Skip to content

The Science of Scouting: Dangers of misinterpretation

As our series looking at the science of scouting reaches its conclusion, Leicester City’s head of technical scouting Rob Mackenzie reveals the dangers of misinterpreting data.

Graphic

There is a line in Soccernomics from Mike Forde, Chelsea’s former director of football operations, when recalling his time at Bolton that seems to cut the heart of the matter regarding the data revolution in the game.

“I sat in many meetings at Bolton,” said Forde. “And I look back now and think ‘Wow, we hammered the team over something that now we think is not relevant.’”

It’s a point that highlights the fact that statistics are in a state of flux as our understanding grows. Certain metrics are likely to go in and out of fashion. It’s something Rob Mackenzie is aware of in his role at Leicester. For while logging assists has entered the mainstream in recent times, Mackenzie is cautious when it comes to overstating the real meaning of a ‘chances created’ tally.

“I think quality of chances created is very important,” he tells Sky Sports. “Shots on target could be 15 versus 12 but how threatening were those 15 chances? That’s where goal expectancy can bring an added level of depth to data or information. Goal expectancy from chances created is important.

“I mean, a key pass could be a two-yard pass to someone on the halfway line who sees the keeper off his line and shoots. That’s nowhere near as positive as a square cut-back to someone in the six-yard box who scores. We need to interrogate the data a bit more and reduce the risk of making a mistake.”

Image: Xavi: Statistician's dream

Passes are another area of the game that have come under scrutiny. Everything from Leon Britton’s pass completion rate to the number of passes that Toni Kroos or Xavi Hernandez can get through in a game have become common discussion points in the modern game. Here too, technical scouts must drill down into the data to identify what really matters.

“Passes are a classic one,” adds Mackenzie. “You’ll go on your Twitter feed and people will be telling you that this guy has completed 50 passes. That’s all well and good but for a holding midfielder who gets the ball off the back four that’s what I’d expect. The real question is what is he doing with the ball? How penetrative is he being?

More from Science Of Scouting

“If those passes are side to side that’s great but is he having an impact on the game? When we code a game ourselves we look at whether passes break the midfield line and where passes break the defensive line.

“That way you’re able to see how many opponents a player can take out of a game with his passing. If he’s playing it from side to side then a team can sit back and defend against that all day long. If he’s picking out gaps and taking out whole rows of players that often indicates effective play."

Monday Night Football, MNF
Image: Monday Night Football

Mackenzie urges caution to Sky Sports viewers too. “Another classic for me is on Monday Night Football when they get the average team position up and people everywhere are soon discussing how reserved a player has been because this is where his average position is for that half or the full game.

“But that position is actually generated by the touches he has when he’s on the ball. So it may not be that he’s been sitting back, he may have been bombing forwards all night but nobody has been picking him out.

“It’s about the guys using it knowing what they’re presenting but also the people at home understanding what it is they’re being shown. I get the impression that some people believe that data is generated by the Prozone cameras that track movement, when in reality it is calculated by an accumulation of a player’s on-the-ball actions.”

When it comes to interpreting data, it seems the key is always to know what the data isn’t telling you before you extrapolate meaning to decide what it is.

Click here to see the full list of articles in our Science of Scouting series

Around Sky