8 Comments
User's avatar
Ben's avatar

Man, these KPI posts are great checks of the eye-test and vibe-test vs. cold hard data. Always one of my favorites from Cannon Stats.

Expand full comment
Scott Willis's avatar

Thanks! I agree it really helps check how things are compared to the "vibes"

Expand full comment
Seth Thomas's avatar

Scott - there's definitely been more Arsenal penalties in this run of matches. I understand why the KPI was built on non-penalty XG, but is this a flaw in the metric - because once you get a penalty, you have less need to score a goal. Just wondering if the awarding of more PKs has a disruptive effect on your analysis or you think the KPI already accounts for this.

Expand full comment
Scott Willis's avatar

There is absolutely I think something to this. If they don't get the penalty (and goal) that does change the way they play.

I strip out penalties because it is a bit random from year to year. I can see both sides of wanting to include them (and I do, both)

Expand full comment
Seth Thomas's avatar

Bournemouth might be the prime example. a 3.6 Arsenal XG is lowered to nearly 2.0 npXG because of the two kicks from the spot. And especially as Arsenal were up 2-0 and then 4-0 on pks, you would expect the offensive stats to look fairly tepid if you exclude the two penalty kick conversions.

But I think the use of a 10 match npXG average helps offset some of the impact of this, because certainly you would expect that over 10 matches the team's performance is not skewed by the award of a penalty in a particular match

Expand full comment
Rory's avatar

20% title odds feels like a really good place to be.

Expand full comment
Scott Willis's avatar

Yeah it's not too bad. City at roughly 50/50 with some chasing teams and Arsenal leading the pack. I will take it for now. Let's see if we can keep picking up points as we get to the halfway mark.

Expand full comment
Reggie Perry's avatar

Considering our starting attack trio have barely played together this season, we’re doing quite well.

Expand full comment