When he was signed last summer from Southampton, most observers thought that Victor Wanyama was an insurance policy to provide cover for Eric Dier in Tottenham’s midfield. The team would continue to play the same best XI (with Son Heung-Min and Mousa Sissoko rotating into the attacking band as needed) that we did last season. Wanyama was just (much-needed) depth.
However, due to injuries, that is not at all how things have worked out. We are yet to see last season’s normal XI play together even once this season. In fact, the New Year’s Day win against Watford was only the third time all season that Mousa Dembele, Eric Dier, and Toby Alderweireld have been in the matchday squad together. They have only played together once this season: in the win against Sunderland when Dier and Alderweireld started together in central defense, Dembele played in midfield with Wanyama, and Jan Vertonghen played left back.
The consequence is that, one way or another, Wanyama has been an obvious starter in almost every game this season. If Dembele is out of the squad, Wanyama is the obvious choice to fill in. If Dier is hurt, ditto. If Alderweireld is hurt, Dier has to drop off to play center half, creating an opening in midfield that Wanyama inevitably fills.
But Wanyama, as we have already noted, does not do all the things that the ideal Spurs midfielder would do, if last year is any indicator of what we need in midfield. If Wanyama is supposed to slot in for Dier, then he needs to shield the back line (check), win the ball in midfield (check), split the center backs (uhhhh), and push the ball forward into the attacking third with aggressive, vertical passing (nope nope nope).
Wanyama’s limitations on the ball became apparent very early in the season. Last season, Spurs averaged 16.2 shots per game (best in the league) and the average shot quality was .1, tied with five other teams for second best in the league. Spurs played only five games this season with Wanyama in the double pivot of the 4-2-3-1 and the results, from an attacking perspective, were not good. Though the team put four past an abject Stoke, they failed to score multiple goals in each of the other four matches.
The xG numbers were also poor: Though the team put up 2.6 and 2.5 xG against Stoke and Sunderland (Sunderland was also Mousa Dembele’s first league game of the season), they averaged only 1.3 xG in the first three games and the shot quality was also quite poor: .09 xG per shot against Everton and Palace and an abysmal .06 xG/shot against Liverpool.
Wanyama and the 4-1-4-1
It was no surprise to see Spurs move to a 4-1-4-1 in the season’s sixth game, a match against newly promoted Middlesborough. The benefit of the 4-1-4-1 was that it flanked Wanyama with two midfield players and took all the attacking responsibilities off his plate. All he needed to do was run around and break things at the base of the midfield trio. The initial returns were good. We beat ‘Boro and then picked up a comfortable 2-0 win against Man City. But the 4-1-4-1 was not a solution, as it turned out.
The 4-1-4-1 was great for launching quick counters b/c it fixed one of our besetting problems under Pochettino, which is the lack of personnel getting forward. It also freed Wanyama from the need to play as much of a passing role in midfield since we now had additional midfield players taking that responsibility.
However, the 4-1-4-1 also did not solve the problem with Wanyama splitting our center backs, so Jan and Toby (or Dier when Toby was hurt) were still not able to be as involved in the attack as they were last season. Moreover, it in some ways made a bad problem even worse when it came to our attack out of sustained possession. That sort of attacking is already a bit unnatural for us and the 4-1-4-1 simply made it even more difficult by packing another body into the attacking third. Now we had two “wingers” who both tended to drift inside plus two central midfielders pushing into the advanced central area.
The result was a lack of space for our playmakers and a maddeningly predictable attack that suffered both from a lack of consistent direct attacking via long passing out of the back and the aforementioned spatial issues for our key playmakers. We would continue to experiment with the 4-1-4-1 for a few more games, but it soon became apparent how limited it was as a system for us.
Wanyama and Three at the Back
It wasn’t long after the 4-1-4-1 experiments ended that we saw the team’s first use of a three man defense since a couple of odd experiments with the system last season. The 3-4-3 debuted against Arsenal at the Emirates, a fixture that Pochettino has always treated with a fair amount of caution. (Recall that his first ever North London Derby came early in his first season at Spurs and was also at the Emirates. In that game, Pochettino essentially set the team up in a Simeone-style low block 4-4-2 with Emmanuel Adebayor and Nacer Chadli as counter-attacking strikers.)
Heading into the game, Arsenal were on a tear offensively. They were averaging 1.7 xG per match and .11xG/shot—an extremely high value (.12 was the best value averaged by anyone for the entirety of the 2015-16 season). So Pochettino set Spurs up with a three man defense with Dier and Vertonghen on either side and Kevin Wimmer sitting in the middle. He then set Dembele and Wanyama in front of the back three and basically dared Arsenal to break Spurs down. The Gunners had some success that day, but their lone goal came via a Wimmer own goal off a set piece. By the standard set by other teams trying to thwart the Gunner attack, it was a successful day for Spurs.
The 3-4-3 returned a few weeks later for the match against Hull. However, instead of partnering Wanyama with Dembele, Pochettino went for a modified version of the formation that almost looked more like 3-3-3-1 with Christian Eriksen moving back and forth between a CM role and an AM role. Spurs won convincingly 3-0 and also posted their best xG total of the season and their best shot quality of the season.
