Contact us for a free demo Contact us
for a free demo
StatsBombIQ StatsBomb Data

Chelsea: Season Preview 2020-21

By Nick Dorrington | September 1, 2020 | Premier League

Chelsea came through their transfer ban season unscathed, securing Champions League football with a fourth-place finish in Frank Lampard’s first year at the helm. That proved to be the understated prelude to a transfer window of heavy spending that has raised expectations ahead of the new campaign.

There are two opposing ways of assessing last season. You either think Lampard did a very good job in steering a relatively young team, written off by some, and shorn of Eden Hazard, their primary attacking contributor, to a top-four finish; or, you think that despite Hazard’s departure, the strength of the squad was undersold, that Chelsea had some clear faults Lampard struggled to fix as the season went on and that he somewhat lucked into a top-four finish due to Arsenal and Tottenham’s woes and Leicester’s late collapse.

There are other considerations on both sides of the argument, but the reality likely lies somewhere in between: he performed about par given the quality of the players at his disposal.

Chelsea secured six less points than under Maurizio Sarri in 2018-19, scoring six more goals but conceding 15 more. You have to go all the way back to 1997 to find the last time they conceded more than their total of 54 goals (although they did let in 53 in finishing 10th in 2015-16).

The underlying numbers offer more support of Lampard’s work. This wasn’t a repeat run of his 2018-19 season at Derby, where his team converted mediocre metrics into a spot in the playoff final. His Chelsea side actually had the third-best expected goal (xG) difference in the Premier League.

Chelsea regularly dominated the shot count. Only Manchester City took more than their 16.29 per match or conceded less than their 8.32 per match. The transitions forward were slightly quicker and more direct than under Sarri, and there was less of an overt focus on possession, although Lampard did continue the shift to a more aggressive high press.

That was the main commonality with Lampard’s Derby team, who were also more aggressive in contesting possession high up the pitch than an average Championship side. Unlike at Derby, this approach was successful in suppressing opposition shot volume, but with the classic tradeoff: the average quality of the shots Chelsea conceded was amongst the highest in the league.

It was on the defensive side that Chelsea severely underperformed their metrics. They were about par in attack, scoring 62 non-penalty goals from 63.06 xG; in defence, they conceded 16 goals more than the expectation: 52 non-penalty goals from 36.03 xG.

We’ve been clear from the start that we believe Kepa Arrizabalaga to be average goalkeeper at best, and the numbers do suggest he was a negative contributor last season. It still isn’t enough to account for all that difference.

It’s also notable that this has been a thing at Chelsea for some time now, even before Kepa’s arrival in the summer of 2018. From the second half of 2017-18 onwards, they’ve pretty consistently underperformed their metrics, with the bulk of that attributable to the defence.

A small part of last season's difference can perhaps be assigned to set pieces, where Chelsea conceded 13 times (only two teams conceded more) from 10.51 xG. That is an area in which you'd usually expect to see most variation from the model. The tape does also suggest that they were pretty open in midfield. The previously bulletproof N’Golo Kanté missed 16 matches through injury, and covering large swathes of midfield territory in transition is not exactly Jorginho’s strongpoint. The backline was often left exposed, given impossibly large spaces to defend.

Now, our xG model already includes goalkeeper and defender positioning, which is a huge help in determining how clear chances really are. Perhaps there is something on the margins of what the model picks up, like balance, relative body positioning, etc, that has made a small (and it would be small) contribution to the difference? Or maybe last season was just the latest episode in a particularly long run of bad luck in terms of opposition finishing?

We won’t get a stable sample with which to test those theories because it will be a very different Chelsea who step out onto the pitch in 2020-21. With purse strings tightening across the continent, Chelsea have found themselves in the enviable position of being one of the few sides of genuine means with funds freely available. They’ve already laid out north of £130 million on three new arrivals, with a further £90 million soon to put down on Kai Havertz. Add to that the free transfer signing of Paris Saint-Germain central defender Thiago Silva, whose wages presumably aren’t cheap, and it's quite the outlay.

