Expected Goals: The story of how data conquered football and changed the game forever

Product Information

Obviously scenarios like this reflects the limitations of expected goals. There are several factors that xG could not take into account in these scenarios : The result is a model that is even more accurate at reflecting goalscoring opportunities that has been used across international coverage of the Women’s World Cup 2023 such as on ESPN and Fox Sports. Teams are often judged by the quantity of shots that they have in a match, or indeed in a season. Media companies will show the stats for how many attempts at goal each side has taken in a game. The central premise of xG is that the quality of those shots is of equal importance as the quantity. Analysts can work out the number of goals that a team would have expected to score from a certain amount of shots of a certain quality. Similarly, analysts can work out which players have scored more chances than they would be expected to. I would have loved to heard more about defense and goalkeepers and how to measure their performance in terms of xG. Also, a bit more about how strikers such as Michu and Benteke were briefly overvalued due to one good season. This is mentioned once but not really explored any further. Understand the defensive performance of a team by assessing how effectively are they preventing the opponent team from scoring their chances.

There are echoes here of the Moneyball technique popularised by the Oakland Athletics in baseball (which is referenced several times) and Tippett draws a connection between it and xG via their facility to refine gauging a player’s ability to one number. Whilst the location of a shot forms the main basis of its danger level, other factors also play their part. Shots which come from crosses are considerably harder to convert than shots which take place when the ball is standing still. Whether the shot is headed, volleyed or hit from the ground also affects its chance of success. So too does it matter whether the effort is taken on a player’s weaker foot. Analysts account for a whole range of such factors in their Expected Goals models.This could be based on the number of games or shots taken but it’s also important to consider the number of actual goals against the expected goals. How xG can be used in betting and scouting The new approach relied on using the data to find undervalued players in the market. It was a massive success and helped Brentford achieve five successive top half finishes despite having one of the division’s smallest wage budgets and fanbases. Meanwhile, Brentford even promoted to the Premier League. There is little attempt to put forward counter arguments, such as a traditionalist's view that not all forward passes etc. are created the same or that there are things data can not show you. These are noted as other elements in the process but the tone is overly evangelical on behalf of the advocates of data science. However, some players have still consistently outperformed their xG over the same period, most notably Lionel Messi, Harry Kane and Sergio Aguero. If you ever want to assess a player’s skill in relation to the average player, then it’s important to use a large enough sample size to omit any variance. Expected goals is one of the first advanced metrics to become widely known among general football fans and so it has inevitably faced its critics over the years (see Jeff Stelling in 2017). A battle between the traditional way of viewing the game and the upcoming world of data analytics. However, before we pass our judgement, it is important to understand how the metric works and how we should be using it.

The increased exposure which the metric has seen over the last couple of years is simply a drop in the ocean of what is to come. The reason why is simple: Every footballing judgement ever made is based on an analysis of the performance of teams or players. And the Expected Goals method offers by far the most advanced, profound and accurate gauge of performance. Expected goals (or xG) measures the quality of a chance by calculating the likelihood that it will be scored by using information on similar shots in the past. We use nearly one million shots from Opta’s historical database to measure xG on a scale between zero and one, where zero represents a chance that is impossible to score, and one represents a chance that a player would be expected to score every single time. As a data scientist, math, numbers, algorithms, and the hidden patterns woven in data, have pretty much guided my understanding of the world around me and given me a fulfilling, comfortable career. But with data now widely available, clubs can create a shortlist of players who have statistics that fit a particular profile, all without having to leave the training centre. xG is a metric that goes beyond the traditional shot counts, but it is important to remember that it is still just a metric. We can use it to evaluate underlying performances, but it is actual goals that are going to win you football matches.

For example, in the 2021/2022 UCL Campaign, the final had Real Madrid accumulate 0.73 xG and Liveprool had 2.9 xG. This is roughly the same value across many models published online (they generally only fluctuate about 0.2~0.3 xG between models). This means that Liverpool was expected to have scored almost 3 goals for the shots that they've taken. xG is a metric which has launched itself during the past years into mainstream soccer terminology with a boost from the clubs that are treating data and analytics serious. It’s being used by media, fans and also sports betters. The xG model is only as good as the factors being input into its calculations. These data inputs are limited by the data we possess today from companies such as Opta. Other factors, such as shot power, curl or dip on the shot or whether the goalkeeper is unsighted or off balance might not be considered in most xG models out there. Due to model being based on averages, the random nature of a football match and the rarity of goals in the sport makes it almost impossible to consider with enough statistical significance all historical factors that can cause a goal to be scored. xG should be used as indicative and supportive information for decision making purposes and generating opinions rather than a finite answer to the performance of a team or player.