Chances Aren’t Enough – Ones Too Good To Miss

I don’t know about you, but back when I was a kid, growing up, Air Supply was a pretty big band, especially when it came to mushy slow dance songs at school, or on the radio, etc. And their song “Chances” is the theme of the day.

Chances aren’t enough

Ones too good to miss

Chances aren’t too strong

A chance is all there is

Air Supply – Chances

Those lyrics are spot on. We talk about chances all the time when discussing football/soccer, and with respect to the quality of those chances, and also those players who are best at creating those chances. The ones that are too good to miss, especially.

Back in January I wrote up a brief article on chance creation quality, using some heat maps, taking a look at the top chance creators in the PL, and also looking at Chelsea. I’m going to look across Europe this time, and use a bit of a more visual approach, and try to break down why chance quality is so important, mapping out the progression along the way.

Assists and Key Passes

Let’s start with the actual results, the things that happened on the pitch, first. Players make Key Passes (KP) and some of these become Assists (A). I wanted to look at the top Assist creators in the Top Euro leagues, and cut off the list at players having a minimum of 8 assists in league play, which also conveniently happened to be a set of 25 players. Data fields I imported were as follows:

• Apps – Not really relevant for the analysis, but still interesting to include (Whoscored.com , WS)
• Minutes – for calculating per 90s (WS)
• Assists – Also by type, for curiosity (WS)
• Key Passes – Also by type (WS)
• Expected Assists (xA) – for expected assists values (Understat.com , US)
• Open Play Key Passes (OP KP) – for tracking open play data (US)
• Open Play Assists (OP A) – For tracking open play data (US)
• Open Play xA – For tracking open play data (US)

Multiple calculations are included as well for analysis, and I’ve included a full table at the end of the article for those interested.

So let’s see how these players do in terms of the actual results on the pitch, the KP and the A. I’ve plotted Assists per 90 vs Assists Per KP below:

A few things obviously stick out:

• Jadon Sancho has a very high Assists per 90 output. Leroy Sane does as well.
• Mertens and Di Maria I would not expect to be right there with Messi either, but there they are.
• The ever-beloved Eden Hazard is next on the list, right above Brandt and Eriksen and Haller.
• What’s odd, though, is that Messi and Hazard both have relatively low (at least in this group) A/KP numbers, meaning they are needing more Key Passes to get each Assist they have.
• And some players have very high A/KP values, meaning they are creating more Assists per Key Pass, and there’s some odd names over there, like Pogba, and Haller, for example.

The thing to be careful of here, and this is why we are going to drop down into the realm of ‘expected’ values shortly, is that these are all dependent on their teammates actually scoring from the chance that was created. Many will say that “well, Eden sets up Morata, and Morata misses the 1v1, so it’s not Eden’s fault!” for example. Which is possibly accurate (probably in those cases).

Expected Assists

Thanks to the folks at understat.com who provide their numbers to the public, I have some expected assist data to play with. I have to, as always, make the assumption that this data is accurate, but so far it’s not led me wrong.

One of the first things to look at, so we can set a proper mindset, is how expected assists, or xA, compare to actual assists, or A, for the players. Who is performing above expectations, and who is below? As usual, I’ve got a chart for this:

Some observations:

• The vast majority of these players are performing ABOVE expectations, as should be assumed for a list of top performers (averages and all that)
• The list of those performing below expectations, is an interesting one. Messi and Fraser are near the top of the list in terms of Assists, and Sterling in the middle, Suso and Depay and Pepe all at the bottom. So at least for this dataset, there’s no apparent correlation here between the actual and the expected, which is good, right? Kind of the point.
• It is telling, though, that three of the top four with respect to xA/90 (Sterling, Fraser, Sane, Messi being those top four) are under-performing.
• Normally you could try to equate an A to xA ratio under 1 to be due to having bad shooters on your team, or just being a not-great team, thus the output isn’t there. But that doesn’t explain Sterling and Messi, who play for the #2 and #3 teams in terms of league goals, respectively.

