Why Tottenham should stick with a back four

Introduction

In late August, Jake Meador wrote here about the two defensive systems Pochettino was experimenting with at the start of this season: a back three and a back four. Since then, Poch has been forced to make up his mind: we’ve started a back four in every match since Huddersfield on September 29, when Jan Vertonghen picked up a hamstring injury that has kept him from playing since then, and his return on Sunday against Arsenal was cut short by a red card. With Vertonghen out, only Davinson Sanchez and Toby Alderweireld remain of Pochettino’s preferred center backs, meaning that for important games, we’ve been more or less stuck playing those two in the center of a back four. In theory, we could add Juan Foyth or drop Dier in to make a back three, but playing three central defenders requires a specific skillset that only Alderweireld, Vertonghen, and Sanchez share.

So what have our performances using a back four shown us, and should we continue to use it now that Vertonghen is available? To find out, I’ve looked at the numbers from last season, and will supplement with some tactical research of my own.

Sidenote: while working on this piece at an unnamed coffee shop in a college town, a boy—a bro, more accurately, and probably a bro of a fraternity—and a girl sat down next to me, seemingly for a date. Through interruptions and rambling, the Bro’s loud bro-voice took over not only the entire conversation, but the entire coffee shop. Because it was so distracting that I couldn’t finish the article, I have transcribed the Bro’s highlights here:

“I’m from Orange County, which is known for its old people, and being Republican. My family pretty much moved there to avoid liberals.”

“Oh yeah I’m suuuper into Elon Musk. Ha ha ha yeah I’d say I’m a pretty big Hyperloop enthusiast ha ha ha.”

“I got my dad to buy me an Elon Musk flamethrower before he retired and it’s sick.”

“I live down the street from Elijah Wood, the blue-eyed guy who played The Frodo in Lord of the rings.”

“I don’t know if you’ve heard, but I’m related to [unclear civil war name]”

“Yeah there’s some pretty funny stuff in the Civil War”

“It’s not good for the alcoholic family history”

“One time I was in a museum and they had a huge rock. Like a huge rock. And I was just staring at it for so long that my dad had to ask if I was okay. So yeah, I’m pretty into museums.”

“You have to make sure you go to a top eight law school. My dad’s a lawyer. An entertainment lawyer. Very competitive, but he knows a lot of movie stars. So yeah, make sure you go to a top eight law school.”

Anyhow. Without any further ado...

The numbers

During the 2017-2018 Premier League season, Tottenham played 11 of their 38 games with three center backs (always Vertonghen-Sanchez-Alderweireld) and the remaining 27 with two center backs, which included Eric Dier at times. I analyzed both the goals scored and the expected goals (xG) from each game to get a sense of how the two formations compared. If anybody wants the full season spreadsheet, I’m happy to share it, but for now, here are the results, rounded to the nearest hundredth.

It’s worth pointing out that although we played just under 30% of our games in a back three, we faced Liverpool at home and Arsenal and Man United away in that formation, as well as a balance of other teams from across the table. This means that, crudely, there shouldn’t be too much of a skew in the results based on the opposition we faced.

By actual goals scored, we were far better offensively in a back four than in a back three, scoring an average of .56 more goals per game—about an extra goal every other game. We were marginally worse defensively, conceding an average of .96 in a back four, compared to .91 in a back three. I’ll delve into this more in the subjective analysis, but it makes sense that numbers-wise, the back four was far better for our attacking returns because it means we have an extra player in the midfield.

The xG confirms what the actual goals suggested: we played better, on average, in a back four than a back three. Again, the culprit was not actually defensive results, where the average xG perfectly predicted the average goals conceded per game at .90 in a back three to .96 in a back four, but offensive performances. We earned an average of .37 xG more when going forward in a back four relative to a back three—equivalent to one expected goal every three games.

