If you’ve been getting stuck into some football coverage of late, you can’t help but have noticed it. Everywhere you look the talk is about Expected Goals (xG). It has featured on Match of the Day, on Sky Sports’ coverage and the vast majority of broadsheet newspapers – and some tabloids – have included features about xG. But what the heck is it?

At some point when watching a game, you may have heard a commentator say something along the lines of ‘….and he scores those chances nine times out of ten.’

Now, that’s just a throwaway comment from somebody that is paid a lot of money to waffle on about the beautiful game. But actually: is there something in that? Can we really give a chance created a kind of score that dictates how often it would be scored?

The answer is yes we can, and that is the genesis of Expected Goals.

The Basics

It goes without saying that a striker with the ball at his feet in the six-yard box has a far greater chance of scoring than a midfielder attempting a snapshot from 30 yards out. But how much greater?

That is what xG measures. This is a visual attempt to describe how often a certain chance created is converted into a goal, using historic data that gives a probability decimal based upon the likelihood of the ball hitting the back of the net.

It can be displayed graphically, and will typically look something like this:

So Napoli won this game 3-1, and the goals scored are highlighted on the graphic in pink. As for the rest of the squares, the basic rule of thumb is that the bigger the square the greater the likelihood of that particular chance ending in a goal. If Team A creates more ‘big squares’ than Team B, they can expect to beat them nine times out of ten (there’s that saying again!).

As you can see from the image, the Expected Goals count of both teams is given an overall numerical figure, which in this case is 4.7-0.8

What it Tells Us

In the example above, Napoli won the game comfortably by three goals to one and also ‘won’ the Expected Goals count by some 4.7-0.8. So they created more chances than their opponents AND scored more goals….that figures.

But actually, one of the best things about xG is that it can tell you when a team has been ‘lucky’, e.g. won a game but created far fewer chances than their opponents, and vice versa when a team has been unlucky. Extrapolate this data over a longer period of time and we can predict future form patters.

A great example already from the 2017/18 English Premier League season was the Chelsea vs Burnley game. Now, you may recall that Burnley shocked the footballing world by beating the reigning champions 3-2, and that Chelsea had two men sent off in the process.

But what you may not have known is that Chelsea actually won the Expected Goals count that day. So while their fans might have been in the doldrums, and Burnley’s on cloud nine, the truth is that there was more than an element of luck in the Clarets’ victory. The fact that Chelsea beat Tottenham in their next game, and Burnley lost theirs, will have come as less of a surprise to xG followers than traditional football fans.

Its Use in Betting

Studying xG actually gives you an advantage over the odds compilers at the bookmakers. Why? Because you can predict when runs of form – both good and bad – are likely to come to an end.

For example, Ipswich Town currently top the English Championship with four wins from four. Impressive stuff you might think? Well….what if we told you that the Tractor Boys had lost all four of their matches as far as Expected Goals is concerned!

You might come to one of two conclusions: Ipswich’s attacking play is super-efficient (unlikely), or that they have been lucky so far in a small data set, and over the passage of time they will regress to their normal level.

So we can ignore the bookmakers’ prices, which will closely reflect the league table and instead focus on this notion: sooner rather than later, Ipswich’s golden run will come to an end.

Here’s another working example from this season. On Saturday, Coventry played Newport County at home. Here’s how the two teams shaped up heading into the fixture:

Coventry City

  • Game 1: Beat Notts County 2-1 at home. Lost the xG count 1.0-1.2
  • Game 2: Beat Grimsby 2-0 away. Won the xG count 1.6-0.8

Newport County

  • Game 1: Drew 3-3 with Stevenage (a). Won the xG count 2.8-1.5
  • Game 2: Drew 1-1 with Crewe (a). Won the xG count 3.5-1.6

So what are the takeaway points? Firstly, a draw would have been a fair result for Coventry from the Notts County match, which would have put them on four points in the table. Newport, meanwhile, have absolutely bossed their two games so far, and perhaps deserve to be on six points.

The bookmakers priced Coventry at 8/11 to beat Newport at home, with the visitor available at 3/1. So what actually happened?

You probably knew the answer without looking, but Newport beat Coventry 1-0.

Now this is just one example, and anything can and invariably does happen in 90 minutes of football. But punters are advised to get on board with xG – it might just make you a few quid this season!