How has Chelsea’s switch to a back three affected the Premier League and has this had any impact from a trading perspective?
After a decade dominated by formations like the 4-4-2, 4-2-3-1 and 4-3-3, three defender formations have seen something of a revival over the past couple of years. In his final season at Barcelona, Josep Guardiola saw the 3-4-3 diamond as the next evolution for his team, while Louis Van Gaal utilised a 3-man defence with Holland at the 2014 World Cup to some success. However, the trend only really started to hit the headlines when Antonio Conte radically switched to a 3-4-3 formation after 6 games of the 2016/2017 Premier League season. Using Stratadata, I will assess how the tactical switch improved Chelsea, while also looking at the potential impact the current trend of 3 defender formations may have on trading markets.
To begin with, let’s have a look at their actual performance over the first 12 games of last season. In their first 6 games, Conte toyed with a 4-1-4-1/4-2-3-1 system, picking up 10 points and 10 goals along the way. Back to back defeats to Liverpool and Arsenal proved to be the spark for Conte’s decision to switch to a 3-4-3 and it proved to be a pivotal point in the season. For comparisons sake, in the next 6 games, Chelsea took maximum points, scoring 7 more goals while conceding none, but more interestingly, Conte’s men went on a 13-game win streak which ultimately helped them on their way to the Premier League title.
So how did the 3-4-3 actually improve Chelsea’s performance on the pitch? One of the main reasons for Conte’s switch to a 3-4-3 was to get Hazard, Pedro and Diego Costa playing closer together. With an additional defender, the wingbacks could afford to stay higher up the pitch to provide width, which in turn, creates more space for the front three to link up in the centre. This proved to be the case as before the switch to a 3-4-3, Chelsea made key entries into the final third from a central position on average, 36% of the time. After the switch, that figure went up to 43%, with Hazard, Diego Costa and Pedro seemingly given more space to attack.
Staying on the topic of key entries, but instead looking at the defensive side of things, it seems that opposition teams tried to counter Chelsea’s 3-4-3 by exploiting the attacking nature of the wingbacks, having more success down the right flank (33% compared to 41% with a 3-4-3) However, this explains one of Conte’s most intelligent decisions; utilising Azpilicueta (a natural fullback) as a right-sided centre back meant the defender could provide natural cover for Victor Moses from a position he’s fairly comfortable and accustomed to. The extra space available to opposition teams on the opposite flank could be explained by the fact that Gary Cahill (a natural centre back) tends to play on the left side of the back three, as the defender probably feels more comfortable defending from a more central position to Azpilicueta.
Interestingly, Chelsea’s desire to play through the centre of the pitch has had a slight effect on their corner count as they’re less likely to force corners from blocked crosses for example and on average, they received 1.5 less corners per game in 6 games after the switch to a 3-4-3, so it’d be interesting to look at this on a broader scale for a future blog.
Moving on to StrataBet’s chance data, there was also a vast improvement in the quality of opportunities Chelsea created. They did have slightly more shots on goal in their 6 games prior to switching to a 3-4-3. However, when you compare the number of great chances they created before and after the tactical adjustment, there’s no denying Chelsea became much more efficient at creating higher quality chances, which naturally lends itself to scoring more goals.
At the other end of the pitch, the 3-4-3 also had the desired effect. Not only did they concede less shots overall, after conceding 5 great chances in their first 6 games before utilising a 3-4-3, they conceded none in their next 6 games after switching their formation. The number of very good chances also decreased so it’s no surprise they kept an incredible run of clean sheets with the addition of another defender.
From a trading perspective, Chelsea’s increased creative efficiency and defensive solidity also had an impact on the pre-match total goals market. All 6 of Chelsea’s opening games of last season finished over the total goals line which ranged from 2.5 to 2.75, but in their next 6 games, only half went over the total goals line despite it naturally falling to 2.25 goals on 3 occasions, which was probably due to their opposition struggling to create anything clear-cut.
|First 6 games of 2016-17 Season|
|Home Team||Away Team||Total Goals Line||Result||Over/Under|
|Games 7-12 of 2016-17 Season|
|Home Team||Away Team||Total Goals Line||Result||Over/Under|
It’s amazing to think such a simple switch in a team’s shape can have such big consequences. After the 1st of October 2016, when Chelsea first lined up in their new formation, more and more teams have experimented with a back three. Most notably, Tottenham Hotspur ended Chelsea’s 13-game win streak with a similar formation. The trend appears to have continued into this season, with the likes of Manchester City, Everton, Burnley, Swansea and Bournemouth all using three at the back at some stage, but has this had a broader influence?
In total, 33 Premier League games this season have featured either one or both teams using three at the back and of those, 15 games have finished over the total goals line (45.45%), while 16 have finished under (48.48%), with the 2 remaining games earning a push. Interestingly, the number of matches not featuring a back three finishing under the total goals line jumps up to 17 games out of 27 (62.96%), with the other 10 games finishing over the total goals line (37.04%). What’s even more interesting is the fact that the average total goals line for games between teams playing four at the back (2.45 goals) is actually lower than the average for games featuring a back three (2.62 goals).
It could be that the bookmakers are still slightly over-estimating the total number of goals in games not featuring a team playing three at the back and this could suggest that although more teams are experimenting with a back three, they may actually be more comfortable and efficient defending in a back four, mainly due to the unfamiliarity of 3 defender formations over the past 10 years. Perhaps it will take some time and more training than expected to get teams defending efficiently in a back three which could open up trading opportunities on under markets whenever a team experimenting with 3 defenders reverts to their usual back four.
Ultimately, it appears Conte’s experience and familiarity with coaching a team to play three at the back at Juventus and Atalanta has paid dividends in ensuring Chelsea’s quick transition and instant results. It’s unlikely other teams will experience the same drastic improvements with such a simple change, but it wouldn’t be a surprise to see more teams using 3 defender formations more frequently and efficiently over the remainder of the season.