Soccer analytics: The newest trends

Unlike the three-point revolution in the NBA and the advent of on-base percentage in baseball, soccer doesn't have that single metric that has changed the way we follow the sport.

Of course, there are trends and tweaks but analytics isn't rewriting the game of soccer ... yet. However, with every advancement in the space, the pros are gaining a deeper understanding of facets of the game in which they never previously had any insight. That is leading to more innovation and setting new trends in motion. 

With so much more to learn, the newest frontier involves player tracking data. 

A new world of data 

“Tracking data is definitely going to nurture the next phase of analytics in the sport," Nashville SC director of strategy and analytics Oliver Miller-Farrell said. "The data itself is very rich. It’ll be about how clubs harness that detail into actual insights.”

Ahead of the 2020 season, MLS signed a deal with Second Spectrum to supply cutting-edge player tracking data to every club. Second Spectrum installed optical tracking systems in every MLS stadium during the offseason, enabling it to leverage those cameras and a cloud-based analytics engine to capture data on the location of every player, the referees, and the ball 25 times per second. 

The data tracks everything and everyone, including the positioning and movement of players off the ball. As the database and sample size grows with each passing game, tactics can be understood on new analytical levels, including movements of players off the ball, team shape in specific states of play and much more.

“With the Second Spectrum deal with tracking data, it automatically generates so many different parts of the game," FC Cincinnati director of soccer analytics Alexander Schram said. "Now we can not only assess individual player performance, we can look at team strategy more because we know exactly where the players who don’t have the ball are. I think that’s going to be a massive step we’re going to take, looking at tactical analysis purely on data."

“You can watch a whole game in 2D with all the players as dots," Philadelphia Union data analyst Dean Costalas added. "It’s like ‘wow, let’s watch how this whole thing unfolded.’ The 2D dots show stuff."

Below is an example of the 2D dots, showing a play during Columbus Crew SC's 1-0 win over NYCFC on March 1, 2020. Crew goalkeeper Eloy Room sends a long ball into space under pressure, springing Lucas Zelarayan in behind and resulting in a Maxime Chanot red card.

A deeper understanding

Total distance covered, sprints, top speed — all this data is now available to every MLS club. For a team like the Union, which adopted a high-pressing system predicated on high energy over the last two years under sporting director Ernst Tanner, those numbers are invaluable. 

“Now we have fitness data," Costalas said. "How much did this player run? How fast did this player run? How many sprints? Just take our first two games of the season. In the first game [2-0 loss to FC Dallas], we ran about 106 kilometers as a team. Then against LAFC [3-3 draw] we ran 120 kilometers. If you watch those games, there was obviously much more transition. Wide-open, back and forth. The Dallas game wasmore traditional and tactical.”

Information about the more fluid aspects of the game, including possession, is now readily available to clubs. What specific actions led to a higher probability of a goal-scoring chance? What areas of the field are being undervalued, what areas are being overvalued? 

“There will definitely be areas that will accelerate, being able to measure possessions and possession values," Miller-Farrell said. "Understanding how discrete actions within spells of possession affect the nature of those spells. That’ll come with the tracking data.”

In those fluid states, analysts can now mine for insights into critical aspects that were previously difficult to assess. It starts with decision making. 

“That's one of the big things for me, I think everyone wants to evaluate decision making," New England Revolution analyst Tim Crawford said. "It’s hard.”

Before tracking data, you could measure success percentages in certain areas of the field, but only if a specific action took place (e.g. an attempted pass, dribble, tackle, etc.). Given the number of permutations at any given moment in a match, that didn't come close to revealing the full picture. What happens if a player held on to the ball when they should have passed or hesitated?

With tracking data, now everything can be measured, whether or not a specific action occurred. Today analysts are able to better understand the probability that a single decision lead to a goal-scoring chance and whether the odds went up or down depending on a specific player's decision.

Event data previously spit out raw numbers, like a certain player making 10 line-breaking passes in a game. But without context, that number isn't entirely useful. The total of 10 line-breaking passes may sound like a lot, but what if that player had the opportunity to make that pass on 30 different occasions?

