For starters, this analysis uses data from December 28, 2024. Teams have played either 17 or 18 matches as we’re approximately halfway through the Premier League season.

For this post, I thought it would be interested to explore the expected goals (xG) statistic for both squads and individual players and compare that against the number of goals that have actually been scored. However, for starters, I’ll explain exactly what xG is.

What exactly is xG?

For those who don’t know, said simply, xG is the probability that a shot will result in a goal. However, there are several characteristics that are taken into account when determining the probability of a shot resulting in a goal. Those characteristics include items such as:

• Where was the shot taken? Angle of shot relative to the goal and distance from the goal.
• Type of play leading up to the goal. Was it from a live play, such as a through ball or cross? Or maybe a dead ball, such as a free kick or corner kick.
• What was the last body part to contact the ball? Was it a header or off the shooters foot, did it rebound or take a deflection?

Characteristics like the ones above, along with several others, are taken into account and compared against a database of thousands of others with similar characteristics. Based on this comparison, a probability is determined to see the likelihood that the shot in question will result in a goal.

For instance, an xG of 0.7 would indicate that 7 of 10 identical shots would result in 7 goals. Similarly, an xG of 1 is a definite goal, while an xG of 0 is a definite miss. At the end of a game, all those individual probabilities are summed up, for both the squads and individual players, which provides one number outlining the total expected goals for the respective squad or player.

All data used for this blog post is from FBref and the model they use is provided by Opta. In addition to typical characteristics outlined above, Opta’s model also considers additional items which factor in defense and goalkeeping, such as:

• Amount of pressure the shooter is under from defenders.
• Clarity of the shooters path to goal.
• Position of the goalkeeper.

It should also be noted that xG does not take into account the quality of players in the play. Rather, it is an estimate, based on probability, that considers how an average player would perform in a similar situation.

Squad Expected Goals

When looking at Premier League squads, there are several comparisons that we can make. We’ll focus on the following two:

• Total number of expected goals per squad.
• xG per squad compared to actual goals scored.

Total Number of Expected Goals per Squad

Top 5

As of the date of the data, December 28, 2024, Liverpool is leading the league with 38.2 expected goals. That is 1.7 xG more than Chelsea, who are second with 36.5 xG. When looking at the top 5, the remaining teams include Bournemouth at third with 35.4, Manchester City at fourth with 32.8, and Tottenham in fifth with 32.7.

Keep in mind, at the time of writing this article, Liverpool is the only team in the top 5 that have played 17 games, the remaining four teams have all played 18 games. Given Liverpool’s quality so far this season, it is expected that the xG gap between first and second place would grow when Liverpool have played equal number of games.

Botom 5

At the other end spectrum, Ipswich Town have the lowest xG in the Premier League with only 15.9 expected goals. This is followed by Everton, who are marginally better with 16.5, Leicester City with 17.5, Southampton with 18.5 and Wolves with 18.7.

Out of the bottom five Premier League squads when focusing on expected goals, the bottom two, Ipswich and Everton, have played 17 games. The remaining three teams have played 18 games.

Full Table

The table below ranks all Premier League squads based on their expected goals metric.

The top five teams average an xG of 1.98 per game, while the bottom five teams average an xG of 0.99 per game. Meaning, out of the top 5 teams when ranked by xG, they are expected to score an average of 1 goal per game more than the bottom 5 teams when ranked by xG. The average for the middle ten squads is 1.47 xG per game.

Next we’ll review how these expected goal metrics compare to actual goals a team scores.

xG per Squad Compared to Actual Goals Scored

Above, we discussed each squad’s expected number of goals scored, but how does that compare to the actual number of goals scored?

To determine this, we take each squad’s total goals scored and subtract their expected goals metric. A positive number indicates that they are scoring more than they are expected to score. A negative number indicates that they are missing chances that they are expected to score.

As seen below, the graph provides a visual interpretation of the goals – xG metric for each squad in the English Premier League.

