This is a quick tie up of my previous two posts, the first about Everton and their apparent ability to shoot better at certain times of the year than others:
|“if Everton take a shot between December and April it’s 29% more likely to result in a goal than if they take a shot between May and November”|
The second being the variation (or lack thereof) of attacking events in the premiership by month:
|“in general the Premiership football you see in August is very similar to the Premiership football you see in May, and every month in between”|
In terms of the former I had a couple of suggestions to explain the apparent improvement of Everton on twitter so I’ll see to what extent I can address them:
|“could it be related to purchase of better strikers in January e.g. Jelavic this year, Kevin Campbell etc in the past?”
I can’t prove this for individual players but there’s several reasons I think we can rule this out. Firstly the timing doesn’t line up: Everton’s sh% takes off at the beginning of December, whereas the transfer window doesn’t close until the end of January (to pull the examples from that tweet Jelavic was signed on January 31st and Campbell didn’t actually join until March).
Secondly if it were the case that Everton were buying strikers in January who are better than shooting than those they already have it begs two questions: a) why aren’t they able to buy these players over the summer? and b) surely Everton can’t be the only team doing this, it must be a league wide thing? Well we a) don’t see a spike in Everton’s sh% at the start of the season and b) don’t see a spike in sh% across the league in either February or September (the end of the first transfer window). If we’re being picky the league sh% in September is actually 1 standard deviation below the full season mean.
| “Simplistic – but what about who is taking the shots (injuries to key players), where on pitch and when shot is taken..for example – winning a game and taking a ‘relaxed’ shot or desperately chasing a game and taking shot under pressure?”
Injuries may have an effect, there’s no way I can tease out the detail with the numbers I have. I’d suggest that over ten years that they should even out to a large extent.
The second half of this tweet is what I’d call score effects (I’ll once again point people towards this excellent article by Gabe Desjardins). Everton win more games between December and April and so it makes sense that they spend more time playing with the lead. Teams in the lead do have an improved sh% and sv%, and so some of the difference can be attributed to score effects. That being said Everton’s sv% only improves from 80 to 81% and I think we could realistically expect the sh% and sv% to be affected pretty equally, hence I’d argue that a large component of the sh% improvement remains unexplained.
If anyone wants to throw out any more suggestions I’ll happily look into them. That being said I honestly think that, as a group, we’re under-appreciating just how much luck is at play here, even over a 40 game sample.