Behind the EPL Scenes: Quick Analysis Using xT Metric and Football Event Data
In today’s analysis, we delve into the ratings of the best EPL players, leveraging the Expected Threat (xT) metric. To simplify, we can view xT as an indicator of the change in the probability of scoring a goal resulting from moving the ball, either through a pass or a carry.
I’ve provided a detailed explanation in one of my previous articles but will offer an visual example here for clarity.
We divided the pitch into zones, with each zone representing a different probability of scoring a goal. When you move the ball from Zone A to Zone B, you alter this probability. This alteration is quantified as xT.
After a game, all xT values for a player are summed to arrive at a final xT score. Finally, we aggregate all xT values across all minutes played to calculate xT per 90 minutes. This estimation is what we use in our rating system.
Let’s begin by considering the Top 25 players at this moment, ranked by their total xT from passes and carries
What we observe:
🔵 Manchester City
- 4 players in the ranking.
- 3 out of the 4 are in the TOP 3!
- Doku and Rodri show anomaly values in the threat they provided (more details will be seen in the next graph).
🔴 Arsenal
- 4 players in the ranking.
- Martinelli — second place in the EPL by xT from carries (see separate ranking for the best carriers below).
- Zinchenko — in the TOP 3 by threat from passes (see separate ranking for the best passers below).
🔴 Liverpool
- 3 players in the ranking.
- Players in the lower part of the ranking (Salah and TAA ended in the TOP 20, Elliott in 25th place).
- Trent Alexander-Arnold — second place in the EPL by xT per passes.
🔵 Chelsea
- 3 players in the ranking.
- Sterling closes the TOP 3 of the best carriers in the EPL (see separate rating for the best carriers below).
⭐️ Sinisterra is in the TOP 5 but might not have made the ranking with a stricter filter. His 613 minutes played are the lowest in the ranking. For comparison, Bruno has played 2232 minutes over 23 rounds, equating to 24.8 full 90-minute matches.
⚡️ Ogbene, showcasing his talent at Luton (currently positioned 17th in the league), has distinguished himself in the rankings with high xT values from dribble advancements. Despite Luton’s overall standing, he individually excels, ranking among the TOP 5 in the EPL for this particular metric.
I have transferred the ranking to a scatter plot graph. Each player is represented by a point in terms of the danger created through passes and carries per 90 minutes of match play.
Players from the TOP 25 are highlighted in bold. The axes represent the league’s average values. Depending on which quadrant a player falls into, the color of the point changes.
This allows us to clearly see the anomaly values of Doku and Rodri from a different perspective.
Almost all of the TOP 25 have predictably landed in the top right quadrant. However, there are exceptions — Trent and Bruno primarily create through passes, while Ogbene, on the other hand, does so through carries.
The following shows a team-based aggregation of metrics. When looking at the efficiency of moments valued through xT, Manchester City’s metrics significantly stand out among other teams, showing a balanced performance. If not for Doku’s phenomenal statistics, Manchester City would be much closer to the cluster of Arsenal, Chelsea, and Tottenham.
Liverpool’s significant lag behind the same area of Chelsea, Tottenham, and Arsenal is concerning. For successful participation in the championship race, it’s crucial to have versatility and the potential to create threats both through team play and individual actions.
In conclusion, separate ranking of the TOP 25 footballers by xT for passes is presented.
Similarly, a ranking for the best ball carriers is provided.
P.s.
If you’re interested in a code breakdown for creating such visualizations, calculating metrics, and working with football event data, please let me know in the comments. In future works, I will be posting examples of code and implementations.
