Everton at forefront of performance analytics usage in Premier League

As teams throughout the league become increasingly progressive in their use of match data, the level of detail available to teams to assist in their pre and postmatch analysis and preparation has become ever more complex. Stat providers such as Prozone and StatDNA supply data in various forms to an estimated 15 of the 20 Premier League clubs.

In Everton's case though, the team has been using data analysis ever since David Moyes first became manager at Goodison in 2002. "Moyes wanted something in place to provide him with extensive information because he's so detailed, thorough and methodical in his work," said Steve Brown, First Team Performance Analyst at Everton. The fundamental basis for Everton's pre and postmatch work is provided by the Prozone software which provides the eventing details, player tracking and physical data from which Brown and his team build their analysis.

The climate of change when it comes to use of data for match analysis is probably best reflected in the transformation in players' attitudes. Today, at Everton it's commonly used and accepted at every level down from the senior squad to the youth academy. "I think initially some players were sceptical toward it," said Brown. "I've had past experiences where people have viewed it as a negative tool and if it's used wrongly it can be damaging in the coaching process."

There's no question that some among the old guard of players within the professional ranks were initially apprehensive or resentful toward the data and questioned not only its application, but its usefulness. "That was where the main boundaries lay," said Brown. "Some of the older generation of players who weren't as familiar with working in a detailed analysis process -- of which we do a large amount of at Everton. At Everton, they've had to buy into the process and if they don't buy into the process, they'll just get frustrated because of the way the manager works here."

Building a weekly game plan

With Moyes leading the way, match analysis has become an integral part of Everton's weekly game preparation and one of the keys to its opposition scouting. "Now we tend to examine at least 5 games [of the team we're facing], break them down, link them up with the scouting reports we have and the stats we get from ProZone," said Brown. "[From that] we build a detailed picture of how we think the opposition play and what we think their style, strengths and weaknesses are, what their intricacies are in terms of what they're doing positionally, and if their individual players have trends and tendencies to in their play."

Brown meets with Moyes and his coaching staff for several hours to develop a game plan, which is then also distilled into a shorter version for the players, emphasizing the key points such as goal or set-piece analysis. Brown will also identify the characteristics of opposition players to discuss with Everton players. "We will identify certain trends of a player, for example a winger, and show our fullback what we think they're more likely to do in certain situations," said Brown. "It's the small detail which we are meticulous in. We examine if the players have certain patterns to their play which we think we can expose or need to be aware of. Do certain wingers have a style of attacking play in order to create a crossing position? Something that we might be able to predict and counter. The idea is to give them a picture of what the opposition [player] is doing."

Beyond that, much of the game plan focus is on tactical shape and pinpointing weaknesses in the opposition to be exploited -- for example, a left winger who doesn't get back into a tight midfield position. "At times we've used certain 'without possession' shapes which have negated the opposition," said Brown. "At times the manager has come up with a specific system, making subtle positional changes which has then allowed us to negate the problem or allowed us to capitalize -- it's those kind of intricacies which Moyes is brilliant at."

When assessing the opposition, Brown reveals that it's often possible to discern which teams are also using performance analytics heavily in their own approach. This can be apparent in teams that work on their tactical shape in a way specifically designed to stop Everton. However for the most part, Brown points out that most teams tend to stick to a noticeable pattern or trend. "You'll see that individual changes and personnel changes affect their weaknesses or strengths or the way they play. For example there are some fullbacks that like to stay in position irrespective of which team they are playing in, but there are some fullbacks that want to go in tight. It doesn't look like a definitive instruction to go tight, but a natural tendency and it's important to pick those things up really."

Working with Tim Howard and Landon Donovan

While some clubs perhaps place more emphasis on individual game plans, Brown also undertakes extensive one-on-one work with the Everton squad -- operating an "open-door policy" where players can come and request anything they require. Two of the keener participants have been U.S. internationals Tim Howard and Landon Donovan (during his recent loan spell with Everton). In Howard's case, as the goalkeeper, it's predominantly a question of profiling the opposition set-piece takers and giving him their various tendencies. "With Tim, it is more about providing him with the information to make him aware of patterns he may come up against," said Brown. "He likes to be as informed as possible going into matches." In contrast Donovan, as an outfield player, received information which was more specific to his own game and the individual weekly matchups he'd face.

"Landon was eager to find out information on the fullbacks and wingers he was likely to be playing against -- especially with him coming into a new environment and not necessarily knowing the players," said Brown. "He was keen to know his positional requirements to aid our defensive work; i.e. making sure he was in the right position to help our fullback and 'double up' on the opposition winger. He also wanted the information on the opposition fullbacks he was coming up against ... at times when Landon felt it was right, he'd come and say 'can you give me some information on their strengths and weaknesses? Where I can get at them?'"

Brown also noted that Donovan was particularly keen to know his physical data and technical stats after games. "Landon was looking to compare where he was in relation to the Premier League average for his position and where he was compared to opposition players. It was just for self-assurance, but I've no doubt that when Landon came in, our data picked up a level. It moved up a notch which we knew it would from previous anyway. We knew what we'd be getting, his work rate is phenomenal."

Placing data into context

However, as Brown is quick to highlight, the key in interpreting any physical data remains the context of that data. It's no use just looking at the raw number of a team covering a certain distance, it's important to examine why when comparing data from different games. For instance, a lower figure in distance covered by a team in one game could be due to the ball being out of play longer, or simply having contrasting possession or the territory share throughout the match.

Different teams also stress varying statistical values, most of which aren't publicly available. For example, some teams covet the high-intensity top-speed sprint stat -- the number of times a player completes a full-speed sprint during the course of a game. West Ham's Sam Allardyce, is among those managers that regard it as one of the most significant statistics in a match. However, Brown takes an alternative view. "We're not too sure about what the siginifcant physical variables are: is it high intensity distance, sprint number or distance or is it total distance?," said Brown. "Whenever we've looked at our data, we have fluctuating variables which have all shown more correlation with a winning performance. Some seasons it is total distance, other times it's been shown to be high intensity distances. I feel that we're still to find a consistent strong relationship in our physical data and winning performances.

"And again, we're always trying to relate it to the context of the match, what the performance looked like and whether we were satisfied with the performance or not. How the ball possession and actual time in possession or player selection and game situation, game plan or tactics impacts the physical variables. These aspects will bring us closer to arriving at a more relevant assessment of the physical information."

Ultimately though, Brown believes there's still far more core work that can be done in the field of performance analytics -- advanced data models and statistical processes that can be developed as more clubs consider the use of in-house statisticians (Everton itself is considering working with various universities). And of course there's the simple matter of determining what exactly is the right data to look at. "Looking at the basic entry-level data (number of headers won, tackle made, final third & penalty area entries, passes attempted and received) is becoming less relevant to us. We are striving to put the data in greater context: pitch location and efficiency of actions attempted.

"How effective are we in reaching our targets? How efficient are we on set plays, crossing scenarios and gaining controlled attacking third entries? We feel that this is the type of data that enables us to become more accurate in analyzing our performance. Trying to make sure that our assessment of performance is relevant to the manager's team philosophy is imperative. Once you're comfortable with that, you can move to the next stage."

Jen Chang is the soccer editor for SI.com. He can be reached on armchairsweeper@gmail.com. You can also follow him on Twitter at Jenchang88 or Facebook.