Thiago’s underlying numbers show why he will be a Premier League star

Few teams have invested so much into the concept of data as Liverpool, both in terms of time and resources.

From employing data scientists, to chasing inefficiencies in the transfer market to mid-match analysis sessions that value metrics, heat maps and 3D designs as much as heart and emotion, Liverpool, from sporting director Michael Edwards to manager Jürgen Klopp Klopp and his staff have embraced it all.

Arriving from the Bundesliga, Klopp had a heads up on some of the competition. No football league in the world is as technologically advanced or data-savvy as the Bundesliga.

Germany has been at the forefront of scouting, coaching and data innovations for the better part of a decade. All Bundesliga grounds are now kitted out with heat-sensitive cameras that track every player’s movement. Everything is logged. Every pass, move, look, flinch. A 3D mapping of how the players fit into one system on the pitch is tracked, too. No longer do teams rely on the guesswork of scouts to decode the shape of the opposition or to track a specific player’s movements. If the left-sided fullback is pushing up to create an asymmetrical shape, it’s there in a 3D rendering, carefully tracked by any number of matchday performance analysts.

The data is kept centrally by the league and is available to all Bundesliga teams. Unlike in the Premier League, where data analysis and matchday secrets are guarded like the Kremlin, the Bundesliga is an information zone. Everything is available to everyone at all times.

The league should be smarter for it, the coaches better, and the team’s given an edge — at least that’s how the theory goes. The better the league as a collective, ultimately, the better the *shudders* product should be.

One data set has taken on an outsized meaning in the German game over the past half-decade: Packing.

You might remember Packing from back in the heady days of the summer of 2016. Back then, packing was all the rage. It was the latest tool in the nerd/hipsters I’m-smarter-than-you arsenal. Only this one was easy to understand. It was intuitive. It seemed to help quantify some of the things that were forgotten or overlooked in other gaudy metrics.

Could it help quantify the rhythmic impact of the most beloved of modern footballers: The deep-lying creator. Could a metric finally help show, with numbers and figures, what made Andrea Pirlo and Xabi Alonso so effective? Could it explain why the rhythmic precision of Thiago Alcantara and Luka Modric is so essential to effective build-up play?

Impect, the company that created ‘Packing’ likes to think of football analysis as a pyramid.

At the top, you have the surface-level numbers: points, goal, etc. The underlying metrics try to pull those things apart, to better contextualise them, and to give more of a granular understanding of what happened in the game, why, and what it all means. That’s where you bring in more rudimentary data like shots and tackles as well as the nerdier stuff, those that add up to ‘chance creation’ and other metrics.

One of the most ubiquitous data sets is Expected Goals. It’s everywhere. Sky/BT broadcasts, columns, social media, an awful lot of air and virtual ink have been dedicated to what was “expected” to happen rather than the reality of what did.

The idea is to bring more objectivity to the proceedings. To take a step back. To view the long game. To try to figure out over the course of a significant sample size whether a player is good or lucky or whether a team is bad or unlucky. Well, perhaps not truly understanding, but to better understand — and from a club’s perspective: To reduce the margin for error.

xG has pushed the use of data in football to a new level. Spin-offs like expected assists, the expected goal chain, and expected points have joined the party.

Today, it is cited in a crude way, with people using it as an example of who ‘should’ have won a specific game. More xG in an individual game does not mean a side should have won.

Expected Goals are better used as a way to track patterns and trends. The more data you chuck into the xG model, the better it is: it’s why expected points, built around the amount of xG a team scores and concedes, is such an interesting measurement over the totality of a season. A one-off sample is poor: perhaps a coach was allowing certain players into certain areas because they’re weaker on their left-foot, for instance. Such efforts might garner a high xG score, but might overlook the fact the player is indeed useless with his left foot or that it was part of a game plan.

Packing looks to dig down an extra level beyond xG. It details how a team was able to move the ball into a position to be able to achieve an expected goal.

There’s another distinction, too. Expected Goals is ostensibly a team metric that is then dressed up as an individual one — the place from which a ball is shot and how the rate of success from that position compares to historic data. From that point, it tries to unpack the cause and who was responsible: The player who took the shot is given the credit; then there are Expected Assists — the person who passed to the shot taker; Then the xGChain — the people who passed to the person who passed to the shot-taker.

Packing works in the reverse order. It is a player-driven metric that combines to show the effectiveness and stylistic makeup of a team. It can build in the tempo and patterns of play. It can isolate which defender is responsible for the most important ball progressions, not just the volume of ‘progressive passes’.

Its root notes are fairly rudimentary, even if sometimes they’re swimming in jargon. The theory is simple: fewer defenders between a player and the goal means a player is more likely to score. Thus the best way to score is to cut out as many defenders as possible with the ball. If you play a pass that bypasses — ‘Packs’ — two defenders, then you have effectively eliminated them from the game and increased your team’s chances of scoring.