This is a fascinating metric because it shows which teams are over or under performing, which teams are taking or missing their chances. Large positive spreads (difference between xG and goals) in this metric can mean a number of things such as outperformance (e.g., scoring more goals than other players in similar scenarios resulting in an above average scoring ability) or luck (e.g., scoring wonder strikes from improbable locations).

Similarly, large negative spreads can indicate the inverse. Examples include underperformance (e.g., scoring fewer goals than the average player in a similar scenario, resulting in a below average scoring ability) or bad luck (e.g., hitting the woodwork from close range or the goalie making a freak save from an otherwise for sure goal).

The full table, ranked by the goals – xG metric, is seen below. For reference, the table also provides each squads league position as of the data date (December 28, 2024).

Wolves: Leading the Goals – xG Metric

Wolves have the highest Goals – xG metric, with 10.3. As of the date of the data (December 28, 2024) they have scored 29 goals but only had an expected goal total of 18.7. This results in Wolves scoring 10.3 more goals than they were expected to be based on the probability model. This is quite a staggering achievement. For reference, second place is Brentford who have a Goals – xG metric of 5.9, which is almost half as much as Wolves.

Wolves’ Goals – xG metric is a bit of an outlier. And it is heavily linked to Matheus Cunha who is having a standout season and on an individual level, is leading the league in the Goals – xG metric. While we’re not going to focus too much on individual performance in this post, we will be writing an article in the future about which players contribute the most and least to their teams Goals – xG metric.

Bournemouth: the Worst Goals – xG Metric

At the other end of the spectrum is Bournemouth, who have the worst Goals – xG metric in the league at -8.4. They have scored 27 goals but were expected to score 35.4. Bournemouth are creating lots of goal scoring opportunities, but they aren’t clinical enough with their finishing.

At the time of writing this article, Liverpool and Chelsea are sitting first and second in the league and they have a Goals – xG metric of 1.8 and 1.5, respectively. This is the 8^th and 9^th highest ranking in the league, so they are both sitting approximately middle of the pack for this metric. They also have the two highest expected goals metric in the league (as noted above), outlining that both teams have players that create goal scoring opportunities, but they also have the quality to finish those opportunities.

Premier League Table vs. Squad’s xG Ranking

Finally, when looking at the same table above but adding the xG Position (ranking), it is interesting to see where the xG ranking is compared to a squad’s actual position in the Premier League table.

For instance, Liverpool and Chelsea are ranked 1 and 2 for both metrics, respectively. Nottingham Forest has the largest difference between the two rankings (3^rd in the league, but 15^th in xG), indicating that maybe their defensive attributes are playing a large role in their ranking.

Lastly, Bournemouth has the 3^rd highest xG in the league but are only ranked 6^th, imagine where they would be if they could convert more expected goals into actual goals.

Summary

That’s it for this week. In this post, we covered expected goals or its abbreviation, xG, which you’ll often see or hear. For starters, we explained that xG is the probability that an average player would score a goal from a similar scenario. To determine xG, characteristics of the chance in question are compared against a data base of goal scoring opportunities with similar characteristics. Based on the comparison, a probability is assigned.

If a team or a player is scoring more goals than their xG, it indicates that they are overperforming. While the opposite indicates they are underperforming (i.e., missing chances that an average player in a similar scenario would score).

After understanding what xG is, we focused on two related comparisons. The first was the total number of xG each squad has in the Premier League. The second was the xG per squad, compared to the total number of goals that squad has scored, and we provided an analysis about some over and underperforming squads.

Like any metric, there is more context needed to better understand what happened than simply a number on the page. In future posts, we’ll take a deeper look at xG on an individual player metric, so see which players are having a larger impact on their squads xG metric.

Furthermore, while xG is an important metric for determining performance related to scoring goals, it is also important to consider the opposing team’s defensive efforts. As a result, we’ll also take a look at xGA in a future post, which is expected goals allowed. This metric provides more insight on the defensive side of things and if squads are allowing more goals than they should be.

So, stay tuned for those future posts and send us a message if there is a specific topic you would like to see covered.

As always, thanks for taking the time to read these posts!

JC