Why patterns and context are so essential to analysis in football

If you’re reading this article, chances are you love to analyze football. So allow me to help with your love for analysis with this call to action. Stop taking information from single-match occurrences, or single-match statistics, even statistics over time on their own, as a mechanism for making inferences about the wider context at hand. Recognizing patterns over time, and the wider context behind those patterns, are the essentials behind analysis in football, allowing you to more accurately assess performance, and improve upon performance problems.

One of the most preeminent problems I see in football analysis lies within inconsistencies in continuity – that is, failing to adequately assess for continuity over time. Match analysts may take patterns occurring in one specific game, and make claims for that action occurring within the greater context of every single game. This isn’t the case, and it’s why ‘team and manager‘ analyses, are the most difficult to study. Why? They take time, and constant recognition of patterns occurring over time, not to mention the wider context behind each action, rather than single-match instances. What we are essentially looking for within any team and manager analysis, are patterns that occur over at least a five-ten game period. Within this, there is a recognition that those patterns may realistically evolve into new and improved compositions over the next five-ten matches and the five-ten afterward. Football strategy is an ever-evolving process within top clubs, and a team may respond in unique and novel ways to each of their opponents, changing tactics on a week by week basis (someone get Graham Potter on the phone!).

So with that, we recognize how difficult it may be to make inferences about a team over time, and identify patterns that occur over the course of a prolonged stretch of time, let alone an entire season. But that is what we want to study if we are to truly assess a team or player’s performance within our analysis. We want to assess for patterns that occur over time. This not only increases the validity of our analysis, but allows us to provide greater context into the team or player’s specific needs.

For example, a player may flounder toward poor finishing in a single match, missing chances for fun. You could bring that evidence to the player via video analysis and clips of their horrid misses to show them what they did wrong, in the quest of helping the improve for next time. Or you could take the opportunistic approach, that also happens to bring in greater context, and show them patterns within their finishing techniques that occurred over the previous twenty games where they scored goals for fun – and show them what they did right in those instances. You could also show clips of times where they missed chances down to the poor passing of their teammates, or pressure applied by the opposition, helping the player understand psychologically, that their misfires are not all down to their raw knack or lack thereof for scoring goals.

This is essentially what we are looking to provide within our Role Continuity Evaluation System. We want to assess player performance over time based on their specific role on a football pitch, rather than just a broad range of statistics that may or may not be relevant to their job. We can then identify a player’s role not just in a single match, but over time (the continuity piece), allowing us to more accurately assess their performance between an array of fixtures. For example, there’s a wonderful ‘Ball-Playing-Centre-Half’ in the Canadian Premier League named Thomas Meilleur-Giguère. In three of his four matches so far, he’s completed over 50% of his long passes according to FotMob, with 10.25 passes into the opposition’s half per 90. We could show the 24-year-old clips of his long passes over time, and the ones that worked against the ones that didn’t quite come off. We could look into the one match in which he reached 45% rather than above that 50% threshold and identify potential performance problems that he can combat to ensure he always reaches above his own standard. We could study his techniques used over time, his direction of those passes, the “packing” behind each pass (the number of players that he beat with his pass), and look for opportunities to expand on that ability within the Pacific set-up.

TMG’s passing map v. Halifax – 90% successful, but only 45% long passes (CANPL.ca)

For all Pacific’s possession, Meilleur-Giguère’s passing into the final third is relatively low (2.0 per game), despite being second on the charts for total passes in the league (70.5 per game) and first for passing percentage of defenders (92.7% – fourth of all players in the CPL). Perhaps he could be more progressive with his passes into central channels, using Alejandro Diaz’s exceptional movement and chest control to caress passes back to the ground closer to goal. Perhaps we could look at the movement patterns of Joshua Heard as he sprints in behind an opposition back-line, and look for opportunities where the left winger could be found instantly in build-up play, rather than Pacific’s patient approach that tends to take centre-stage, as they take advantage of their right-sided overloads. In essence, Pacific have a player with an incredible ability, and they could do more to harness that player’s power. By studying patterns over time, we can not only come to this conclusion, but more accurately assess how both the player and team can work together to improve. Single-match video footage or video clips may be completely irrelevant in that quest, as it doesn’t take into account the greater context of what may be happening over time.

