This article should more accurately be titled ‘Why I’m a bit sceptical about Arsenal being title contenders but they might be but I think they probably won’t be and not because they can’t keep their first team fit’. But that’s a bit long.
A few weeks ago, I wrote this http://eastbridge-sb.com/are-arsenal-really-title-contenders-thursday-29th-october-by-etnar_uk/ on the same subject, which ended with a rather piping hot take, so I should make it clear that I don’t have an agenda against them. Actually, Manchester United aside, they’re probably my favourite of the traditional Big 4 contenders.
This season, they have good shot and expected goals numbers, hinting that their good form over the first twelve games may be genuine ability this term. However, I recently introduced the idea of what I’m calling ‘statistical possessions’ and ‘statistical attacks’ here https://everyteamneedsaron.com/2015/11/08/extending-the-chain-of-conversion/. Better teams, largely, seem to turn possessions into attacks more often and attacks into shots more often, and do the reverse on defence (which is what leads to good shot numbers). I think, to an extent, they also suggest a team’s dominance as shot ratios do too.
From the four full seasons of data I have (top 3 leagues 2014/15, and Premier League 2013/14), statistical possessions are converted to statistical attacks around 31% of the time, and s. attacks to shots around 21% of the time. As of having played 11 games of the season, 15/16 Arsenal are super-performing on 3 of the 4 ‘metrics’, I guess you could call them, namely: possession% (possessions converted to attacks); attack% (attacks converted to shots); opposition possession%; and opposition attack%.
On possession% Arsenal were at 37.13%, Chelsea (interestingly) second on 34.36%, a fairly large difference when the league average is at 32.5%. Within the 4 full season sample, that figure is 3rd, behind on Barcelona and Real Madrid in 2014/15. On attack% they would come 2nd in the 4 season sample, behind Real Madrid again. Offensively, Barcelona and Real Madrid are historically far more dominant in their league than Arsenal are in theirs. Maybe Arsenal truly are this dominant, or maybe those figures will settle down a bit soon, leading to fewer shots taken.
On the defensive side, their opposition possession% isn’t so great at all, their opponents converting 34.91% of their possessions to attacks, which is worse than the average (and 70th in the 4 season sample).
Their opposition attack% though is a far different story, at only 16.55%, which would be 3rd in the 4 season sample behind Atlético Madrid and Bayer Leverkusen. No team so far has had anywhere near as big a difference between their attack% and opposition attack% – Arsenal’s was close to 12% after 11 games this year, the biggest in the sample is under 7%.
I’m still tentative about drawing solid conclusions from these new statistics, because I haven’t investigated them in nearly as much detail as things like shot ratios and expected goals models have been, but from what I am gleaning from these, Arsenal’s shot numbers don’t seem like they’ll continue. Everyone else in the 11 game numbers had figures that I’d generally probably consider to be normal, so Arsenal’s stood out to me. Of course, their injury problems probably mean that they’ll fall off a bit anyway, but there ya go.
*I use the word ‘statistical’ in here because otherwise it might be confusing to otherwise talk of possessions and attacks in this way. I also say ‘introduce the idea’ because I’m not 100% certain that this is the best way of doing it, and I’m aware that you could subtly change how you calculate ‘statistical possessions’ and it probably wouldn’t make much of a difference, it’s not definitive as such like ‘shots’ are.