Andre Villas-Boas - Head Coach

[BrainOfLevy]

Re: ***The Official AVB Discussion Thread***

No you're getting the wrong end of the stick.
The Sports Illustrated Cover Jinx was referenced earlier just because its a commonly known example of something regressing to a mean. Just like if you analyse the outcome of 180 managerial changes very non changes its just gonna average out to not making a difference.

What they are missing though is we're not trying to make 180 changes (at least i hope not), we're making one change.
We all know managers can make club's fortunes and we all know they can **** them up. Just because on average they make no difference doesn't there is no better manager out there for us than AVB.

This is again a misunderstanding of it, or just a firm refusal to acknowledge the truth behind it.

There is absolutely no difference, historically, in the success of a club sacking their manager vs one that doesn't. That is a statistical fact. Therefore there is likely no need to sack AVB, because results will likely improve anyway. To say that 'it proves clubs were equally right to sack their managers' is a misunderstanding of those findings.

Of course you can say there are better managers out there for us than AVB, but what are you basing it on? How much time are you willing to give them when things go tits up for a bit? Why would you be any more likely to expect better results from a different man when the evidence suggests results will likely improve with the current one anyway?

 

[braineclipse]

Re: ***The Official AVB Discussion Thread***

I'm going to PM NWND, this regression to mean thing is not big deal at all. a simple two way betting example over a large population will show how simple yet pointless that analogy can be at times.

Not a new scientific discovery and quite frankly its actually being used wrongly here

we're using THAT as some kind of evidence sacking a manager is not worthwhile or that keeping the manager is not worthwhile? Its a one sample test man

edit:- something that would be interesting though would be to get some actuarials on the case to find out the liklehood of success say over 5 years if we stay with AVB or Go with someone else (i guess you would have to specify an exact pool of options for that test to be fair)

It is relevant if the argument is being made that a new manager effect could help us out or if people are pointing to specific other clubs that succeeded after changing their manager.

A larger sample set of data will be more informative than either side just listing clubs that either failed or succeeded after a managerial change in turn.

 

[superfaisty]

Re: ***The Official AVB Discussion Thread***

Missed the game but united won 5-0 away yesterday- am worried big time now for Sunday.

AVB needs to find a way to keep Rooney quiet and st the same time get us playing attacking football and scoring goals.

Any thoughts what we think AVB will do? Can sandro do man marking job?!

 

[NWND]

Re: ***The Official AVB Discussion Thread***

This is again a misunderstanding of it, or just a firm refusal to acknowledge the truth behind it.

There is absolutely no difference, historically, in the success of a club sacking their manager vs one that doesn't. That is a statistical fact. Therefore there is likely no need to sack AVB, because results will likely improve anyway. To say that 'it proves clubs were equally right to sack their managers' is a misunderstanding of those findings.

Of course you can say there are better managers out there for us than AVB, but what are you basing it on? How much time are you willing to give them when things go tits up for a bit? Why would you be any more likely to expect better results from a different man when the evidence suggests results will likely improve with the current one anyway?

Just one last post, but I just don't believe in applying statistics to football, or at least of using them as the sole basis. Because statistics can be manipulated to show whatever anyone wants.

form may recover 'anyway', but really, statistics can't demonstrate this, because the managers who got sacked, never got to demonstrate that THEIR form would improve.

It is likely that the form of the other teams that didn't sack their managers, didn't dip to the level where the board felt like sacking them.

Also, different teams have different objectives and expectations. So while Norwich, for example, are happy to let Houghton go through horrendous spells of form, confident that they WILL recover enough to meet their objectives and stay up, a run of 4 games with 4 losses is a disaster for a title-chasing team, as it probably knocks them out of the race.

Our form just now is in danger of cutting us too far away from our rivals for the top 4, which is our objective, so while AVB may eventually recover our form, if this form continues for a few more games, its likely that we won't meet our objectives.

Therefore, he should face the consequences.

Regardless of whether form returns regardless of sacking, for some clubs, the fact that it dipped in the first place is unacceptable.

Also, stats don't tell you the true quality of form on the pitch. A team maybe playing well, creating chances, but not getting results, or a team may be scraping by like we have, eaking out lucky wins, but also losing games. It is always IMO more likely that form recovers in the first example, even though the stats would show that both teams performed at about the same level. The 2nd team, once the poor performances turn into bad results are more likely to sack their manager, because the board and fans cannot see any signs of recovery or progress.