“Tracking data will allow MLS teams to value space creation and better understand decision-making," Toronto FC director of analytics Devin Pleuler said. “I call it the denominator problem. Now it allows us to much better contextualize things that happen on the field.”

“It’s the dream for me and this is where it’s headed — to start saying how often players are making good decisions," Crawford added. "Then it’s helpful to teach them better decisions, and see if they agree or not."

A new look at possession and passing

Pep Guardiola's FC Barcelona and Vicente del Bosque's Spain national team experienced heralded success with a tiki-taka style, predicated on short passing and maintaining control of the ball. It wasn't long before this possession-based style became the hot trend in world soccer. Lately, though, we're learning more about possession and its influence on match outcomes. 

Analysts have now found that possession on its own doesn't mean much. It's not an independent statistic that correlates to the final score. Jurgen Klopp has found great success with both Borussia Dortmund and Liverpool playing a high-pressing brand of soccer. Jesse March won a few Supporters' Shields while high pressing with the New York Red Bulls and trophies have been regularly hoisted by teams that were not possession-oriented. Klopp has famously said, “No playmaker in the world can be as good as a good counter-pressing situation."

“I like to look at possession because I think it’s interesting, but I definitely don’t think it means much," Crawford said. "It’s all contextual. I get excited when we have 35 percent possession and win because it’s hilarious. A lot of people have valued possession for a long time, but it’s not at all something that will win a game. It can have a purpose, for sure, but people are realizing possession is contextual. It doesn’t have a meaning on its own.”

The tracking data is what provides clubs that contextual framework around each action.

“We look at value vs. difficulty," Pleuler said. "In general, the more difficult a pass is, the more valuable it is. But it’s not a perfect relationship. Some passes are disproportionately valuable given their difficulty.”

For instance, a 40-yard cross-field switch is a difficult pass, but it may not have a ton of value. A lot typically needs to happen for that pass to lead to a goalscoring opportunity. Conversely, squaring the ball from the endline at the six-yard box back to the penalty spot isn't a difficult pass in a vacuum, but it's hugely valuable.

“What you find, the better teams are the teams that find those passes more — the ones that are disproportionately valuable given their difficulty," Pleuler said. "That’s kind of how we view it, the risk/reward tradeoff.”

Another example of how passes are not all created equal: Some coaches around the world may set passing completion targets for players from game to game. (e.g. a defender must complete 90 percent of passes, a midfielder 80 percent and forwards 70 percent) according to info like the one in the table below.

2020 MLS Expected Pass Completion % by Position

Data from Second Spectrum. Minimum 50 passes attempted per player to qualify to be counted.

“You can have a very high passing percentage, but does that make you a better player?" said Colorado Rapids first-team analyst Jase Kim. "There was a team I know who used targets [percentages for completed passes] ... Putting that target gives players a bit of a fear factor where they can’t lose it. But you can be making all these passes that don’t mean anything [just to hit the target numbers].”

Taking a look at some of the best creative performers in recent MLS history shows the complexities of evaluating pass completion percentages.

In MLS last season, Maxi Moralez led the league in chances created from open play (76) and yet he had a 79 percent pass completion rate. Carlos Vela had an even lower completion percentage (75 percent), while still breaking the record for combined goals and assists. And Sacha Kljestan also completed 75 percent of his passes in the New York Red Bulls' uber-vertical system in 2017 when he led the league in assists with 20.  

“If you have those sort of targets, that might hinder players," Kim continued. "Some players might have swagger and say ‘I don’t care’ like Paul Pogba. He has no fear and thinks he’s going to do stuff that’s going to be effective. That’s where the numbers might not work. That’s just an example of using data incorrectly." 

“In football, there’s so much permutation it’s impossible to use data [exclusively]," Kim added. "It helps, but you need experts still to pick out the moments.” 

« Benefits
Soccer Analytics Guide
Who's your club's expert? »