So, I wanted to say a few things about xG overperformance when a small number of matches have been played.

Firstly, overperformance on xG can be the result of just one match at this stage of the season.

I can illustrate this with Crystal Palace (see next tweet).

#cpfc

1/
Prior to Sunday's match at West Bromwich Albion, Crystal Palace had an UNDERperformance on expected goals of 0.4 this season according to the @Gracenotelive model.

After Sunday's match, Palace now have an OVERperformance of 3.5 goals from their attack.

#cpfc

2/
What if we look at the expected goals table and actual table?

Palace are 11th on the actual table and 13th on the expected goals table (incl. Sunday's match).

So, is it really fair to talk of Palace as overperforming this season?

I would argue that it isn't.

#CPFC

3/
Similarly, almost all of Tottenham Hotspur's overperformance in attack is due to 2 matches - their 6-1 win at Manchester United, where they admittedly created 4.4 expected goals and the 5-2 win at Southampton.

This accounts for ~4.5 of their 5.1 goal overperformance.

4/
Southampton, on the other hand, do look like they are genuinely out performing their offensive expected goals.

#Saintsfc have scored more than their expected goals in 8 of their 11 matches.

The question is why? Luck is a possibility but are there others. Southampton fans??

5/
What about the other end of the table? Sheffield United's goal total is about half of their expected goals.

On the basis of expected goals (& without having seen the games which is always a bit hairy), Sheffield Utd arguably did enough to beat West Brom and draw with Leeds.

6/
These extra 4 points, if they really were deserved, would put Sheffield United second bottom so not THAT much difference from where they are.

Worth noting too that West Brom scored early against them which can have a big effect on a match.

7/
Another point - particularly this season - is penalties as @Torvaney has pointed out to me.

Leicester have had 8 penalties this season. If we are trying to draw conclusions about a team's quality, we have to take these out unless we can see the Foxes getting 27-28 pens.

8/
Leicester's 8 penalties produce an expected goals total of 6.32 goals. They have scored 7 times from the spot.

That overperformance is roughly 25% of Leicester's total xG outperformance this season!!

Penalties account for 18% of Brighton's defensive xg underperformance.

9/
So, in short:

- Be careful with making conclusions on future performance from expected goals after only a few matches.

- Look deeper if you can. Is the over/underperformance due to 1 or 2 games? Penalties? Something else?

- Try not to use the word 'luck' without digging

10//
You can follow @SimonGleave.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled:

By continuing to use the site, you are consenting to the use of cookies as explained in our Cookie Policy to improve your experience.