• last month
Scientific data is changing the game for football clubs and their players. But how does the action on the pitch get turned into the data we see off it?
Transcript
00:00As the cost of securing football players continues to rise, teams are increasingly turning to
00:06data science to make more informed decisions.
00:09Smaller clubs within Europe use it to try and gain a competitive edge over their big
00:14budget rivals.
00:15So how has the rise of data in football transformed the sport in such a short space of time?
00:26Hello and welcome to Football Now from the La Salle Stadium in Qatar.
00:31When it comes to the way that football clubs analyse matches and scout players, there has
00:35been no greater revolution in the modern game than the collection and application of data.
00:40Nowadays, scouting a player with just your eyes is actually seen as a little bit old
00:44school.
00:45Most teams now employ data scientists to interpret thousands of games logged by data specialists.
00:51So just how does the action we see on the pitch get turned into data off it?
00:56So all those data points are being collected manually.
00:59So per game, we have two operators who are being supervised by a team lead as well.
01:04So there's one for the home team, one for the away team, who collect all that information
01:07manually.
01:08So every shot the player takes, it's been manually input by an operator.
01:12We collect passes.
01:13The computer's going to calculate if this was in fact a long pass or a short pass.
01:17That means that every specific moment in the game, when you have 22 players, you only collect
01:21information about one player.
01:23And if you then think about the fact that when you have a 90-minute match, each player
01:27would have around two minutes.
01:29The ball is in his possession.
01:30So that means you cover actually only a fraction of what this player is doing.
01:33So that's what we call on-the-ball metrics.
01:35The focus is shifting more towards the off-the-ball metrics, that the camera will track players
01:39automatically.
01:40And we have operators who identify the players correctly.
01:42So we know that we monitor the correct players, assign the right data points to the right
01:45players.
01:46And that's a pretty automatic process.
01:48That's what teams want to know.
01:49So what is my team doing?
01:50What is this individual player doing when he has the ball?
01:53Of course, it's also changed the game for those watching, with phrases such as expected
01:58goals or XG now very much part of the vocabulary of an average football fan.
02:03But what is it?
02:04And how is it calculated?
02:06What the expected goal value is telling you is that, okay, given a specific opportunity,
02:10so taking into consideration lots of different factors, such as the distance to the goal,
02:15the type of assists, if it's a header or a shot, if there's players in the way that could
02:18block the shot, the position of the goalkeeper, all these factors make it a very big chance
02:23or rather a small chance.
02:25So from zero to one, or you could also put it in a percentage, like from zero to 100.
02:29So we've been collecting this information, the data around shots for decades already.
02:33So we have a huge database of shots that have been made in the past.
02:37And we use historical data to predict performance that's happening right now.
02:41Let's now take a look at a game that on paper looks equal.
02:45West Ham and Aston Villa drew 1-1 at the London Stadium last season.
02:49Both teams took 13 shots, each with five shots on target.
02:54But the data collected by expected goals tells the full story of the match.
02:59Aston Villa had an XG of 0.66, while West Ham had an XG of 2.02, meaning that they failed
03:06to capitalise on better quality of chances to win the game.
03:10Looking at the data this closely gives us extra context to understand why West Ham will
03:14feel, given the chances they created, why they should have won the game and taken all
03:19three points.
03:20Due to the difference in budget, some smaller teams around Europe are turning to data and
03:26analytics to try and gain an advantage.
03:28FC Midtjylland in Denmark, along with Brentford and Brighton of the Premier League, are just
03:33some of the clubs who've invested heavily in data science to try and gain that competitive
03:37edge, to varying degrees of success.
03:40So, how is data levelling the playing field between the haves and the have-nots?
04:10Not only do we see teams use data to help win matches, but players also rely on information
04:40specific to them to improve their performance and also prolong their careers.
04:45Former USA international Jay DeMeritt spent his twilight years with MLS team Vancouver
04:50Whitecaps and says that data on his physical recovery was imperative to making sure he
04:56was in peak condition on match day.
04:58I remember as I got older, you know, 33, 34, those were kind of the last years of my career.
05:04They used the data to make sure that I'd be fit on the game day.
05:08to go, hey, I think maybe on Thursday, we're going to keep Jay really light because his
05:12recovery based on the data is showing that that hard day, Tuesday, Wednesday, is showing
05:18that Thursday should be a light day for him because he's an older player.
05:21That's where I think the data becomes even more important from a layered approach.
05:25It's not just like, how many duels did you win?
05:27What's your top speed?
05:28What's the distance covered?
05:29It's like your recovery is longer right now on the Thursday, so let's keep him light.
05:34Let's not put him in that hard training session.
05:36As data continues to reshape the way football is played and analysed, will the beautiful
05:41game become something more of an exact science?
05:44Let us know if you think the reliance on science is here to stay using the hashtag FootballNowDataRevolution.
05:51That's all we've got time for from the La Salle Stadium here in Qatar.
05:54We'll see you next time.
05:55Bye for now.

Recommended