StatsBomb MythBusting: Christian Eriksen vs. Newcastle
This past weekend in the Premier League was a crazy one, and no single game was crazier than Spurs vs. Newcastle*. Newcastle looked dangerous early on, and actually took the lead in the 13th minute on the back of a Paulinho mistake, sparked off some surprise pressing from the Geordie side. What followed were 77 minutes plus stoppage time of hell for the Newcastle defense.
Final score: Tottenham Hotspur 0 – Newcastle 1
* Those of you in Swansea and West Bromwich wanting to make cases about the referee robberies in your games should make contact with the appropriate Premier League authorities as soon as possible.
I mentioned after the game wrapped that Tim Krul was out of his mind good (14 saves in a match almost never happens), and also that Christian Eriksen was really tremendous for the home side.
In 77 minutes against an entrenched defense, Eriksen had 9 key passes, 3 shots, 3 shots on target, 71% passing completion. That’s a damned fine day for an attacking midfielder. The passing percentage is a bit low, but the rest of it… impressive. He didn’t do a ton of defensive work (1 tackle, 1 Int, 1 Aerial Duel won), but Spurs have an entire team built for that. Defensive work is not his job. Eriksen’s job is to find his teammates with passes that they can turn into shots, and boy did he do that.
Thus I was a little surprised at the reaction of Spurs fans to my Eriksen opinion. The abuse that I got from Spurs fans was loud. And kinda angry, really. A lot of these aren’t from dumb guys – one is from one of the Spurs writers I respect the most - but boy were they hot about the loss.
Visually, this is what Eriksen's performance looked like from the FourFourTwo StatsZone
So I said Eriksen was awesome, and other people said…
“Outstanding?! He was one of our worst players. Missed a sitter, bad set pieces, lots of wasted possession.”
“ was ineffective and inefficient.”
“You don’t know what you are talking about. You’d have a different perspective if you watched him in the stadium. Nobody in the stadium thought he was good.”
“ was good – at least he was trying his heart out.”
“His set piece taking was awful. Went into hiding. Fit the stereotype of the scouting reports.”
“He’s made a MASSIVE jump in league standard.”
I will note I did watch the game in detail while it was on. I actually watch a lot of Spurs games these days, partly because I like some of the players, partly because I like the AVB system, but mostly because I really need to know why they aren’t scoring more goals for work.
I could be wrong about what I saw. Obviously I have a bias toward using data to examining football from the perspective of data, but what if Eriksen’s performance really wasn’t that special? I decided to look at how often a performance like CE’s appear, and also at some more contextual stats to better examine what happened on Sunday. I also rewatched the video with a focus on what Eriksen was doing.
Time for some mythbustin’!
Passing is Misunderstood
One of the first pieces I wrote about football stats involved looking at assists as a key indicator for finding good players. Somewhat to my surprise, I discovered the ability to create high numbers of assists at a young age typically translates to stardom.
I used assists in that analysis because key pass information wasn’t publically available for earlier seasons, but assists have been tracked in certain leagues since at least 2003. However, key passes are a part of the process that leads to assists, and occur more frequently, so when possible, I try to use that information instead.
Why? It has to do with separating outcomes from processes.
Example: A player passes the ball to an open attacker at the 18 yard box, who then shoots, and kicks the ball into the upper deck.
The outcome there is bad (missed shot). The process, however (finding the open man 18 yards out for a shot), is good.
Is the missed shot the fault of the passer or the shooter? That type of analysis is difficult, but let’s just say that most of the fault for the miss probably lies with the shooter. If that shooter is particularly bad at getting shots on target, regardless of who passes to him, then he’ll drag down the assist stats of the passer as well. However, key passes will still get registered.
That’s why key passes are interesting to track, and high levels of KP usually translate to high levels of goals/assists.
Before I get too far ahead of myself, we probably need a definition. What is a key pass? Opta defines it as:
The final pass or pass-cum-shot leading to the recipient of the ball having an attempt at goal without scoring.
Key passes are correlated to assists, which are goals resulting from a pass, and are a fairly strong measure of whether a player is a good offensive passer.
The top 5 in EPL for key passes last year were David Silva, Baines, Suarez, Mata, and Steven Gerrard. The year before was Mata, Silva, Modric, Nasri, RVP, and Valencia (13 assists for United that season in 22 starts!). It’s an imperfect, but damned useful stat.
So key passes matter. In fact, they usually tend to indicate players who not only create chances, but create +EV ones. I would far prefer to have a bunch of guys on my team who pass well enough that they create shots with scoring probabilities of 30-40% (open looks in the box, 1v1s with the keeper) vs. the 10% average. Championship winning teams are built on this principle.
And yet teams still don’t seem to value great passers enough. Real Madrid sold Mesut Ozil, the best creator in Europe over the last five seasons. Goal scorers get all the money and the attention. They also get the adoration of fans.
How Rare Was Eriksen’s 9 Key Pass performance?
I wanted to look at this because maybe this type of performance happens all the time. I may have found it impressive, but if it’s a common occurrence, then we can rule out my impression and just write it off as something any decent footballer can do.
