a player but using TAM to pay down his wage to the max (and potentially further to free up budget space for other players) isn’t a terrible combination. I’d be more worried about sending a boatload of TAM away to then turn around and use more TAM to keep a player’s wages compliant with the rules.
- With all of the new money flowing into the league being TAM (and the rule being that you can’t use TAM and GAM on the same player), the fact that only $300K of this deal is TAM is a positive development. It would appear that the club hasn’t even dipped into the discretionary TAM available, and that this deal won’t hamper the club’s further spending. If it’s true that Nagbe in and $650K of GAM out doesn’t mean another player is headed out (and that the GAM was saved from 2017), that’s good.
- Something, Something International Spots? I’m honestly confused about international spots in the new MLS. It would appear that high-end domestic players have suddenly become unicorns with higher valuations than before. And I’m not totally sure why. Perhaps it’s the fact that $64M of additional discretionary MLS fees/wages has just been approved for clubs to chase after high earning players (players that are often internationals) but without an increase in the number of international slots in the league (aside from the 8 new ones LAFC carries in). Perhaps, Atlanta is transitioning into a future state where a team cannot count on the market for international spots magically clearing every year at a relatively nominal cost. For instance, if Atlanta continued to roll into every season with 9 or 10 international players, perhaps teams would start to ransom them in a way that hurts, or perhaps the USSF presidency being up for grabs has spooked MLS into wondering if new constraints on international players are coming. Who knows. That all being said, Atlanta a) is good at converting internationals into domestics after a year of residency and b) if international spots are suddenly more valuable, why are Atlanta United trading all of theirs away? Watch this space I guess.
- It has been reported but not confirmed that the $600K of incentives in the total transfer fee for Nagbe are the type of incentives that are hard to hit, meaning it’s very possible the fee is just the $1M. Incentives around Nagbe scoring a bunch of goals or assists when he’s going to be dropped back into midfield and doesn’t have a history of creating these anyways, are a stretch. And Nagbe winning league MVP or Atlanta United winning the league are the types of things that are both a) unlikely, and b) if they were to occur, I would suddenly not give a damn what the fee was.
I’m ready for Nagbe to be successful in Atlanta and I will 100% be rooting for this to be a successful move on ATLUTD’s part. I still have some concerns, but ultimately after reviewing all this stuff, I don’t think there’s so much downside in this to where there’s a huge negative tail risk. If it doesn’t work out and he doesn’t fit the system, the club have given up some of the initial expansion GAM and not much else. Tata can always revert back to a Larry/Carmona midfield partnership, and who knows if Nagbe were to leave in a year or two perhaps there’s a transfer fee coming Atlanta’s way. If it works out well and Atlanta continue to improve the squad with some of the rumored international players they’ve been linked to (or they bring back Yamil Asad), I could see a very high ceiling for the team.