So I saw a link to this today:
This part:
“ATTACKING MIDFIELDER: Willian
Criteria: Shots on target and key passes “
And specifically this part:
“Of course we wanted this to be Maddison, Sterling or Christian Eriksen, but it turns out that Willian creates a shedload of chances (it helps if you take the corners) and takes a fair few shots too. That he has claimed just three goals and six assists in the Premier League this season is a mystery only partially solved by WhoScored’s description of his weaknesses being simply ‘finishing’. Still, Barcelona have been chasing him for years for a reason. “
Now, I have to question this on several levels:
- Why not just pull out the corners then? What does taking corners have to do with playing the Attacking Midfield position?
- Why not include some quality metrics, like xA? I know Understat.com is easy to check, as I did….
- Shots on Target…. and then mentioning Raheem Sterling among others, you know, he of a 50%+ on target rate, and of 10 assists in league play to go with his 17 goals, and I mention that…
- That Willian’s only got 3G/6A is no mystery at all. He sits on the ball, he waits too long to shoot or pass, and he often ends up ruining good attacking plays more than he generates them. When Maddison is on 7/7, Eriksen is on 8/12, and Sterling is on the above-mentioned 17/10, the absurd notion of even considering a player on 3/6 is just outright ludicrous.
I decided to take a look at these four players, since the article writer mentioned those other three specifically. And I looked at two specific items:
- Key Passes – including quality, and looking at Open Play
- Shots on Target – An actual look, not just a casual mention with no real data or anything to justify the selection whatsoever.
Here’s a nice little set of tables showing all the metric breakdowns, and using some quick heatmap coloring to make it all the more obvious why Willian should never have been this selection:
Now, for those not sure, OP = Open Play, XA = Expected Assists, and all of the shot data columns should be clear enough. The overall numbers are up top, and it’s interesting, the other three are all playing nearly identical minutes, but poor ol’ Willian is much lower. So to help clear up any misconceptions, I also broke down in the second table, the Per90 numbers for consistency. Apples to Apples, no matter how rotten one might be.
The last table is a quick peak at KP quality, using the XA data, and also Shooting percentages. As can be seen, the only green in any of Willian’s rows is on his initial artificially inflated (which the article even mentions!) KP/90 number. Because he takes all of our hilariously absurd corners (how many fingers this time, Willian? I see one…)
Sterling is the clear and obvious winner here, and his exclusion is either an attempt at angry page views and clicks (well played if so), or a horrible attempt at objectivity, failing completely.
I would hope football sites that are going to do things like publish team of the year articles would do a better job of selecting players, if they are going to bother actually listing metrics they allegedly used for the selection.
