Expected Goals – Explained
While the term “expected goals” (xG) has entered the football lexicon over the past few years, the concept has been well established for much longer. Analysts and researchers have been using the metric to assess the performance of football teams since the early 1990s.
What exactly is the meaning of xG in football?
So what exactly do we mean by xG in football? Fundamentally, it is a straightforward metric that relates the number of shots at goal with the number of goals scored. It applies to individual players and teams. It can be broken down into specific situations, such as xG for penalties, for corners, for open play, in fact, it can be broken down to even more granular basis such as distance from goal and shot angle.
Essentially, xG can help us understand the game if we use it correctly. But as a tool, it has its limitations, for instance, used by itself it isn’t a reliable prediction of results. What it does provide is an indication of the probability that a specific shot will result in a goal.
The xG model
In fact, there are several different xG models, but we won’t go into their subtle differences here. What is fundamental to all of them is that the most critical parameters are the distance from goal and angle of the shot (though there are various definitions of these). Other important factors are free kicks, through balls, headers, corner kicks, time of game, and so forth. A complete xG analysis might well include many additional factors.
Developing an xG model involves collecting a great deal of information over time and analysing it using various statistic methods to determine the xG for the multiple shots and field positions. For instance, we might want to assess the probability of a player who is 8 metres away from an open net, scoring a goal. The xG model assembled using 1,000 similar instances in which 800 goals were scored in similar circumstances might assign a probability of 0.8 that the player would score, but that would be unrealistic as many more factors need considering. How did he receive the ball? What were the weather conditions? How long had the player had the ball?
To put all those variables together, we need to use a method such as “logistic regression” to find a formula that takes account of all that information to compute a realistic probability of scoring a goal.
How useful is xG in football betting?
xG is a valuable way of assessing the current form of a team an individual player. But can it be used to improve our football betting strategies? Let’s look at what it can tell us:
- – How well a team is performing compared to its usual form
- – Individual player performances
- – How well teams play from specific situations such as corner kicks
- – The xG for both sides in a match
xG can help us put real numbers to these and increase our betting edge. It can also help with specialist markets such as corner kicks and over under markets. However, don’t expect that introducing xG into your betting strategy will yield instant results. You will only see the value over time. xG is a long term historical statistical measure and provides only a guide to future performance. Remember, many factors can influence the outcome of a game, so don’t put too much weight on xG without considering all the other factors also.
However, the more information you have, the more successful your betting is likely to be. The best way is to see if it works for you.
What are the limitations of xG in Football?
While xG is an insightful tool for evaluating scoring opportunities and team performance, it does have its limitations. It’s important to understand these to avoid over-relying on xG as a sole predictor of match outcomes or individual player abilities.
Firstly, xG doesn’t account for the full context of each play. For instance, it doesn’t consider factors like player skill level, defensive pressure at the moment of the shot, or external conditions such as weather, pitch quality, or stadium atmosphere. These factors can significantly influence the likelihood of a goal, yet they aren’t explicitly represented in the xG value.
Secondly, xG models vary, and each has its own approach to calculating probabilities, meaning xG values can differ depending on the source. Some models include only basic elements like distance and angle, while others attempt to incorporate more nuanced details, like shot speed or player positioning. This variability means that xG is not a standardised metric and should be interpreted with caution.
Another limitation is that xG is retrospective; it assesses the probability of scoring based on historical data rather than real-time conditions. This means that while xG is excellent for analysing past performance trends, it’s less reliable as a predictive tool. xG can indicate the quality of chances created in a game, but it doesn’t guarantee future results, as football is influenced by countless variables that can change from one match to the next.
Finally, xG doesn’t cover defensive strength or how well a team prevents scoring opportunities, which is a critical aspect of overall team performance. Metrics such as expected goals against (xGA) help to fill this gap but even combined, xG and xGA don’t capture every aspect of a game. For instance, a team could have a low xG but still win due to strategic defensive play or capitalising on rare chances.
Leave a Reply
Want to join the discussion?Feel free to contribute!