What is Expected Goals (xG)?

Football has changed massively in recent years. On the pitch, new tactical developments such as the rise of gegenpressing and the increased popularity of zonal marking have made watching football a different experience, and off the pitch, coverage of the sport has altered dramatically too. Perhaps the biggest change of all is the increased importance of data within football, and when we watch the sport on TV these days, that's something that's clear to see. The statistics and visuals that flash on our screens during and after games, and the figures mentioned by pundits and managers alike when discussing results and performances add a whole new dimension to football.

But while most fans have accepted these changes, there's one aspect of the data revolution that has been a little bit controversial at times. We're talking of course, about Expected Goals, or xG. Expected goals plays a major role in tactical discussions now, but there are a lot of people that get confused or even irritated when this metric is used, and it's fairly easy to see why.

That's why in this article we'll be giving you an in-depth explanation of the expected goals model and the role that this metric plays within the game. We'll answer the question "What is xG?", we'll explain how xG is calculated, and we'll also talk about the benefits and downsides of Expected Goals. Once we've finished, you should be confident about what's going on the next time you see the famous letters "xG" pop up on your television.

What is xG?

It's hard to enjoy football these days without coming across Expected Goals one way or another, but often people will see the stat come up without really taking in what it means. So what exactly is xG?

Essentially, Expected Goals (xG) is a metric that's intended to measure the probability of a shot resulting in a goal. The purpose is to show when a player should be expected to score from a particular opportunity, by basically rating how good of a goal-scoring opportunity it is.

You might be slightly confused about how the quality of chances in a football match can be determined from a statistical point of view, and this is a completely valid concern. What the xG model does is use historical information regarding similar chances and shots to create a picture of how likely a goal is to be scored. In order to really explain this, let's spend some time digging into the details of how the metric actually works.

How does Expected Goals work?

xG is calculated using a serious deep dive into the past. To measure the likelihood of shots being converted into actual goals, xG uses historical information from thousands of shots with similar characteristics to estimate how likely a goal is on a scale between 0 and 1. 0 would mean that on average, if 10 shots with similar characteristics had been taken in the past, none of them had gone in, whereas 1 would reflect that all of those 10 shots had gone in.

An xG of 0.3, for example, would mean that out of 100 similar attempts, 30 had been converted, so we could expect 3 in 10 of these types of attempt to go in.

In many ways, xG is simply about reinforcing things that we can already see on the pitch. For example, even the most casual of supporters can tell that a shot from six yards out in a central area is more likely to go in than a 40-yard effort from the left wing. What an Expected Goals model does is provide a statistical framework so that it's possible to systematically work out the xG value, aka shot quality, of any given opportunity.

One thing that's important to note is that different xG models can be used by different organisations and competitions. Each model has its own characteristics, although they generally all rely on the same major factors: distance to goal, angle to goal, body part with which the shot is taken, plus the type of assist or prior action (eg. cross, through-ball, set-piece, short pass, dribble etc.) Models use all the information they have on shots with similar characteristics to come up with a mathematical value relating to how much a player would be expected to score the relevant chance.

This way of statistically figuring out the likelihood of shots being converted into goals has already made serious headway within the game, due to several key benefits. In the next section of the article, we'll briefly take you through some of those positive aspects of the Expected Goals metric.

The Benefits of Expected Goals

The fact that most of the world's sports data experts and tactical masterminds have embraced the power of xG shows that it can be extremely useful. Here are a few of the key benefits that Expected Goals can provide for a coach, analyst, player and team as a whole.

• xG shows us the importance of shot quality by emphasising the low chance of scoring from disadvantaged positions.
• It allows coaches and their backroom staff to identify particularly good finishers. Most players will generally convert chances at an average rate, but xG helps shine a light on players who are operating at a higher level.
• Expected Goals also highlights the fact that crosses aren't the most effective way of scoring. Generally speaking, not many goals are scored from this type of chance.
• xG is also great for team analysis. It can be used to give a stronger idea of underlying team quality, providing a context for recent results and highlighting whether a team is over or under-performing their expected numbers. This leads us on to the next point...
• Predictive potential. Using xG to analyse performances and results on a deeper level can help coaches, pundits and journalists predict where results may soon begin to change, and why. If a team is out-performing their xG, they can generally be expected to experience a drop-off in results further down the line when they stop getting the rub of the green.
• Expected Goals can also be extremely useful for scouting purposes. It's used during recruitment processes due to its ability to accurately judge the finishing skills of certain players.

The Weaknesses of Expected Goals

Expected Goals can be extremely helpful when used correctly; however, due to the fact that this is a very new form of data within football, it's not always taken advantage of in the right way, and many people are still unclear about exactly what xG can offer. Here are a few key weaknesses that are worth keeping in mind.

• Using xG in relation to a single game can be misleading and ultimately useless. This is due to the relatively small sample size - ultimately, individual football matches are all completely unique, and any number of random, unpredictable events can happen, so using xG to describe single game events isn't always particularly useful.
• It's often used as a descriptive metric rather than a predictive one. For example, xG numbers will pop up alongside other statistics at the end of the game, where it doesn't offer much. Where xG is most useful is for clubs in preparation for matches during the broader context of a season. Looking at how they're performing and what the underlying numbers are can be very helpful in this regard.
• There are limitations in terms of the data available. For instance, there's a lack of information on the exact state of play when a shot is taken. As the years go on and the data continues to get better, these limitations will gradually be removed or at least reduced.

You'll probably have noticed that most of these weaknesses relate primarily to the implementation of an Expected Goals model rather than the metric itself. Ultimately, it's still early days in the use of xG in football, so it may take some time for people to have a better understanding of how it can be harnessed effectively.

Data in Football

The role of data in football is set to continue expanding as the game moves forward. It's highly probable that those within the world of football analytics will soon come up with new forms of measuring chance creation, set piece efficiency, defensive performance, and other aspects of the game, and football parlance will develop even further as a result.

After all, clubs in the Premier League and other elite divisions are harnessing analytics and data more than ever before in order to boost recruitment and operate as efficiently and effectively as possible.

Due to its increased role in the game, data is something we've covered on the Jobs in Football Blog before. So if the world of data in football is a topic that interests you, make sure you check out our article on 7 easy steps to get started in football data and analytics. And for some information about a group of clubs that have used the power of data to their advantage, head over to our deep dive into the Red Bull Philosophy to find out all about how this well-known franchise operates within the game.

About The Author

Fred Garratt-Stanley is a freelance football writer, Norwich City fan, and amateur footballer for South London side AFC Oldsmiths. For Jobs in Football, he's covered subjects including the rise of set-piece coaching, the role of xG in football, and the growth of tactical ideas like gegenpressing and zonal marking. He's also written about football for publications such as British GQ, VICE, FanSided, Football League World, and more.