Weekly Dribble: Are Chelsea worse than we initially thought?

Chelsea, another odd tweet, and finally some goals

Chelsea: Bad luck, Moisés Caicedo and upcoming fixtures

If you've been following my content since pre-season, you know I was quite optimistic about Chelsea’s prospects this season. Especially regarding their defence—I predicted it to be one of the best in the upcoming Premier League campaign.

To provide you with the most surface-level uncontextualised data, they currently rank 10th in the xG conceded table. They also have the 4th-worst xG per shot conceded and have allowed about the league’s average volume of shots.

Given a favourable fixture run right from the start, you’d expect them to be able to pick up some momentum. Not this.

On the other hand, their 9.40 non-penalty xG means that only Manchester City, Crystal Palace, and Manchester United (see here for context, as it's needed) perform better.

To a large extent, it’s a combination of significant misfortune and only 6 Gameweeks of sample size.

There’s one clear reason their defence feels somewhat unstable—Levi Colwill’s ACL injury in pre-season was a decisive blow, leaving Chelsea without their leader and left-footed centre-back.

The Blues are feeling his absence in:

• on-pitch leadership

• aerial ability

• good decision-making off the ball

• passing skills

• threat during offensive set pieces / defensive security defending them

• … or the ability to drive the ball forward

He’s a complete package who improves the team in all phases of play.

Tosin has been injured, and several other players are facing fitness issues. The defensive stability depends on Moisés Caicedo, who was one of the most influential players in the last Premier League season and remains crucial in the current campaign. It’s almost unthinkable if they had lost him.

There’s a lot of hope that Benoit Badiashile will step up and build on the Benfica game—he has put in several excellent performances in the Premier League during his career. It’s a reasonable hope.

Additionally, Liam Delap’s early injury, which ruled him out until December, and Cole Palmer's groin injury caused a significant shift in dynamics.

The squad lacks sufficient depth to sustain these events.

Additionally, it’s worth noting that Chelsea received a red card in 33% of their league games.

All of these factors mess with data in a sample size of just 6 games.

They started the season with three solid defensive displays, and the dip afterwards was mainly due to circumstances. The fixture list looks pretty favourable after that Liverpool game:

• GW8: Nottingham Forest

• GW9: Sunderland

• GW10: Spurs

• GW11: Wolves

• GW12: Burnley

• GW13: Arsenal

• GW14: Leeds

Overall, this is an excellent run of fixtures for defenders from teams in the upper half of the Premier League table. So, what should you do with their defenders?

I like the move of those who went with Trevoh Chalobah (£5.2m, 8.6% TSB) in previous weeks. He has historically been a major threat from set pieces, both in movement within the box and finishing, and he’s also averaging 9.81 defensive contributions per 90 minutes. If you can weather his suspension, there’s no need to sell him, and if Caicedo stays fit, he's excellent for that upcoming fixture run. Apart from that red card, he’s had an excellent start to the season, especially when you watch closely what he does on the ball.

Marc Cucurella, who’s still owned by 23% of players, making him the fifth most popular defender in the game, isn’t much value at £6.1 million. I wouldn’t consider him or any other Chelsea defender apart from Chalobah from GW8 onwards.

As for attacking players, that’s a discussion for another time.

Florian Wirtz and the most outrageous tweet of the week

Alright, I know some people in the FPL community have already pointed this out, but enough is enough:

I realise it’s only Thursday, but you probably won’t come across a more misleading post on social media this week.

You know I’m trying to educate my readers about data analytics in football and how to use it properly. But it’s too damn difficult to spread the message when FOOTBALL DATA COMPANY posts this.

And conclude with the word “AGENDA” (!).

Because the agenda is in this post.

I don’t think OPTA are some evil company—honestly, I believe they do a great job most of the time. But this is just unnecessary engagement farming. And I fell into that trap…

The problem?

It assumes that all chances are of the same quality.

It’s the same with shots. If two players each have 100 attempts per season, they might have very different goal counts, as the quality of these attempts matters.

As for the chances, you have open play passes and those resulting directly from set pieces.

You have a drilled pass into the 6-yard box, and you also have a short pass that leads to a shot outside the box.

We have a wealth of other data metrics that can provide a pretty accurate picture of a player's creative influence. 

You can also combine these metrics to create more effective narratives. 

Or even better, you can examine how some metrics affect others.

Note that this rant isn’t aimed at Florian Wirtz. I admire him as a player, and I believe he will come out good eventually—he actually looked much better in the last three or four games. Liverpool are in a rebuilding phase, and they need more time. When he makes some tactical adjustments to his game, we will see the difference immediately.

This is aimed at creating false narratives with data “analytics”. It still happens too often for my liking. And it’s a shame that it comes from someone who should treat data more responsibly for many reasons.

After all, it’s your business.

More goals, more fun

Let’s wrap this up with a brief follow-up on last week’s newsletter.

If you didn’t read it, I discussed the lack of goals in the first five gameweeks, presented some data, and explored reasons for the year-on-year decline in the quality of chances, as well as why I believe the goals will come.

You can read it here:

Weekly Dribble: Poor rank in FPL? It’s not your fault — it’s the players

Compared to the first five gameweeks, last weekend was much more exciting:

• Gameweek 6 goals: 33

• Gameweek 6 non-penalty xG: 29

Over the first five GWs of this season, players averaged just over 21 non-penalty xG (and 21 goals), compared to an average of 27 non-penalty xG (and goals) over the previous three Premier League seasons.

Another reminder that everything is still distorted at the moment.

Onto the next!