Does the coach even make a difference?

[ballantyne]

This article from the Guardian...
https://www.theguardian.com/footbal...-clubs?CMP=Share_AndroidApp_Copy_to_clipboard
...reports on a large study of the Bundesliga, and concludes "depends".
Good and bad coaches make a difference to points won, but most don't, notwithstanding they may make difference in other areas such as youth development. This helps to explain why when you look back at highly-regarded, long-serving coaches, you often see the win/lose close to 50/50.
Very interesting though: former star players are bad or mediocre coaches. One theory is that FSPs walk into top flight jobs, while proper good coaches earn the top jobs by learning their trade in the lower leagues. Kosmina, Magilton, Farina, Okon, Butcher, Aloisi, are ones I can think of. If this is true, the implications for CCM are that we cannot hold on to a star coach, like we can't a star player, and that we must not employ Montgomery, Zwaanswijk or Hutchinson as coaches.

 

[Antlion]

I think this would apply to coaches within the bell curve - which is most of them.

We've seen first hand the outer ends of the bell curve and what those coaches do
eg. Graham Arnold v Mulvey/Walmsley

 

[ballantyne]

I think this would apply to coaches within the bell curve - which is most of them.

We've seen first hand the outer ends of the bell curve and what those coaches do
eg. Graham Arnold v Mulvey/Walmsley

It's a U-curve if you plot coach influence against points, but yeah, we've perhaps been down the wrong end. We won't know until Moss, Walmsley, Okon and Mulvey have each gone through c5 more appointments. This is why a large study like this is so valuable.

 

[nebakke]

This article from the Guardian...
https://www.theguardian.com/footbal...-clubs?CMP=Share_AndroidApp_Copy_to_clipboard
...reports on a large study of the Bundesliga, and concludes "depends".
Good and bad coaches make a difference to points won, but most don't, notwithstanding they may make difference in other areas such as youth development. This helps to explain why when you look back at highly-regarded, long-serving coaches, you often see the win/lose close to 50/50.
Very interesting though: former star players are bad or mediocre coaches. One theory is that FSPs walk into top flight jobs, while proper good coaches earn the top jobs by learning their trade in the lower leagues. Kosmina, Magilton, Farina, Okon, Butcher, Aloisi, are ones I can think of. If this is true, the implications for CCM are that we cannot hold on to a star coach, like we can't a star player, and that we must not employ Montgomery, Zwaanswijk or Hutchinson as coaches.

Tbf, Hutch at least has sort of been taking the apprenticeship route, arguably so has Monty and definitely Zwaanie - I'm not sure they're the sort of star-players that they're thinking off here.
Like him or not... Well not... But still... Muscat is another player who has moved on from being a star to the team, did an apprenticeship and hasn't been a complete failure

 

[Forum Phoenix]

I believe the impact and difference between a good and bad coach is night and day and have observed it many many times.

But there are so many variables beyond the metrics used in that paper that I think you can still only draw very limited conclusions.

High level athletes not necessarily having the skill set to transition to high level coaches should be anything but a surprise imo.

 

[ballantyne]

I believe the impact and difference between a good and bad coach is night and day and have observed it many many times.

I think what the paper is saying is that you haven't: that most of the time the coach has little to no influence on points won.

 

[Forum Phoenix]

Yeah, but conclusions are only as good as the data you feed in. And imo there are still many other factors than those they have used to be able to make such a statement, which makes that conclusion still not clear or verifiable.

I can give multiple examples if you’d like, but I’m in a rush right now.

 

[Antlion]

I think what the paper is saying is that you haven't: that most of the time the coach has little to no influence on points won.

Yes, but that's statistically over a large number of coaches at a similar elite level.

I don't think anyone would say they would as happily take Mike Mulvey over Graham Arnold given a 50/50 choice of next coach.

 

[Insertnamehere]

Essentially in theory GA would have 10 more points with this squad than Mulvey did over the same period.
Couldn't tell you if that's accurate. I personally think coaching in the HAL makes a massive difference. Mainly in man management and teaching skills perhaps more than tactics.

I don't think the analysis is very strong by in large. There's too many variables unaccounted for, for mine.
As at the end of most any Honours or Master's work, more research needed.

 

[Wombat]

I think what the paper is saying is that you haven't: that most of the time the coach has little to no influence on points won.

Clearly thats a load of tosh.

 

[ballantyne]

Clearly thats a load of tosh.

You say clearly, so presumably your list of citations is on the way?

 

[Ancient Mariner]

Shows trends, no significant results, we'll give it to the news.

 

[Wombat]

You say clearly, so presumably your list of citations is on the way?

Are you completely stupid??????

Are we not the perfect test case to prove how hollow this drivel is. If you are still in doubt let me know and I will waste my time for your education.

 

[Antlion]

It's a dataset from one place in one league across one period of time. Valid, but doesn't extrapolate to explain what's going on in other leagues in other periods of time

 

[bikinigirl]

. i thought maybe Charlesworth had commissioned the report

. now, even he has realised it is bullshit so he has had it published

 

[ballantyne]

Are you completely stupid??????

Are we not the perfect test case to prove how hollow this drivel is. If you are still in doubt let me know and I will waste my time for your education.

You say "test case", so your contention seems to be that a dataset where n=21 "clearly" invalidates a dataset where n>6000
A more reasonable conclusion would be the one is contained within the other, perhaps?

 

[Forum Phoenix]

 

[Wombat]

You say "test case", so your contention seems to be that a dataset where n=21 "clearly" invalidates a dataset where n>6000
A more reasonable conclusion would be the one is contained within the other, perhaps?

Exactly.

 

[SuperHans]

Have a read of "soccernomics" and the conclusion is that the manager generally makes little or no difference. The players are already motivated at pro level. Instead, success is determined by players' wages. Manager sackings usually occur when the team is doing poorly (below average) in a cycle of ups and downs and after the sacking the stats predict that the teams fortunes are on a return to normal (average) in the cycle anyway. The replacement manager rides that success, but isn't necessarily the cause of that bounce back.
Ranieri and Leicester are also used also to illustrate the lack of cause and effect of manager and success. He had a poor record in Italy. He was dumped by Greece after losing to the Faroe Is. He was appointed to Leicester and their success was actually due to a squad of good players, yet the media attributed it to him and he was hailed as a genius. He then left Leicester and won something like 5 out of 25 games. It concludes that Leicester was successful despite Ranieri, not because of him.
A manager sacking is a ritual sacrifice of a scapegoat, its purpose is to deflect attention away from the lack of investment on quality players by club owners. It really is worth reading the book, but here are a couple of relevant links.
https://www.fourfourtwo.com/features/soccernomics-does-sacking-manager-actually-make-difference
https://www.soccernomics-agency.com/?p=1001