It is likely that you could point to a number of managers who probably would not have been able to recover their teams form. Would you have kept Ramos on after his 2 points from 8 games, for example? Would you have taken that chance with the club rock bottom of the table and in danger of being cast adrift?

Anyway, its been (mostly) good having these debates, but I think you're all aware of my views on the matter now, and credit where its due, you're all still sticking to your own guns, whereever they may be pointed. So I think its time to call things a day.

 

[Rorschach]

Re: ***The Official AVB Discussion Thread***

Why is some other teams mean higher than our mean?

 

[braineclipse]

Re: ***The Official AVB Discussion Thread***

Has anyone claimed that statistics should be used as the sole basis for making decisions? Would surprise me...

 

[milo]

Re: ***The Official AVB Discussion Thread***

Why is some other teams mean higher than our mean?

They spend more money

 

[Rorschach]

Re: ***The Official AVB Discussion Thread***

They spend more money

I guess we are too mean.

 

[braineclipse]

Re: ***The Official AVB Discussion Thread***

Missed the game but united won 5-0 away yesterday- am worried big time now for Sunday.

AVB needs to find a way to keep Rooney quiet and st the same time get us playing attacking football and scoring goals.

Any thoughts what we think AVB will do? Can sandro do man marking job?!

I hope he will bring Vlad back into the side. United don't have the physical presence up front that City had so there's no real reason to play Dawson and Kaboul together anyway. Vlad looks a lot better suited than both when dealing with good technical players along the ground.

Otherwise he should set us up in a similar 4-3-3 to the one we played against City imo, perhaps Dembele for Paulinho or Holtby and Townsend on one of the flanks. Ade if fit and ready is of course an option too. Point is that we gave away goals to City with individual mistakes. I don't think a massive re-haul of the setup is needed to keep us tight at the back, the players just have to stop making silly individual errors and we should be solid enough.

Getting goals against a solid United team will probably prove more of an issue.

 

[Jurgen the German]

Re: ***The Official AVB Discussion Thread***

The vibe Im getting from most AVB supporters on this thread, seems to be implying that if he gets sacked and a new manager comes in and improves the football and the results, then AVB is STILL blameless, as he would have improved us eventually anyway and he simply wasnt given enough time.

Also it seems to be that they are writing off the new manager already, without knowing who that might be or what success he may bring us. Its as if they are saying that AVB is our only chance of success, even though he has yet to bring us any.

To predict doom and gloom if we sack him is beyond baffling. It will be doom and gloom if we dont.

Very similar to what people said about Ramos in my opinion. Can't put a number on it, but some posters felt he could have achieved top 4 with our squad if he had been given the time despite the results pointing to the contrary. He made some baffling decisions during his time here, like playing Ledley King in the UEFA Cup when we bottom of the league and his knee condition meant he couldn't play two games in a week. I'm not saying AVB is another Ramos just to be clear, he's clearly much better than that.

 

[milo]

Re: ***The Official AVB Discussion Thread***

The vibe Im getting from most AVB supporters on this thread, seems to be implying that if he gets sacked and a new manager comes in and improves the football and the results, then AVB is STILL blameless, as he would have improved us eventually anyway and he simply wasnt given enough time.

Also it seems to be that they are writing off the new manager already, without knowing who that might be or what success he may bring us. Its as if they are saying that AVB is our only chance of success, even though he has yet to bring us any.

To predict doom and gloom if we sack him is beyond baffling. It will be doom and gloom if we dont.

i'd be interested in seeing the posts that have lead you to this conclusion.

 

[stinkfoot]

Re: ***The Official AVB Discussion Thread***

I can't help but think of this...

[video=youtube;BKHoMi-U8g4]http://www.youtube.com/watch?v=BKHoMi-U8g4[/video]

 

[africanspurs]

Re: ***The Official AVB Discussion Thread***

The test has already been done, there is no difference between the success of the sacking club and the success of the club that keeps the manager.

You could then maybe do it over say 5 years, but in that time we are talking about so many other variables coming into play that could impact success of failure.

In this instance, where our form has only tailed off for a handful of games, it is absolutely the right study to be using, especially since people are calling for the manager to be sacked!

There could not be a better example/study to use.

