Goalkeepers: The hardest position to analyze?

Data from Fotmob and Wyscout. Photo from Claro Sports.

Thought: Shot-stopping is assumed to be the most important concept to analyze in a goalkeeper’s game. However, if most keepers rank near to the average for preventing goals, and if their performance levels can change drastically from one season to the next, perhaps focusing on other aspects of goalkeeping is the best strategy for decision makers.

What if I told that the best goalkeeper this season is a 37-year-old that’s played less than 10,000 minutes in his whole Liga MX career? Well, if you think the most important part of a keeper’s game isn’t their quality with the ball at their feet or coming out to claim crosses, but is their ability to stop shots, Alfonso Blanco has been the stand-out in Clausura 2025.

With the goals prevented metric, we can put a goalkeeper’s shot-stopping ability into a single number. The metric compares the difference between xG post-shot, which takes into account factors such as the shot location, body part, defender positioning and the end location of the shot, against actual goals conceded. In theory, the goals prevented metric for an average keeper would be 0, not positive or negative.

The five best goalkeepers in the metric during Clausura 2025

On the surface, it appears an effective metric to analyze shot-stopping abilities. The method makes sense, the results are easy to understand, and it seems simple to use the metric in order to make future decisions. However, there’s an issue. The metric isn’t repeatable.

The scatter graph shows the correlation between a goalkeeper’s performance in the goals prevented metric, from one season to the next, from 2021 onwards. The correlation is just 0.08, very weak.

There’s much greater reliability in many key metrics for other positions, for example, progressive passes from midfielders.

Therefore, I can say that Blanco, Malagón, Guzmán and Unsaín have been effective shot-stoppers this season, and have helped their teams to success (all four play for sides that are currently in the top-six). However, I can’t predict whether or not they’ll continue to play at the same level next season. This complicates squad building decisions.

Do we need to analyze deeper?

One theory for the changes from season-to-season could be that keepers have specific strengths and weaknesses. One goalkeeper may have good reflexes, another may be better at reaching shots in the top corner, whilst there may be an impact of facing a shot towards their left or right side. Facing constrating types of shots in different seasons may affect the goals prevented metric, and more context over the shot types could perhaps improve reliability.

Goalkeeper xG is a company that provides specific keeper data, focusing on goal prevention given varying contexts. For example, a few seasons ago their model showed that Liverpool’s Allison was the most effective Premier League goalkeeper in 1vs1 situations. Under Jürgen Klopp, Liverpool played with a high defensive line, and perhaps signed Allison with the knowledge that their system was likely to concede a few 1vs1 chances.

Whilst I’m yet to see evidence that their metrics, or that a similar concept, creates repeatable results, the idea makes sense. Teams should consider the types of shots that they’re likely to concede within their style of play, and scout accordingly.

Normal distribution

A danger when analyzing players is to assume that there’s a linear distribution of talent, and create rankings of best, second best, third, fourth, etc…, considering there to be the same difference between each rank. However, the theory of normal distribution suggests that when analyzing talent within a population, the majority rank close to the average, with a few outliers.

The normal distribution “bell curve”

With an analysis of total prevented goals over a longer period of time, in the next graphic, we can see that the distribution is fairly similar to the normal distribution theory. 29 of 32 goalkeepers, with a minimum of three seasons with more than 600 minutes played, are in the orange zone, close to the average. There are three outliers; Jonathan Orozco ranking very poorly, with Guillermo Allison and José Rangel far above the average. It’s a surprise to see Allison ranking so well, whilst Rangel is the best young Mexican keeper right now. This graphic may suggest that Rangel has the talent to play in a stronger league, but he hasn’t yet played for that many seasons, and his performance levels have dropped during Clausura 2025.

With the majority of keepers close to the average over a larger sample size, it makes sense that we see fluctuations over the short-term (from one season to another). It isn’t likely that an average keeper will perform well-above, or well-below, average for two consecutive tournaments.

Furthermore, a larger sample size can be used to help make better decisions. Between the 29 goalkeepers in the orange zone, the best performer (on the right) would be expected to concede three fewer goals than the worst (on the left), across a 17 game season. Not an insignificant difference, although with the small sample size of a specific season, there could be major fluctuations in their performance levels, impacting team results.

Should teams focus more on other concepts?

If it isn’t possible to predict the goals prevented metric in future seasons with much confidence, and/or teams only have goalkeeper options that rank close to the average historically, perhaps it is better to focus on other aspects of goalkeeping.

Making saves may be the most important concept, but other factors do matter, and they may be the only ones that we’re able to analyze with confidence right now. Goalkeepers can impact matches by; reducing the xG of shots by being better positioned, prevent shots from happening in the first place by coming off their line and intercepting crosses and through balls, and add value in the build-up phase.

Best performing goalkeepers in specific concepts during 2024-25

Metrics available to measure those concepts are more repeatable. Therefore, and depending on a team’s style of play, it might be the best decision to focus on a keeper’s passing ability, or cross claims, as long as they haven’t historically been too bad at preventing goals.

Thank you for reading, if you’d like to contact me regarding any questions, comments or to consult my services, you can contact me on:

• Bluesky - https://bsky.app/profile/tomh36.bsky.social

• Linkedin - https://www.linkedin.com/in/thomas-harrison-a682a2175/

• Twitter/X - https://x.com/tomh_36

También, se puede leer este artículo en Español- https://cambiodejuego.beehiiv.com/