To turn the idea from a theory into a metric, the creators assigned a score value. A player receives a point for every opposition player that their pass or dribble bypasses.

Below, player 1 stands with the ball and rolls it across the pitch to player 2. Though the ball may have moved forward, he has not bypassed any defender, so he’s given a Packing Score of zero. Player 2 then fizzes a ball between the first-line of the defending team’s press, bypassing three defenders — he is given a Packing score of three.

The added bonus of Packing as a player analysis tool: It awards a Packing Rate to both the passer and the receiver of the pass. Rather than just give points to the player who passed the ball, it also awards value to the player who moved to create the kind of space that would allow a defender to be bypassed by the passer.

Think of Liverpool’s classic 4–3–3 look: Roberto Firmino drops off to create a midfield diamond, the two wide forwards drive from outside-to-in to narrow the shape; the fullbacks initiate play from outside before the midfield options bounce the ball through the middle. The key component of that traditional set-up was Firmino and his movement — it’s why he was dubbed ‘the system’. But trying to lock down in a tangible way the benefits of his movement has always been tough. Understanding why the one-one-one look is so effective, and Firmino’s movement so pivotal, has always been a scouting/coaching/opinion exercise:

Packing assigns a value to that movement — perhaps not whole, but better than the alternatives. Let’s return to the previous example. The pass from player 3 is given the same score value — four — as is given to the withdrawing forward — play 4 — who is able to pick up the ball between the lines.

They’re both awarded credit. The data set (‘KPI’ in Impect’s parlance) takes into account the importance of players in advanced positions being able to identify and occupy spaces between the lines. A progressive pass that gains a high packing score is less likely to be effective if the ball is played into a congested area and is not completed.

(The data will also differentiate between bypassed ‘opponents’ and bypassed ‘defenders’ — those in the last line of defence. The latter being weighed as more valuable — creating a chance for a striker clean through on goal — then the former — one defender rolling the ball to the other to beat a lone-pressing forward)

Two of the original demi-Gods of Packing data now set-up shop in Liverpool’s midfield: Naby Keita and Thiago Alcantara. Keita was all-vertical-stuff, all-the-damn-time when he was playing Germany: he was a crafty, roving, on-the-move playmaker who flitted between the lines.

The eye test said Keita was the real-deal; the numbers backed it up.

The same was true with Thaigo, though his brilliance was less of the jump-out-of-the-screen-and-hit-you-in-the-face, best-player-on-the-playground variety and more of a droning, melodic, sophisticated excellence.

Packing was not the only data set that highlighted the pair’s brilliance, but it helped bring an objective measurement to their intoxicating styles; it helped quantify the effectiveness of those between-the-line shuffles and no-look passes.

Over his time in Germany, Thiago developed into a different player than the one that first burst onto the scene at Barcelona. He kept all the metronomic precision of Xavi and Iniesta and those Barca midfielders that defined an era, but he added tenacity and bite to his game to go with it.

He loved to head the ball. He played with energy. He shoved people out of the way. He had no time for dawdling.

Thiago was held in such high esteem by Pep Guardiola — was considered the focal point of the entire Pep Takes Germany project — precisely because he could blend both the traditional, melodic, symphonic style that was then viewed as PepBall and the heavy-metal, quick-transition Ragnick/Klopp style that had by that point taken over the German game.

Thiago remains that blend today. It’s what makes him so effective. He can play in any style, happily togging to whichever is needed on a possession-to-possession basis. Want me to slide into a back-three to set the rhythm of the build-up play? No problem. Want me to slide into a two to better protect against counters? Happily. Want me to push up into more of an eight role so that I can better play with the right-back and right-sided forward? I’d be honoured.

And yet there have been undeniable struggles early in his Premier League experience.

An overinflated reputation? A result of the struggles that have torpedoed everyone in a Liverpool kit this season? A difficulty to grasp Liverpool’s system? What is it?

In his final season in the Bundesliga, Thiago was without question the most influential midfielder in world football. He didn’t just top the Packing metrics, he nearly shattered the system. Bypassing opponents, receiving the ball by bypassing opponents, measuring which players at the heart of the action were most likely to do the two most essential things a midfielder can do, Thiago came out on tops.

He wasn’t good. He was the best — by a decent distance; and this in a league loaded with talent in the midfield strata.

And despite some early struggles in a Liverpool team crumbling all around him, Thiago is still posting some pretty massive Packing metrics. On the ball, he is as good as ever. Thiago is averaging 54 bypassed opponents per game, according to data provided to Liverpool.com by Impect . That’s more than Jordan Henderson, Kevin de Bruyne, and Rodri. It’s double that of Ilkay Gundogan. It’s almost three times more than Gini Wijnaldum — Wijnaldum averages 19 bypassed opponents per game.

