Dom’s Comms: Week 10.5 — Fullbacks
The new tactical trend, just like last season, is fullbacks. I’ve spoken on the shift in my last article, but I wanted to provide more insight today. After watching, and coding (certain aspects) of the 0–1 loss to Augsburg, I have created a brief data set attempting to quantify and demonstrate the difference between both players' roles.
The primary reason for this article is that the most intriguing element is the “side fluidity”. Across the three public preseason games, Steve Cooper has opted for the same asymmetrical setup with his fullbacks, initiating games with a generic 4–2–3–1 or 4–3–3 formation on paper.
One of the fullbacks inverts into a wide-CB role, creating a de facto back-three in possession, the other is afforded the freedom to advance into the last attacking line and function as a makeshift winger, but unlike last campaign’s clarity — where LB became LCB, and RB became the right #6, the side to push or sit has fluctuated with each fixture.
In the most recent game, the 0–1 loss to Augsburg, the starting fullbacks were Victor Kristiansen (LB) and James Justin (RB). As you might expect, Kristiansen played the “attacking” role, while JJ inverted as the RCB.
If you watch the game through this microscopic lens, it’s clear to see the movement happening. But, perhaps other Leicester City fans paid less attention to an asymmetrical FB setup, and more to the general play of the friendly.
That’s where my niche data collection task and specific fullback analysis will attempt to highlight what was happening, and what has been happening so far this preseason.
First, we have the .csv data re-formatted into a table for your viewing pleasure, with colour highlights indicating which player outperformed the other on each metric.
Inside the “Info” section, you can see that both players have “73” for “minutes played”, despite Justin playing the entire game. This is because once Kristiansen was substituted the system changed.
The idea for the asymmetrical FBs is to have one join in attacking sequences and be a creative player (LB), and one to help in the build-up (RB). These roles are illustrated clearly in the “Possession” segment of the table, but I’ll show how the on-paper 4–2–3–1 formation transforms for this to happen.
Justin’s deeper role allowed more touches (202) of the ball than Kristiansen’s (137), with both numbers seeming abnormally high for two individuals if you tend to track game-to-game data, and it is.
Generally, data providers use the metric “touches” as an all-encompassing term for “actions” on a pitch, each pass, dribble, tackle, etc. rewards them with a single “touch”. For me, I coded every individual touch they both took, so Justin touched the ball 202 times.
JJ also completed more passes (78) and with a higher pass completion (94.87%) due to the shorter, less threatening pass selections he opted for. Often receiving the ball, dragging Augsburg’s press away from the other CBs, and then passing back into them for a route of progression.
Whereas, Kristiansen’s lesser touches resulted in more “crosses” (8), an equal amount of “key passes” (1), and an uptick in “touch p/pass” (2.98), as the space opened up to carry possession through the thirds before releasing it.
I also tracked passes into the penalty area in the first half, with a heatmap associated with a player’s general positioning…
While Kristiansen’s passes into the penalty area often resulted in a turnover shortly after being completed (75%), they added a good, slightly different threat to the attacking unit. None of the visualised passes included his “crosses” into the box, the passes coded were targeted balls into the area.
To improve shot creation from these moments, the Foxes needed to commit more players into the spaces around the box recipient (either Patson Daka or Stephy Mavididi) to interplay and move Augsburg’s fairly rigid last line to open space for a shot.
Yet, for his lower volume, JJ was the player who managed an “Open-play key pass” (1), i.e. a pass leading to a shot, with his direct line-breaker for Daka’s speculative volley in the opening 10 minutes of the game.
It seems a secondary role, to the primary recycle or progress via the double pivot, for the CBs is the ability to play long/direct, especially those utilised in the CCB — Jannik Vestergaard or Conor Coady — role.
If you look at the “Defensive Actions” section of Table 1, you can see that an advanced position (IP) for Kristiansen aided his contributions off the ball, particularly when the Foxes were in transition from attack to defence.
While the team shifted between their shapes, which I will showcase the out-of-possession shape below, the Dane could preemptively step up and initiate, while JJ’s role exclusively placed him in the last line and that limited his numbers.
When Cooper’s side had fully retreated into their OOP shape, if Augsburg held possession for some time, then both FBs would be in the last line of defence, as shown above.
The other factor was Augsburg’s reliance on their attacking right side when on the ball, which wasn’t an indictment of Kristiansen nor a target of his side, just a natural progression line that saw more sequences function on his wing, and therefore allowed more defensive actions.
In the end, with the combination of jumping into the advanced line IP, and that positively impacting him off the ball, Kristiansen’s game was quite heavily about the defensive metrics, 17.86% or 1 in 5 of his total actions were associated with defending.
On the flip side, while JJ’s adjusted role to a wide CB seems to be defensively-minded, and I think a large reason for his inclusion is due to his ability to defend spaces, only 5.95% or 1 in 20 of his total actions impacted the defensive side of the game.
Instead, with increased reliability in possession, Justin completed just shy of 95% of his passes, and 94.05% of his total actions were on the ball. To play the sitting FB role, you need to patiently find the right passes, which lends itself (data-wise) to more touches and higher pass completion (%).
I hope that sheds some light on how the roles impact the individual data.
With secondary data providers (OPTA) producing work for the Premier League season, I’d be able to identify the frequency and location of all these passes as post-analysis. But, instead, these were the metrics I quickly assorted whilst live-coding the friendly on Saturday afternoon.
This data comes with its shortcomings, but, in a pinch, and with a lack of in-game data across preseason, I wanted to provide at least a brief analysis on the key tactical implementation Cooper seems to be set on ahead of next season.
Unfortunately, because I spent a large % of the game coding data, I didn’t get to truly analyse the overall performance. That will have to wait for the RC Lens game next Saturday, as the final match before the Premier League 24/25 season gets underway.
Who knows, between now and then, Leicester City could’ve signed or sold multiple first-team players and that might have a consequence on the tactics, structure, and playing style.
At this stage, everything is still very much up for debate.
