Could 'bad data' be affecting Aston Villa's performances?

Incorrect data analysis in 21st century football and the ethical considerations of inaccurate data recording are amongst the biggest problems in the sport. Fans who cannot analyse correctly can use data to back their chosen narrative at the cost of misleading others. Teams who cannot track data correctly risk making terrible decisions.

A document leaked to 7500 to Holte by an Australian source shows that Aston Villa may have a problem when it comes to tracking positional data via GPS systems provided by Catapult Sports due to the potential for stadium roofing and overhangs to disrupt the passage of GPS signals to recording devices.

The document is a tender put forth by Catapult to secure a deal with the Australian Rugby Union. In the pitch presentation, there is an admission that data can be affected under certain conditions.

This news relates to information brought out by David Ornstein of The Athletic who states that Premier League clubs are paying over-the-odds to a variety of data companies to chase ‘marginal gains’.

As for the Catapult tender, there are typical interference issues associated with GPS usage, but also non-typical issues.

In an interview with Training Ground Guru, the Brisbane Broncos Head of Performance, Paul Devlin, noted a number of GPS based concerns:

"Although NRL teams have been wearing GPS units (All Catapult) in games for more than six years, the overhang at some stadiums - namely the Allianz (Sydney), Suncorp (Brisbane) and ANZ (Sydney) - mean most GPS providers can’t get a signal. STATSports’ Apex unit is the only one I’ve found that gives a perfect signal from 17 satellites at those stadiums. The other teams can’t get accurate data in those stadiums with the other units."

GPS units in modern sports science calculate important sporting data. Data, which if used correctly, and if accurate, can transmute into analytic golddust. A GPS tracker in sport can calculate acceleration speeds, positioning info, decision-making speed and distance covered. For fitness purposes, a GPS can identify which players are at 100-percent, and identify who is lagging behind. There is a strong argument that the GPS tracker is amongst the most important items in the arsenal of a football club.

Correct GPS usage in the chase for marginal gains could be a big thing for smaller clubs looking to find some kind of competitive advantage. Incorrect usage of any kind of data, however, is fools gold amidst fancy graphs.

With Devlin's quote that many GPS systems track incorrectly due to stadium design in covered grounds, and Catapult's admission of data errors via signal interference in the tender document, how does this effect Aston Villa?

For one, we don't know which stadiums are affected by tracking issues. Every Premier League stadium that overhangs like the ones in Australia mentioned by Devlin could be affected by data issues, or none at all could be affected. We don't know, but we do know that some GPS trackers are more secure than others, while certain other trackers may be affected by situational context.

Here are Catapult's tracking data images from July 2018 and September 2018 sourced from the tender document, and from both images there seems to be something of an accuracy issue, even without much of an overhang present in the stadium.

The big question for Villa: Can Catapult deliver constant and reliable data from every single stadium? Is it 100 percent accurate? Can Villa translate any issues in the data if there are any? If there have been any data issues, has this affected performance and fitness analysis? If so, do they know? Why not use an accurate tracker?

Furthermore, are remote stadiums affected by a tracking accuracy issue? Is London especially susceptible to any kind of interference? The document raises plenty of questions, and we do not know for sure that Villa's data has been at all compromised by tracking errors. The possibility does stand, however.

What we do know is that Catapult's GPS system can be affected by a wide range of situations.

GPS data analysis and data quality aren't the only reason why a football team would lose games, but there would certainly be an interesting correlation between late-game fitness levels and post-match performance analysis. If tracking systems aren't at all reliable, are Villa able to accurately identify the reasons as to why they lose out in the second half of games?

Ornstein’s report notes that a number of companies are in dispute with Premier League clubs over data and he states in example that ‘one Premier League club recently wrote off an annual six-figure sum they were paying to a firm for performance data that was eventually found to be riddled with mistakes.’

If there is an issue with Aston Villa’s data, it certainly would not be limited to Catapult, and data issues certainly would not be limited to Aston Villa. Again, we cannot accurately state that Aston Villa’s GPS data has been compromised at all - but there is a sign that Catapult’s systems aren’t at all 100% accurate. This leads back to Paul Devlin of the Broncos who compared Catapult to a competitor to find errors in certain stadiums, which is revealed closer in the pitch to the Australian Rugby Union.

Hypothetically, if Villa are trusting in data to achieve marginal gains to avoid relegation, and are relegated, then there will be questions about the data and it's accuracy - a picture already brought up by a member of the Australian sports fraternity.

One of the funniest Aston Villa stories of the season relates to (another) kind act from Villa captain Jack Grealish. After a match, Grealish handed his Villa shirt over to a fan. A par for the course action for Jack - of course - but the humour doesn't result from Jack's shirt, but from what lay inside the shirt.

Jack had passed over his Villa shirt, and the internal vest with a GPS tracker nestled inside.

It's an important device, and one we cannot lose track of.

But is it losing track of Aston Villa?