Going beyond data scouting

One of the key questions about data scouting is how a player’s individual output is affected by the team they play in.

It is unfair to expect a potentially highly skilled passing central midfielder to put up the same type of output playing for a direct style team in League 2 than they would do playing for Barcelona.

But can we pick out key statistics that show the types of players who will be able to play at a higher level?

Basic data

A quick glance at Semi Ajayi’s passing profile from his season at Rotherham would not give the impression of a ball-playing defender. He barely made any passes and when he did they were inaccurate.  So how would be transition to playing for West Bromwich Albion?

Ajayi has gone from one of the least accurate, and lowest volume, passers in the league up to well above average for most metrics. It wasn’t that he was unable to play in a passing style but that the tactics of the team he played in was affecting his statistical output.

But if you look more closely into the data from the previous season you can pick out signs that perhaps Ajayi was not as bad a passer as the raw statistics suggest. On our creative passing metrics there are some interesting findings.

Ajeyi was around the 75th percentile for various different types of creative action, such as completing passes into the penalty area and the final third. Now this could be a function of volume, if he hit his passes long towards a tall striker you would expect that he would be making more passes to the penalty area than other centre backs on teams who play through midfield. But he is sufficiently high in enough metrics to warrant a closer look.

Advanced Data

With our advanced data we can look more closely at his passes and it becomes a bit clearer. Ajayi played a significant number of minutes in different positions. With our pass mapping tool we can separate out his passes played from different roles.

His good creative numbers are likely to be a function of his minutes played in a more advanced role where he played a lot of passes in the opposition half. His poor overall passing numbers also then are at least partially explained. A centre back has far more space to operate in than a midfielder. Their forward passes are generally into less crowded areas and completed at a higher rate. 

So with Semi Ajayi we needed to go into advanced data to see why his passing performance was so poor. It really wasn’t, our basic aggregate data wasn’t telling us the whole picture. As we can see below he is now settled in at right centre back and playing more typical passes for his position.

Making allowances

Another interesting case study is that of Joelinton. He went from a dominant attacking team in Hoffenheim to team struggling to create chances. His output will have changed, won’t it?

At Hoffenheim, he was a creative attacker, great at participating in build-up play but with slightly below goal output for a striker.

At Newcastle, we can see the immediate impact of moving from a dominant to a less dominant side. His shooting and attacking play output has dropped considerably from the 40th to the 20th percentile. 

What he does still have are decent build-up and creativity numbers, almost at league’s best levels of passing accuracy and expected assists.

If we imagine the transfer in reverse, with Joelinton leaving Newcastle for Hoffenheim then would we be surprised that the player who barely took any shots, and offered no goal threat suddenly became more threatening and creative?

We shouldn’t be. Strikers need service, if you only judge them by current goal output, without adjusting for the relative dominance of the team they play on you will miss out on all sorts of good players.

That is why the MRKT Insights Workflow for scouting includes both data searches AND scouting searches. We look at all the numbers for all players, and with advanced data where possible, if a player isn’t getting chances, but is putting up good creative numbers they may also need consideration. Context is always key.

We also view hundreds of hours of video, although great believers in data scouting there are good players who don’t show up in pure data searches. Sometimes a search can simply be on age, if a young player is breaking into a first-team they are probably worth a look regardless of their data profile. We have developed a huge database of promising players at all levels of affordability.

Our scouting workflow is designed to ensure we always include the key concepts of considering the team style of play and the individual player attributes required to play that role in that system. We then work with a combination of basic searches, advanced data, and our unique scouted player database.

We only work with non-competing clubs to ensure each client gets suitable recommendations. If you would like a chat, or demonstration project, please contact us.