TMG passing map v. Valour – 94% pass completion, with 67% long passes (CANPL.ca)

But we can dig even deeper within this concept. In my previous example, I only made reference to statistical information as the starting point for the analysis. A wider context is always at play, and new patterns can be elucidated upon further examination. For example, TMG’s long passes may be more likely to be received by a player with exceptional ball control (e.g. Marco Bustos), or a player that has time and space to receive (often diagonal switches of play to the other side). Perhaps they are less likely to be received by players running into space (creating more of a 50/50 challenge), such as passes down the line to his own left-back Nathan Mavila. The opposition is also important to the equation. For example, did Thomas Meilleur-Giguère make any of his long passes whilst under pressure from the opposition, or were they all uncontested plays? We can then study this contextual information to identify opportunities to improve upon TMG’s long-passing under-pressure, his ability to play passes down the line into advancing runners to the same strength as his ability to play switches of play to open players. We can even identify opportunities for the 24-year-old to keep the game ticking through short passing and simplicity, as opposed to those expansive moments that may unlock the opposition’s defense.

TMG’s passing map in the 0-0 draw against Edmonton, 97% completion (CANPL.ca).

Taking notes on the direction of his passes also becomes important. Quite outstandingly, TMG completed 97% of his 72 passes against Edmonton, before being replaced at half-time, much to our confusion. The two misplaced passes followed a similar trajectory. They were both longer than average, and were spread from right to left diagonally. Rather than being received, they both landed out of bounds. That means the centre-back didn’t have a single pass in the match that found its way into the feet of the opposition. When he slipped up, it was through passing too close to the width of the field…too close to the touchline, which naturally has more room for error. But without the patterns of his other passing maps or the wider context at play, we still cannot make inferences. Perhaps those two misplaced passes resulted from poor touches of his teammates. Maybe on both occasions, his teammates slipped in mud! More likely, for all you know, he was under pressure in those situations, and had to more quickly make up his mind, resulting in error.

Meilleur-Giguère’s passing map on the opening weekend vs. Forge (CANPL.ca)

Digging even deeper and elucidating the vast array of questions that can now arise, those passes could have also been exceptional passes, and for whatever reason, they didn’t pay dividends. On the other end of the spectrum, it could be very possible that a few of his accurate passes resulted from complete errors from the opposition. He may have pulled off a wonder pass that resulted in Pacific’s best chance in the match, but that doesn’t mean that exact replica of pass and movement patterns to receive would work every single time. That is again where patterning becomes imperative. By studying the wider context of a series of matches over time, we can then assess moments where TMG can pull off those wonder passes to success, and moments where he would be better off playing it safe. We can assess the direction of his passes, and look at the maps over time to determine that a significant number of his incomplete passes head toward the touchline. On this matter, we can even adequately assess at this present time that his centre-back partner in Amer Didic shouldn’t continue to attempt the same number of hoofed passes up the field as TMG. He lacks the same exceptional technical quality, not to mention timing of decision making with his long spreads. While TMG’s feel intentional, Didic’s feel more hopeful. On one specific match, the former Edmonton man may complete more of his long passes than his partner (P.S. it hasn’t happened yet). But even if the wider context reveals that Didic found himself under greater pressure, and that his teammates were incredibly poor at receiving these passes that still came off, we cannot make inferences of his quality based on that single-match information.


In a research article by German researchers Andreas Grunz and Daniel Memmert, the authors discuss how pattern recognition can arise from simple statistical clustering, or more in-depth methods of “neural networking”. In data science, neural networking is defined as a series of algorithms that recognize relationships in a set of data, through mimicking the way the human brain operates. What does this mean? It means it takes into account the greater context behind clustering of data than more simplified forms of analysis, such as “similarity analysis”. In order to gain inferences over patterns that occur over time, the type of action, players involved, and success of those actions can be studied, in addition to metrics like pressure applied or speed/direction of the action. A machine could do this within a matter of seconds, but the human brain can also use the same methods (hence the point of calling it a “neural network”). The machine can adequately study in a way that mimics our brains, and we can simultaneously mimic the machine’s processes in our quest to identify patterns.

So next time you assess a player’s decision making processes, or the success and/or failure of their actions on a football pitch, be sure to include the wider context through some of the metrics listed above, if not more.