The three new attackers all look a good fit for the play style, coming off of some of the most aggressive high-pressing teams in their respective leagues.

Timo Werner is real thoroughbred, a striker all the continent’s biggest teams should have been looking at. He combines relentless running into the channels with high-volume and high-quality shot output. He is a direct upgrade on Tammy Abraham, who himself scored a very respectable 15 goals last season, and also someone who could easily play alongside him.

Hakim Ziyech is a high-usage, high risk/reward passer and shooter. You might think his high shot count last season (4.89 per 90) was conditioned by Ajax’s strength relative the rest of the Eredivisie, but he got off even more on their run to the final four of the 2018-19 Champions League. In fact, he’s always been a high volume shooter, even back to his Heerenveen and Twente days. How he’ll work out on a team who had a fairly equitable share of shots between their players last season remains to be seen.

Havertz is a tall, skilled and deceptively quick player who fulfilled a variety of roles at Bayer Leverkusen last season. He saw minutes as a right winger, central attacking midfielder and even as a false nine. He provides both shot and creative output, tallying 11 goals and six assists from 14.61 xG and xG assisted last season, and might be ready to explode.

Those three will all have to fit in around Christian Pulisic, who produced over a goal or assist for every two matches he played in terms of both expected and actual output in his first season at the club. Lampard seemed hesitant to start him during the early part of the campaign, but he got two good runs of starts either side of his mid-season injury, and established himself as a key player down the final stretch.

Lampard now has a superb set of options for the four forward-most positions, and there is enough wiggle room in their respective skillsets to fit them all together into a somewhat coherent system. There is also natural symmetry there, with Pulisic and Werner generally preferring to receive towards the left; Havertz and Ziyech, the right.

Elsewhere, there is no replacement yet for Kepa nor a capable ball-winner in midfield. In addition to question marks over whether Lampard's system makes best use of Kanté's attributes, there may be concerns that last season’s fitness issues were the early signs of physical decline. The money spent in the final third might prove a little superfluous if those needs are not also addressed.

There have been reinforcements in defence. Ben Chilwell is a good, solid Premier League full-back, better going forward than he is at defending. He's not a huge upgrade on what Chelsea already had on the left, but he is both homegrown and younger, which may be justification enough. He’ll likely be a relatively dependable performer, and he still has time to refine the defensive side of his game, but the feeling persists that they might have found someone a little more inspiring at that price point.

Thiago Silva obviously has great pedigree and even at this late stage of his career probably represents an improvement on last season’s starters. But at 35, soon to be 36, any physical decline is likely to be a bit more obvious at Chelsea than it was at PSG. Malang Sarr, another free transfer arrival, is an the opposite end of the age curve. Now 21, he's been a starter at Nice since he was 17. There are some suggestions he might be loaned out, and this could simply be Chelsea taking advantage of value in the market with an eye on future profit, but he seems to roughly fit the profile of what they might be looking for in a central defender.

The incomings raise doubts over the future and development of some of the younger players who got good game time last season, particularly Callum Hudson-Odoi, who produced solid output when he made the pitch but never seemed to totally convince Lampard. It is also likely that Abraham and Mason Mount will see their minutes reduced. The enforced use of internal options has, though, had clear benefit. Chelsea now have good, first-hand information on a group of young players who, if nothing else, have had their market value raised by a year in the first team.

So where should expectations be set for Chelsea this season? The points gap to Liverpool and Manchester City is probably too large to overcome in a single campaign, even if all of the new signings do immediately click, but the level of their spending suggests that reducing it will be the minimum requirement. A third-place finish seems the most likely outcome. If that isn't achieved, and previous problems also persist, Lampard's job may well be at risk.



If you're a club, media or gambling entity and want to know more about what StatsBomb can do for you, please contact us at

We also provide education in this area, so if this taste of football analytics sparked interest, check out our Introduction to Football Analytics course

Follow us on twitter in English and Spanish and also on LinkedIn


Article by Nick Dorrington