After Toby Alderweireld went back down with a mysterious injury, the team reverted to the 4-2-3-1 shape with Dembele partnering Wanyama. But then earlier this week the 3-4-3 returned in the match against Watford. Once again Eriksen played a hybrid CM/AM role and, once again, it worked spectacularly well. Spurs again posted their best xG and best average shot quality numbers of the season.
Why does the 3-4-3 work so well?
Before we get too enthusiastic, we need to note that the answer to this question might be as easy as “because it was used against Hull and Watford.” But even taking that into account, there is a logical explanation for the improved results.
First, as we noted before, the 3-4-3 system naturally splits the center backs, making it possible for them to become more involved in the attack. There have only been two matches all season where Spurs’ shot quality stats matched those we averaged last year—and they are the two most recent games with the 3-4-3 where the defenders have become more involved in the attack.
Second, with a deeper passer behind him and Eriksen or Dembele next to him, Wanyama has virtually no attacking responsibilities.
Third, the system makes it easier for our fullbacks to get forward, providing necessary width for the attack, and allows the advanced midfielders to do what they naturally want to do anyway—play more narrowly.
Put another way, the 3-4-3 does the thing for Wanyama that the 4-1-4-1 did and solves the attacking issues that the 4-1-4-1 made even more pronounced and crippling. It’s no surprise that it has worked extremely well for us so far: It retains all the things that made us so good last season while successfully compensating for the change in personnel from last season to this season.
Should Spurs use the 3-4-3 against Chelsea?
The answer here is almost certainly “it depends.” The 3-4-3 we used against Hull and Watford would be a brave and probably foolish system to try against the hottest team in England right now and the probable champions of the league come May.
However, a 3-4-3 like what we used against Arsenal but with Toby stepping in for Wimmer might be perfect. We know from the Arsenal match that the 3-4-3 can stifle even the most devastating opposition attacks. We also know that the biggest danger in going 3-4-3—having Dier, Toby, and Jan isolated against Eden Hazard, Diego Costa, and Pedro or Willian—will be ameliorated by playing the Wanyama-Dembele duo just ahead of them. Both Wanyama and Moose are quite comfortable dropping deep and helping out the defense.
The other intriguing thing here is how this approach could unsettle Chelsea’s midfield. When we defeated Jose Mourinho’s eventual champions 5-3 on New Year’s Day, the key to the win was using Christian Eriksen’s runs off the ball to drag Nemanja Matic out of position. There may be potential to unsettle Chelsea’s current duo of Matic and N’Golo Kante in similar ways.
The Chelsea 3-4-3 works for a simple reason: Kante covers every square inch of midfield. Because of that, they can afford to only go with one other proper midfielder and to use wingbacks plus an attacking trident up top. Wingbacks plus the trio of Hazard, Costa, and Pedro up top leaves them light in the center of the park, but Kante’s range is so good that doesn’t matter.
Spurs, however, may be a more challenging order for Kante. First, the duo of Dembele and Wanyama may be the only Premier League midfield duo that can physically match up with Matic and Kante. So the Chelsea duo can’t simply count on out-muscling their opposition as they would against probably any other Premier League opponent. Second, if the 3-4-3 is working the way it ought to, we’ll see Dier and Vertonghen stepping forward to play long passes into the attacking third, skipping over midfield entirely. In that case, Kante’s activity won’t necessarily matter as much because we’ll be pushing the ball directly into the Chelsea defensive third more regularly and challenging Gary Cahill and Cesar Azpilicueta to deal with the in-form duo of Dele Alli and Christian Eriksen.
There is certainly a huge risk in using this system against a team as scorching hot as Antonio Conte’s Chelsea. Chelsea’s first full game in the 3-4-3 came 13 games ago against Hull City. Since the move to that system, their record is an obscene (wait for it) 13-0-0. In those 13 games, they have scored 32 goals and conceded four. Stretched over a full season, that translates to 94 goals scored and 12 goals conceded. It’s entirely possible that this Chelsea could go down as one of the all-time great Premier League teams thanks to Conte’s shift to the 3-4-3.
But the quality of the opponent may end up being a good argument to play the 3-4-3. We know what we have with a 4-2-3-1 system that uses Dembele and Wanyama in the double pivot. It’s perfectly serviceable, but not spectacular. (We could also trot out last season’s best XI with Son Heung-Min or Moussa Sissoko replacing Lamela, but trotting out the Dembele-Dier midfield for the first time in seven months against this Chelsea seems suicidal to me.) So, for my money, I think the 3-4-3 we played against Arsenal is probably the way to go. It won’t be as spectacular going forward, but if we’re going to get a result against Chelsea we pretty much have to keep a clean sheet because the odds of us getting more than one goal against the Blues are basically nil. A defensively sturdy outfit that frustrates the Chelsea attack and allows us to play a direct attacking style and get the most out of Dele and Eriksen is likely the best and perhaps even the safest move for this vital fixture.