Still not over that performance from yesterday, so I thought I'd throw myself into a little look at our defence (+ GK) over the course of this season to distract myself. For context I also took a look at both Manchester City and Liverpool -- so some notes on them a bit later.

In the 49 games I looked at, only Azpi (RB) & Christensen (RCB) have started 20 games or more in a single outfield role.

NB. I have split the CB position into right centre back (RCB) & left centre back (LCB). I am treating CBs in a back 3 as different roles (RSCB/MCB/LSCB).

Liverpool, unsurprisingly, have VVD (LCB/46), Trent (RB/41), Robertson (LB/41) & Gomez (RCB/25), with a minimum of 25 starts. While City have 3 players (Mendy, Walker & Otamendi) who have started 20 or more games in a single role, with Fernandinho starting as a LCB 19 times.

Here is a breakdown of the roles Chelsea players have fulfilled this season, with the number of starts and whether we were playing a back 3 or 4.

Here is a look at every combination Chelsea have played with as a back four this season.

A look at every combination Chelsea have played with as a back five this season.

Liverpool, however, have had virtually the same back four start in ~70% of their fixtures. The only changes they make are at right centre back. Consistency is key.

City, much like Chelsea, have not been able to settle on a clear set of defensive personnel this season. Their top selection has only been deployed FOUR times → Mendy (LB), Fernandinho (LCB), Otamendi (RCB) & Walker (RB).

Perhaps Chelsea's biggest issue is not settling on a centre back pairing with defined roles. 7 different combinations in all competitions (6 if we remove Guehi's starts). Switching CBs from one side to another is also slightly confusing.

Every Liverpool "starting" combination of van Dijk at LCB with Gomez/Matip/Lovren at RCB has been used individually more than any Chelsea pairing this season. Klopp has more consistency while rotating than Chelsea do looking for the answer.

Even the Fernandinho (LCB) and Otamendi (RCB) combination has started 11 (!!!) games for City this season. It is also worth mentioning that in 2018/19, when City won the title, their defence had far less fluctuations in team selection. Laporte/Stones/Kompany featured regularly.

From a Chelsea perspective the defensive frailties we have seen this season naturally go beyond the defensive personnel. We have largely had mediocre/average holding midfield play and an objectively poor goalkeeper this season. These have huge impacts on defensive structure.

Nevertheless, the lack of consistency in back four (or even back five) selection should be cause for concern. Liverpool are the benchmark when it comes to playing an established group. City in 2019/19 were similar and Chelsea's title winning teams were the same.

Not settling on a CB combination has been a huge factor, as this tends to be the foundational platform of your entire defensive structure. Equally, the relationship between RB/RCB and LCB/LB has hardly had time to take shape.

There is very little chance to establish a centre back partnership and build chemistry between centre backs and full backs, when the same defensive setup has only been picked (at most!) 6 times all season. This naturally filters through into midfield and forward play as well.

We must try to find some semblance of continuity for the rest of the season - which is difficult given form and fitness. However, going forward it must be a priority. There is little chance of establishing a rhythm in possession (and out of possession) with the flux we have seen.

Chelsea still need to look at a goalkeeper and the profile of who is playing in the holding role. There are still enormous issues in the final third -- did we create anything consistently of note against Sheff Utd? Or look like scoring?

If there is one thing to be taken away from this thread, it's that consistency of selection in terms of defence generally correlates with your team performing well. We haven't had that this season, & it should be something we address for the remainder of the campaign + beyond.