Now, let’s have another look at that chance quality chart, but with expected assists this time, instead of actual assists:

Some Observations:

• Messi really stands out here now. He’s no longer looking poor in terms of chance quality, and he’s through the roof with respect to his xA/90 number. It’s not even close to anyone else. In the first chart, Messi was easily in the bottom 40% in terms of A/KP, but here he’s now in the top 20%. He’s a great point to build off of, because obviously everyone knows his quality.
• Notice that once again, Sane and Haller are high in terms of xA/KP, just like they were with A/KP. Same with Kostic. But notice that Kimmich and Sterling are pulling ahead of the pack. As high as their numbers are, they probably should be even higher.
• Sadly Eden Hazard is still in the back of the pack, worse actually than his A/KP numbers. He’s in an interesting cluster, too, with Brandt, Sancho, and Suso. These are ‘volume’ guys, as I like to refer to them. They see the ball a lot for their clubs, and they have a high KP number, but not a very high xA value. What’s also interesting, that beyond Sancho, the other three are all in similar situations with their clubs. Milan is in 4th, Leverkusen in 6th, and Chelsea in 6th. And all three are struggling to score goals as a club too. Perhaps asking too much of their star power, or perhaps simply the players trying to do too much on their own to overcome roster deficits. Would need further analysis.
• I want to take a moment, too, to point out that only one club has two players on this list, if I recall correctly, and that’s Manchester City. And neither of them are de Bruyne, or the Silvas. What a luxury to have two wingers both so capable of creating goal scoring opportunities. #RIDICULOUS

Open Play

Now, chances created are well and good, and the quality of them is also important, but there’s one more distinction I want to make here, and that’s with respect to the state of the game when these chances are created. Every club needs a good free kick and/or corner taker. I get it. And I agree. But to me, that clouds this discussion a bit, or at least it potentially does. What I want to see is how these numbers play out in Open Play, and thankfully I’ve got data for that, too. First let’s look at how their OP A compares to their OP xA:

It’s the same folks on the list below the line, for the most part, Fraser, Sterling, Messi, and this time Jony as well. To me, this is a good thing, as it means that these quality chance creators aren’t being reliant on free kicks and corners to drive their numbers up. Some of them definitely get an overall boost in terms of actual chances created and assists, but their expected open play results are pacing just as well, and that’s the least we can ask.

Let’s crack open the xA/KP vs xA/90 visual:

Some Observations:

• Notice how even though Messi is still in a league of his own, Sterling is starting to catch him a bit in terms of xA/90. I’m assuming it’s because he doesn’t take the corners or free kicks, but still. At this point, I’m putting this data in my pocket for the next time I have to hear “Sterling is Overrated” nonsense. These numbers are unreal, and are new to his game, above and beyond the incredible scoring numbers he’s been putting out. Lest we forget, he’s sitting on 15G/9A with 8 matches left to play, following an 18G/11A season last year. And he’s only 24.
• Sane, who by the way is also playing at a ridiculous level, on the opposite wing of Sterling, is only 23. Unfair. Second highest chance quality mark in the group.
• Should we even talk about the fact that Manchester City are the ones who SOLD Jadon Sancho? #SMDH
• Kostic – Who? I’m going to get to him in a bit.
• It’s no wonder we tried to lowball a bid for Sarabia, which not-shockingly he turned down. Good creator.
• Can someone please explain to me what the story is on Ryan Fraser, and why everyone isn’t after him? I don’t care if he plays for Bournemouth and is probably piss-poor defensively, he’s basically a young Aaron Ramsey at this point, right? Maybe this season is an anomaly, but you’d think someone is looking.
• Haller doesn’t get nearly enough interest in my opinion. Sure, it’s tough playing with Jovic, who’s gone nuts, but Haller’s a big part of that reason. I’d love to have him at Chelsea. Scores, creates, young, would be a very incredible replacement for Giroud IMO.

Filip Kostic. He’s not remarkable or anything, I’ve done some digging on other metrics, and a bit historically after seeing this ridiculous numbers. This is what Filip Kostic is: A cheap, non-HG alternative to Barkley, if Chelsea were looking. He’s on 3G/8A in league play for Eintracht Frankfurt this season, playing at LCM, and he’s got another 3G/1A in the EL. Historically he’s in that 4-5/4-5 range for a few Buli clubs, and seems to be coming into his own this year.

His passing isn’t great, which is a black mark, but he’s obviously a good creator, can score a goal or two, and is a solid defender. Comfortable with crosses, long balls, tackling, etc. He’s actually a bit of a poor man’s Kovacic, too, just with scoring/creating ability. If we are able to sign players this summer, and decide not to keep Kovacic, I know we probably won’t look at Kostic because we already have Barkley, but he’s not a bad rotation option. And even though he just signed for Frankfurt this season, he probably wouldn’t cost more than €25M. Honestly if Spurs end up losing Eriksen he could be an interesting target for them, too. Or Arsenal for that matter. I think he’s going to come good. He’s playing above his station right now, that’s for certain.

I want to also point out, that this is why looking at the details, and looking at expected numbers is so important. Certain players tend to rise to the top, like the cream, and other players tend to sink a bit. It’s not just important to create a LOT of chances, it’s important to create GOOD chances.

For those that want the data table, here it is. Enjoy!