tmesisSCIAGRoyalBlue (we didn't get the lift that you normally expect from a change of manager)
No such thing. https://www.fourfourtwo.com/features/so ... difference
That's a pretty terrible bit of analysis.
The thinking behind it is based on the fact that clubs average 1.3 points per game, but sack a manager when averaging about 1 point per game, yet three months later will be on about 1.3 a game again.
The assumption appears to be that the dip to 1 point per game is therefore a blip, and if clubs did nothing, they'd bounce back to 1.3 a game again.
Also, the new manager lift is generally seen as a short term thing, much less than three months stated there, when the players are trying harder to impress the new manager.
If anything, the stats there suggest that if clubs sack a manager when on 1 point per game, they'll have improved to 1.3 a game within three months.
Of course, from that you have no idea how well the clubs would have done if they'd not sacked a manager, but the assumption that they'd return to 1.3 a game again seem highly dubious.
Regression to the mean isn't an assumption, it's a well documented statistical phenomenon (some would say a statistical inevitability) across a range of fields including the form of football teams. Yes, it would be nice if this analysis actually contained a control group of clubs who go through a bad spell and don't sack their manager, but there are very few of those around these days - Bournemouth and Burnley, Morecambe, any more? Maybe you could also look at clubs who change manager in good form and whether things get better for them. But frankly the principle that bad runs of form eventually come to an end is uncontroversial, so this seems unnecessary.
Point being, there's no evidence to suggest that sacking the manager will improve results. Yes if your manager is actually holding you back (Mourinho at United) then it probably will, but that's about improving, not just changing. Situations like Coppell offering to resign when we were struggling in the PL because he thought a new manager would "get a bounce" are just stupid - no reason to think the old manager would be less likely to get a bounce if all else was equal.