A statistical comparison of Willian and Wilfried Zaha

There’s been some chatter about Wilfried Zaha and Chelsea over the last couple of days, which I’m really hoping is smoke and mirrors, as I’ve never really been a fan. Sure, he dribbles a lot, and he runs fast, but he’s never even been able to excel at Crystal Palace, let alone be worth top dollar — £80m rumored asking price! — for a top club.

The other half of the argument is that (supposedly) he would be an upgrade on Willian, with which I also disagree. And I’m not really a fan of Willian all that much either, though I recognize the quality he does bring to the table. Do I think Willian is good enough to be a starter for Chelsea? Nope. Haven’t for years. But do I think he’s better than Zaha? Yup.

I tend to go right to data when making comparisons, as the “eye test” skews the debate through inherent bias. As always, I started with understat.com data, and their cute little radar diagrams. I’m normally anti-radar, mainly because they’re often misinterpreted in analysis, but it’s not the absolute worst option for a comparison. First, let’s take a look at the last five seasons combined:

Understat

From a per-90 perspective, Willian is shows better in every category except Goals and Expected Goals, where the two players are very nearly identical.

So let’s take a look at just last season:

Understat

Last season, Zaha definitely scored at a higher clip, although the Expected Goals were dead even. And everything else is still higher for Willian, with the KP90 (passes that lead to shots per 90) much, much higher. Score some points here for Willian.

Moving beyond the radar, let’s dive into a detailed look at shots and key passes from open play, to try to negate any benefit Willian might have had from free kicks and corners:

data sourced from whoscored.com

I highlighted the totals where the better metric was, but obviously each season has its own values as well. Zaha is definitely the better shooter and scorer in open play, but he’s not the better creator.

That said, there’s a lot more to football than just shooting and creating. There’s all the dirty work, the little bits, the ‘in-the-trenches’ work. I tried to capture what seemed relevant from whoscored.com for this dataset, such as defensive actions, passing, dribbling, and so on.

data sourced from whoscored.com

So, let’s break these down a bit, shall we?

• Zaha dribbles more, but at a much worse success rate, which is really a bad thing
• Zaha is a far worse passer, and at a much lower volume as well. It’s tough to ignore how good of a long ball passer Willian always has been. His ability to switch the ball consistently is a big benefit.
• Their on-target percentage for shooting is nearly identical, 36.0% vs 36.8%. Zaha gets blocked more, which is shocking.

I created a couple of metrics to try to show some overall perspective: Offensive Touches and Bad Offensive Touches.

• Offensive Touches = Dribble Attempts, Total Passes, Unsuccessful Touches, Times Dispossessed, and Total Shots. This is pretty much 99 per cent of what can happen with the ball.
• Bad Offensive Touches = Failed Dribbles, Inaccurate passes, Unsuccessful Touches, Times Dispossessed, Off Target Shots, Shots hitting the woodwork, and Shots Blocked. Essentially when something positive didn’t happen.

This is where it gets interesting. Willian is only at 2.5 Bad Offensive Touches per 90 less than Zaha (14.4 vs. 16.9). But also notice that Willian’s Bad Touch Ratio is 24.56%, about 1 out of every 4 touch, to Zaha’s 43.45%, nearly half!. The initial counter argument might be that Willian makes a lot of passes, including short passes, and this is true. But Zaha also has twice as many UT+DISP to Willian’s, in a lot more minutes, and has just an absurd amount of failed dribbles. AND his passing percentage is worse. Whatever the circumstances may be, Zaha does something BAD with 43 per cent of his touches. U - G - L - Y.

• The other metric I came up with is Defensive Good Touches. This is simply Tackles Won, Interceptions, Passes Blocked, Crosses Blocked, and Clearances. They are wingers, so it’s not much, but Zaha does contribute over 3.5 of them per 90 (as compared to Willian’s 2.8). Top sides do expect everyone to defend, not just the back seven.

Now, one of the things that Zaha has going for him is that he’s young, or at least younger than Willian. I don’t really consider 27 (next year’s age) all that young, but so be it. So perhaps, if we look at the details, we might see that Zaha is actually improving with time, and we could expect better numbers? Nope. Not really. Here’s a breakdown of some of the ‘bad things’ at a per-90 level, and the only one improving is the number of failed dribbles:

data sourced from whoscored.com

The other argument, which isn’t really possible to validate or invalidate through publicly available data, is that by playing for a better club, Zaha should ‘by default’ look and play, well, better. But this seems to be more of a case-by-case situation, and most established players don’t magically become superstars by playing with superstars, even if they might improve marginally (see also: Danny Drinkwater).

Also, let’s not forget that we already have Christian Pulisic, we already have Mason Mount, we already apparently have Callum Hudson-Odoi (any day now, right?), along with Ruben Loftus-Cheek, Ross Barkley, etc. Whom are going to bump for an £80m player? What would be the point?

To sum up, I can’t think of a single reason that this transfer would make sense. Zaha is not good enough, his fee is too high, he’s too old, and thee’s no room for him in the squad.