Moneyball XI: Finding undervalued players with overlooked statistics – data analysis

If there is one club that is the envy of English football at present, then its Brighton. Not only do the Seagulls have a young, inventive manager and a team playing attractive football, but their player recruitment drive is raking in hefty transfer fees.

One of the summers transfer sagas was Moisés Caicedos move to Chelsea from Brighton, which was reportedly over £100 million. The South Coast side had initially bought the Ecuadorian for allegedly just £4.5 million. This is one of many examples of transfer deals that Tony Blooms club has been accustomed to of late, where young, unknown players have been bought cheaply, only to be sold on for a sizeable profit.

The means of identifying such unknown players have been achieved by taking advantage of vast amounts of data collected at all levels within world football. However, Brighton isnt the only club utilising data for transfer market activity. Premier League rivals Brentford have long been advocates for using statistics and it was such methods that played a part in the West London side reaching Englands top flight and staying there.

Imagine a scenario of a side looking to play possession-based football in a 4-3-3 formation. In this data analysis, we will use overlooked statistics to create a starting XI of undervalued players who would conform to the tactics of the 4-3-3; all players considered for analysis will be under the age of 26 and not be currently commanding a large transfer fee.

Goalkeeper

In the modern game, it is often required of goalkeepers to be comfortable in possession. Whilst this is a desirable attribute, it should be a secondary thought to a keepers ability to save efforts. However, when considering a keepers save percentage, this may not be the most proactive.

For example, a keeper in a team that is high up the league is likely to be protected by a strong defence, and hence, efforts at goal may have a low expected goal value. This, in turn, suggests the keeper is more likely to make the save, increasing his save percentage value.

As a starting point, we have looked at the post-shot expected goal values (PSxG), which are the expected goals based on how likely the keeper is to save the shot. We have then deducted the goals conceded (including penalties) from the PSxG value.

From here, we assessed the secondary criteria, such as possession, and decided Famalicão, Luiz Júnior, makes our team.

At the time of writing, the 22-year-old Brazilian keeper had a PSxG value of 22.5, but when deducting the goals conceded, the Famalicão keepers value is +5.5, which is bettered only by Farenses Ricardo Velho in Portugals top tier. Although Júnior does not rank top, his possessional play outperforms his Primeira Liga rival.

The radar chart shows that although similar save statistics between the two goalkeepers, Júnior is facing more shots per 90. While Velho does not concede as many goals, he tends to opt for a long ball upfield, which is not desirable for a side looking to play out from the back.

It is shown that Júnior is much more comfortable leaving the pena