If you’ve skimmed to this section in the article, you may have missed the over-arching point. So allow me to say it again. We cannot take statistics out of a hat, or single-match occurrences, and use them in our analyses without providing (or at least understanding) the greater context. To give a final example, I’ve identified Halifax’s Samuel Salter to be quite a poor passer of the ball in recent Tactical Reviews. But perhaps he’s attempting a far greater deal of difficult, forward-thinking, progressive, long passes, particularly ones closer to goal where the alarm bells are already ringing for opposition defenders. The greater context needs to be considered, and the context needs to be considered over time. Salter had a better day in this weekend’s fixture against Edmonton, completing 80% of his 20 passes. The movement of Zachary Fernandez underlapping and overlapping around him was spectacular. In other matches, the movement of those closest to the situation may have been substandard.

Above are two separate passing maps for the Halifax man this season. The first shows a game against Atletico Ottawa, where he completed 65% of his 20 passes. The second shows this past weekend, where he completed 80% of his 20 passes. Against Atletico, we noted at the time how Halifax struggled to adequately find connectivity between the thirds and progress through the centre of the pitch, against Gonzalez’s stern 4-4-2 defensive structure – not to mention a very strong defender in MacDonald Niba as Salter’s direct opponent. Against Edmonton, with not only a weaker left-hand defensive set-up combatting his movements, but improvements on the movement and positioning of his teammates, Salter’s passing stats rose to the occasion. He’s still not a passing guru, but this greater context, taking into account more than just the stats alone, provides a greater recognition of the patterns that could help Salter to succeed in the future, or continue to hinder his passing stats.

With this example, you can see how a vast array of opposition traits and characteristics, combined with a vast array of traits and characteristics associated with the actions surrounding Salter’s own team can bring greater context into a sweeping statement within an analytical framework.


In the analyses you find on this website, there is an intentional approach to never take anything in isolation, even within Match Analyses. The wider context of the situation, and patterns that occur over time, must continue to be considered, in order to improve the validity and quality of the analyses. This, of course, is not the easiest task in the world. There are a multitude of aspects to consider within pattern recognition and the wider context at play. But you must scuba-dive into the unknown and dig deeper when making sweeping statements, particularly when doing so on the basis of one single inference or statistic. If you can add this to your analytical framework, you too will enhance the validity and quality of your analysis.

So there it is! The importance of not only understanding the wider context, but exploring patterns over time when analyzing football matches. If you’re looking for Opposition Analyses, Player Analyses, Scouting & Recruitment, or Performance Analyses at the youth or professional level, first, you’ve come to the right place. But secondly, I’m open and available for freelance work, whether it be video or live in person, working with players, managers, or soon to be managers as they develop their ideas about the game or solve performance problems. Simply contact rhys@themastermindsite.com for more information.

Thanks for reading and see you soon!

This article is part of our Introduction to Football Analysis – Online Course with Rhys Desmond. See what else is in store for the course and navigate to the next section below.

See the entire course curriculum and jump around by visiting:
Introduction to Football Analysis – Online Course with Rhys Desmond

Success! You're on the list.


The advantages of having a player who doesn’t defend

We often look at a player who doesn’t defend as a detractor. Players are told from a young age that they must defend, regardless of their position on the field. In many respects, this is true. But when you get to the professional level, roles can evolve in novel ways, and even incorporate a player…

How to organize the defensive line as a goalkeeper

With a wonderful view of the action, goalkeepers must be more than just shot stoppers, and participate actively and vocally in communicating with their mates. They should not be solely responsible for organizing the entire team, but have the ability to do so through the vantage point they have on the field, and often the…

My top ten articles of 2022

2022 was another record-breaking year for TMS, and a great year of firsts for my work and the site. In 2022, we launched our Role Continuity Evaluation System, a batch of brand new Game Model Ebooks, and our Introduction to Football Analysis Course. This year was also the year that I had multiple clients from…

The half-spaces as a key chance creation channel

It has long been hypothesized that ‘Zone 14’ is the holy grail of chance creation. The ideas around this concept were built around a study from the late 1990s that specified that successful teams had a higher frequency of getting into this zone when compared against their peers. Since opposition clubs often compact central channels…

The multiple layers of Trent Alexander-Arnold

Trent Alexander-Arnold is one of the best right-backs in the world of football, and possibly one of the greatest the Premier League has ever witnessed. So for all the extremities that he accomplishes in attack, it becomes very normal for fan-bases across the globe to try and poke holes in his game and conjure up…

11 thoughts on “Why patterns and context are so essential to analysis in football

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s