I only have access to a shortened data set on this one (last year plus this year), but out of more than 1800 matches played in the big 5 leagues last season, there were fifteen times where a player had nine or more KP. Players who produced those performances included: Pirlo, Ozil, Diamanti (who also had 14 shots in the same match, one of the most absurd statistical performances I have ever seen), Boudebouz x 2, Obraniak, Farfan, Diego x 2, Tobias Werner (who?), Valeron, Susaeta, Rakitic, Jose Antonio Reyes, and Koke.
There has also been one 9+ KP game this season, produced by the ageless Francesco Totti.
So 16 total instances in more than 2000 matches. That’s less than 1%. It’s fair to say that’s rare.
“He’s made a MASSIVE jump in league standard.”
This is true, and I'm actually not certain how big the jump is. Is Eredivisie 60% as good as the Premier League? 70% I have no idea, really.
However, not only is a 9+ KP performance rare, it hasn’t happened at all in England either this season or last. That’s nearly 500 matches without one. I checked with Duncan Alexander at Opta, and he said the last person to do it was Leighton Baines in September 2011! So we're actually looking at like 2 in 800 matches.
Eriksen made a massive jump in league standard…
…and still hung one of the better offensive passing performances in recent league history on a Newcastle team that is actually pretty good.
Clearly, he’s shit.
Not All Key Passes Are Alike. Also, 71% Passing Success??!
Both of these are fair points. All KPs are not alike. Some create shots that are practically worthless (like almost any ball delivered to Andros Townsend). Others create shots that are almost goals in and of themselves. Still, Eriksen created nine of them, including a couple of peaches in a free header from a corner that was off the bar and the exquisite layoff to Paulinho in the center of the box.
Eriksen also had a set piece delivery directly to Soldado’s head that resulted in a shot on target that Krul made a good save on.
Now 71% passing isn’t very good, but it’s something that I thought was strange too. Eriksen is normally very clean with the ball, but passing through a stacked defense for the entire match is hard.
Then I dug into the numbers and noticed 13 crosses in the match. Crosses are usually completed at a 20% clip. Eriksen’s non-cross completion was 88%, but he was 2/13 on crosses.
Edit: Dan Kennett points out below in a comment that passing stats already have crosses removed, which I did not know. Read his point below for more illumination.
Check out this old image from Gabe Dejardins about expected completion rates for passes based on location of the target. Assume for a second that Soldado was generally in the blue spots and it gets easier to understand.
Eriksen also had three shots in the match and three shots on target (again, very good). Krul managed to save all three, just like he did with the 11 shots on target from Eriksen’s teammate. Sure, Spurs players made some of them too easy for him, but in general Krul played like an octopus with keeper gloves on.
“ went into hiding, fitting with the stereotype of the scouting reports.”
Really? Dude had 81 touches, or 14 more than any other Spurs offensive player. He was also only dispossessed one time. Townsend had 4 in a full match, Sigurdsson had 4 in 69 minutes. This statement is just silly.
“ was good – at least he was trying his heart out.”
Andros Townsend has one fluke, completely-by-accident goal from 45 shots this season!
His average shot distance is 26 yards!
He has 0 assists in 10 games!
Get. The fuck. Out.
Townsend’s performance so far this season is exactly what Shakespeare was thinking of when he wrote, “Full of sound and fury. Signifying nothing.”
Okay fine… I’ll come clean.
I actually thought Andros was decent on Sunday. He had some good lateral passes to setup dangerous shots and his shooting was better than it has been this season. Let’s not pretend he deserves to be in the lineup more than Lamela though, okay?
This Was Spurs Best Offensive Performance of the Season
And Eriksen was all over it. Spurs had one match last season with 14 shots on target, the 1-0 at home against Sunderland on the final day of the season (which included a red card). They also had 13 as part of a 3-2 win at West Ham. Sunday they produced 31 shots, 14 on target, and got beat 0-1.
I’ve looked at a couple of different models and they thought the score probably should have been 3 or 4-1 for Spurs. Shit happens.
Spurs have only scored 6 open play goals in 11 matches, which is dreadful and is likely not the result of just being unlucky. AVB has frequently sold out on defense in personnel selections so far this season, and it’s reflected by that goal return. On Sunday though, the team had more balance, and they were outstanding offensively against a Newcastle team who are actually pretty good.
Outcome: Spurs lost 0-1 at home to Newcastle.
Process: They had 31 shots and 14 shots on target. Christian Eriksen did something that hasn’t been done in the Premier League in at least 500 matches. Spurs lost, but they were actually really good.
If Spurs play like this in future games, they are going to win a lot of them. More performances like this will definitely leave them in contention for a Champions' League spot. But Sunday's performance doesn’t happen without the allegedly inefficient and ineffective Eriksen. His passing was incisive. He consistently made his teammates look better. And his use of space is extremely important for a Spurs team that too often makes themselves easy for opponents to mark.
This might be controversial, but based on the rarity of that type of performance and how he's performed over his career, Christian Eriksen is quite possibly one of the best attacking passers in the Premier League already. Who somehow has only started league 5 matches so far this season.
If AVB ever gets his lineup selections right - which includes bringing Adebayor back from Siberia - this team could be almost unstoppable.
Secret Bonus Stat For Those Who Made It This Far
In the list above where I noted guys who had 9+ KP games, I actually left one out. There was one more guy who produced a feat like that last season in the big 5. Spurs actually already own him.
Erik Fucking Lamela.
6 open play goals in 11 matches! Lamela can’t get on the pitch in a Premier League game!