It’s not just about positioning, either. Only Kevin de Bruyne (12.1) and Jordan Henderson (9) average more bypassed defenders — the last line of defence — per game than Thiago among Manchester City and Liverpool players, the two clubs with the most bypassed opponents and defenders this season. Thiago is currently averaging 7.4 bypassed defenders per 90, a good if not otherworldly total.

As a comparison, Wijnaldum bypasses 2.7 defenders on average per game, Gundogan 4.9, and Curtis Jones 3.9.

Where Thiago’s numbers have taken a slight dip is in receiving the ball between the lines. He has rarely received the ball beyond the last line of defence but he has alway bossed the ‘bypassed opponents as a receiver’ total. It defines his game: give-go; pass-slip; pass-relocate, gather again, play again. Shut your eyes and you can picture him strutting and moving between the lines, and moving the ball at pace.

Thiago’s numbers are still good — very good, in fact — but they’ve tumbled from his Bayern days. He currently averages 37 bypassed opponents as a receiver, second only among Liverpool’s midfielders to Curtis Jones (Jones is really, really good). It’s a high mark, but it’s some distance away from De Bruyne (56 per game) and Gundogan (59 per game), the two players leading this season’s totals.

The underlying data points to the same conclusion as the eyes, but perhaps a bit more forcefully: Thiago is the same player with the ball for Liverpool that he was with Bayern, and he’s been a different player without the ball.

For Klopp to get the Thiago — the Bayern Thiago, if you will — he faces a choice: Either he completely re-models his classic 4–3–3 look, handing Thiago a free role at the base of midfield alongside essentially two holding midfield players. Or he keeps Fabinho in midfield, adds a consistent shuffler on the left-hand side and frees up Thiago to move wherever he likes from a right-side starting point.

So far this season, Klopp has tried both. Early on, he stuck Thiago at the base of midfield. It was a bit of a mess; neither the conductor nor the orchestra could get on the same page. Thiago is running, on average, twice as far in Liverpool games this season as he did for Bayern the year before — Thiago’s strength is in subtle movements, being able to find a sliver of effective space and then getting rid of the ball ASA-and-P, not having an ocean to drive into or to cover.

Part of that is about the stylistic differences between the two leagues and the two teams he played for; part of it was about coming to grips with the idiosyncrasies of Klopp’s counter-pressing system; part of it is because Liverpool’s defensive shell has so often ruptured that the midfielder has been forced to cover a a bunch of ground. He hasn’t helped the team, and the system has been built to get the most out of him.

Throughout his career, Thiago has proven to be a switched-on and solid screening midfielder. He wins with anticipation and speed. But he does so in short spaces. He isn’t the man to cover up holes sideline-to-sideline at the base of midfield. When he was given the free, lurking role at Bayern, in came in one of three forms: Thiago alone on one level with five or six attackers ahead, allowing Bayern to press the ball from the front; Thiago with a shuttling midfielder alongside in a two-man base; Thiago in the centre with a pair of inverted fullbacks drawn into the middle of the pitch to help cover the space.

Klopp has tried some creative solutions to bridge the divide. He has rolled out a fungible 3–2–5 system that sees Liverpool morph in shape while on the move: Five players occupy the five attacking zones across the pitch; Thiago and Gini Wijnaldum set-up base camp in the middle of midfield; two centre-backs stand at the back; Trent Alexander-Arnold is given a free role to zoom from the backline (the third player in the back-three) all the way up to the forward line, or he can stand on the midfield plane and act as a creator/counter-attack protector.

It’s shown signs of promise. But the true the-lights-have-gone-on moment came in the middle of the week against RB Leipzig. With Fabinho moved back into midfield and Wijnaldum operating as the back-and-forth commuter on the other side, Thiago was free to do Thiago things — at his own pace, in his own time, finding his own space. It was his most all-around performance of the season to date; Chelsea away probably remains his best ‘Thiago’ performance.

The Leipzig game, like the Chelsea one, was a jump-off-the-screen moment for Thiago in a Liverpool shirt. As the data suggests, the same old, run-the-show-in-the-Champions-League final is right there, burrowed under the surface. Now, it’s up to Klopp and his staff to figure out a way to put the most effective midfielder in Europe to use.

This is part one of a three-part series based on the ‘Packing’ data set. Part Two, on Joe Gomez and Liverpool’s defensive line, will be out on Sunday morning. Part Three, on Packing’s use in scouting and the future of data will be out on Monday.

Originally published at https://www.liverpool.com on March 13, 2021.

Senior Football Analyst at Cox Media’s sports vertical’s: All-22 (NFL) and SEC Country.

Senior Football Analyst at Cox Media’s sports vertical’s: All-22 (NFL) and SEC Country.