"The test has already been done" Do you even know what the test was? Do you know any parameters except for the number of clubs that are involved? the definitions tested against? was the test to prove or to disprove?

its okay if you want to talk to me like i dont get it btw, ignoring my back ground in stats and decision modelling, i am curious to know how this model is a tested good model to use in this scenario

if you dont know or cant be bothered i'll accept a link

edit:- i've just realised that i dont even know what we are testing against :lol:

 

[africanspurs]

Re: ***The Official AVB Discussion Thread***

*****

 

[milo]

Re: ***The Official AVB Discussion Thread***

"The test has already been done" Do you even know what the test was? Do you know any parameters except for the number of clubs that are involved? the definitions tested against? was the test to prove or to disprove?

its okay if you want to talk to me like i dont get it btw, ignoring my back ground in stats and decision modelling, i am curious to know how this model is a tested good model to use in this scenario

if you dont know or cant be bothered i'll accept a link

Please can you try and be a little more polite, there is no need for this hostility.

 

[africanspurs]

Re: ***The Official AVB Discussion Thread***

*****

 

[scaramanga]

Re: ***The Official AVB Discussion Thread***

"The test has already been done" Do you even know what the test was? Do you know any parameters except for the number of clubs that are involved? the definitions tested against? was the test to prove or to disprove?

its okay if you want to talk to me like i dont get it btw, ignoring my back ground in stats and decision modelling, i am curious to know how this model is a tested good model to use in this scenario

if you dont know or cant be bothered i'll accept a link

edit:- i've just realised that i dont even know what we are testing against :lol:

I've already linked to the BBC report in this thread but:

Clicky <--- Brief report
Clicky <--- Discussed as a part of the BBC's "More or Less", where they debunk crappy stats and point you towards interesting ones (like this one)
Clicky <---- This is a summary piece but with a link to their own paper and a handful of references
Clicky <----- The paper itself
Clicky <---- Publication of said paper

I believe the standard reference in this field is Bruinshoofd, A, and B Ter Weel, B (2004), "Manager to go? Performance dips reconsidered with evidence from Dutch football", European Journal of Operational Research 148: 233-246

 

[braineclipse]

Re: ***The Official AVB Discussion Thread***

The number of variables to consider must be astronomical to be really informative to make definitive conclusion BoL is making

There is a mean for a reason , its an average which tends to be a central tendency anyway..and its actaully an average over a controlled group over a finite length of time.

Do you think outliers and anomalies are included in this? I wouldnt know, people are throwing in this theory which may or may not include 'lucky and unlucky' coaches..or the 'best and worse' worst coaches etc

another question i should be asking is when was that average actually done? was there an average , index/figure/whatever it is, before the experiment? or after? are we measuring that success or failure against that marker? how do you even determine whats success or failure anyway? is that results over a specified period of time? whats the length of time to test for success?

is the model of regression being used even standard/ Linear? or is it stochastic? the reason i ask cause the experiment type is important to understand what and how they are testing this stuff against.

i dont see how a model that has a generalised result can then be applied at the singular level...as a predictor it wouldnt be safe...if you wanted to predict that over the course of 360 combined sackings and retainers that results would have a small Stdev or null then fine

but at the singular level how do you use generalisation for that?

I'm not sure I agree with the definitive conclusion BoL is making.

I haven't looked at the study myself, but even accepting that there are some issues that makes it less than perfect (isn't there always?), I think it most likely has some validity in the situations I pointed out in the post you quoted. It seems to me to at least offer some data and results instead of the claims I pointed out as not particularly informative.

 

[africanspurs]

Re: ***The Official AVB Discussion Thread***

Please can you try and be a little more polite, there is no need for this hostility.

hostility?!

The last few pages i see people telling others that dont agree with the theory for this scenario that they cant grasp it or dont understand whats going on.

I happen to think you cant use that test to make a deicision on whether or not to kick a manager out in one particular instance...so pre-emptively , before people say "you just dont get it, you dont understand" i'm volunteering that people educate me from scratch. cause quite frankly its one thing having a masters degree in the stuff , its another actually applying it and i havent read the book before

actually you know what...i'll just delete them

 

[africanspurs]

Re: ***The Official AVB Discussion Thread***

I'm not sure I agree with the definitive conclusion BoL is making.

I haven't looked at the study myself, but even accepting that there are some issues that makes it less than perfect (isn't there always?), I think it most likely has some validity in the situations I pointed out in the post you quoted. It seems to me to at least offer some data and results instead of the claims I pointed out as not particularly informative.

i can agree with this.

i'll just leave it at that