Why a back four is better for Tottenham

Although defending is far nervier than attacking, and thus defensive mistakes stick more vividly in our memory, we actually played quite well on defense last season, conceding the third-fewest goals in the league (36). In individual games, we only conceded two or more goals in nine out of the 38 we played. The deciding factor in many games last season was our ability—or lack thereof—to score goals. Against Inter last week the same concern surfaced. Despite controlling possession and maintaining a strong position in the Inter half, it was only with the introduction of Christian Eriksen and Heung-Min Son around the 70th minute that Tottenham created real goalscoring potential. In the Pochettino era, we have often lost big games 1-0 or 2-1, and these one-goal margins reflect a strong defense with less offensive emphasis. The numbers show the same. It is also worth pointing out that the Spurs defensive game is just about as good as it will get, whereas I see plenty of opportunity for growth in the offensive game. Therefore, since we defend just as well in both formations, we would do well to draw on the back four’s attacking returns.

With the right personnel, a three-defender system can be extremely effective. Variants of it have been winning big games across Europe, but the success of the tactic depends on a team’s ability to fill specialist roles in defense. For a more in-depth read, check out the excellent SBNation article examining the three-defender system’s renaissance in football. In short, the system requires three center-backs with excellent positional awareness, speed, tackling consistency, and skill on the ball, because they will be under more pressure and their mistakes will be more catastrophic when they happen. Spurs have these three in Toby Alderweireld, Davinson Sanchez, and Jan Vertonghen. However, to really make the system work, a team also needs wing backs who can play the position well. The greatest advantage of a back three is that the fullbacks are able to push up the field into a wingback role, allowing them to function as midfielders during the attack. Arsenal did this successfully on Sunday, with Hector Bellerin and Sead Kolasinac attacking Spurs in our own half so effectively that once, Kolasinac played a cross in to Bellerin, meaning that both wing backs were playing as full attacking players. For another reference point, Kyle Walker might be the ideal wing back (although he hasn’t been used as one lately): he’s fast enough to track back when he loses the ball, but gifted enough in attack to function as an extra midfielder.

However, a traditional back four is offensively superior to a back three plus wing backs in Tottenham’s case for pretty simple reasons. We don’t have a stable of fullbacks who are good enough to play in a wing-back role. Trippier and Rose would probably be our best bet, but they play—and press—pretty much the same way in a back four as they would ahead of a back three. Crucially, unless your wingbacks can truly function as part of the attacking midfield, playing with four defenders creates a numerical advantage in attack relative to a back three plus wingbacks. This advantage occurs because without a third central defender, a team can field one more midfielder who can either operate in a double pivot in front of the defense, as Poch likes to implement, or cover another midfielder’s defensive duties to allow them to take up space further up the pitch. Several of our midfielders (Son, Lamela, Moura, Alli) are adept at playing both wide and narrow as needed, and can cover the parts of the touchline that would be occupied by a wing back, and with a back four, our front four—Dele, Eriksen, Son, Kane being a classic example—can be supported by a double pivot of Dier and Sissoko or Dier and Dembele.

As our defense changes and matures, the answers to Tottenham’s defensive tactical questions could change. How Juan Foyth and Davinson Sanchez develop, for example, could revolutionize our backs. Both have lots of raw potential, and although Sanchez is quite good already, there’s no sign that he’s slowing in his development. If he takes on new attributes, like precise distribution from the back or new positional sensibilities, it could free up players elsewhere for Poch to play in a different formation. Still, the deciding attribute of our defensive tactics will continue to be our fullbacks. Rose and Trippier might have the necessary attributes to play as wingbacks, and Aurier and Davies definitely do not. This isn’t to say that they’re bad players, but rather that they don’t contribute much as wingbacks that they don’t already as fullbacks.

In Conclusion

Spurs’ offense and goalscoring ability are key if we are to reach our aspirations this season. Last season, we defended quite well, but were frustrated in attack. The same was true on Sunday against Arsenal although we were touchy in defense in addition to attacking poorly. The defining feature of the North London Derby was the utter inability of the Spurs offense to link meaningful moves together and find shots on goal. We defended in both a back four and a back three, but neither formation altered the general dynamic of Arsenal’s defense stifling Tottenham’s offense. We need to consider how our defensive shape and transition to attack can contribute to more goalscoring opportunities, and although Jan Vertonghen’s return (now postponed by a red card) offers the possibility of playing a back three, to do so would be a mistake. Stick with a back four, figure out where we can improve that formation, and get the job done.