Scientifically Judging a New Manager

[avonspurs]

Don't get me wrong - I love statistics - but I can see a few flaws in this:

1. Doesn't take into account the difficulty of the games being faced
2. Doesn't take into account the starting period for that manager (ie: did they have a full pre-season or none; was pre-season disrupted by major tournaments (ie: world cup/Euro), player turnover during the pre-season, etc.).
3. Does it take into account European involvement and distances travelled to these games - which affects time spent on training ground/recovery time for players
4. Weather conditions - more mistakes are bound to occur in poor weather conditions (unless you are Spurs last season, in which mistakes occurred daily!!!)
5. Player injuries.


And I'm sure many other that I cant think of standing in an empty cold flat!!!

 

[nightgoat]

Quite possibly, but like I say, Tim didn't have the benefit of playing scrubs in the Europa League or early cup matches against non-PL opposition.

I reckon a PL loss percentage of 27% is amongst the lowest of our PL managers. A 36% loss ratio, do you reckon many have beaten that?

I initially thought Ramos might manage that, but he only had an overall 32% loss ratio. Gross had 38%.

 

[mattstev2000]

I think we need to focus on league only, it is our bread and butter and allows for better comparability. Sherwood, as an example, didn't have the luxury of coaching us through the piss weak early rounds of the Europa League and both of his cup games were against Premier league opponents, not Forest and Brighton.

I think we should also exclude any league matches against teams that beat us when Sherwood was in charge. That will make his figures look even better.

Ideally he can have some extra percentage points for 'gillet wearing', 'uses of the words heart, guts or fight to describe a football match' and 'making a prat of himself in the press'. Then he would without doubt be statistically the best manager we've ever had.

 

[parklane1]

I think we should also exclude any league matches against teams that beat us when Sherwood was in charge. That will make his figures look even better.

Ideally he can have some extra percentage points for 'gillet wearing', 'uses of the words heart, guts or fight to describe a football match' and 'making a prat of himself in the press'. Then he would without doubt be statistically the best manager we've ever had.

Well done and thought out post Tim Were would we be without the stats lovers.

 

[Kendall]

I think we should also exclude any league matches against teams that beat us when Sherwood was in charge. That will make his figures look even better.

Ideally he can have some extra percentage points for 'gillet wearing', 'uses of the words heart, guts or fight to describe a football match' and 'making a prat of himself in the press'. Then he would without doubt be statistically the best manager we've ever had.

Groan. Hardly the same thing.

With the league you're more in line with comparing apples with apples. With cup games you can be comparing apples with piss covered shit.

 

[taricco]

I think we should also exclude any league matches against teams that beat us when Sherwood was in charge. That will make his figures look even better.

Ideally he can have some extra percentage points for 'gillet wearing', 'uses of the words heart, guts or fight to describe a football match' and 'making a prat of himself in the press'. Then he would without doubt be statistically the best manager we've ever had.

To be fair, I did mention in the OP that it should be league only. And I definitely don't have Sherwood bias. However, it does make the basis for a far fairer comparison (obviously).

 

[taricco]

Don't get me wrong - I love statistics - but I can see a few flaws in this:

1. Doesn't take into account the difficulty of the games being faced
2. Doesn't take into account the starting period for that manager (ie: did they have a full pre-season or none; was pre-season disrupted by major tournaments (ie: world cup/Euro), player turnover during the pre-season, etc.).
3. Does it take into account European involvement and distances travelled to these games - which affects time spent on training ground/recovery time for players
4. Weather conditions - more mistakes are bound to occur in poor weather conditions (unless you are Spurs last season, in which mistakes occurred daily!!!)
5. Player injuries.


And I'm sure many other that I cant think of standing in an empty cold flat!!!

Some good points. I think overall, however, it should be looked at as a sound structure for having sensible discussion over how long to give a manager. Obviously if they have an injury crisis against Chelsea away, on a waterlogged pitch, one day after a European game, we can put this into perspective during the discussion.

However, over the length of six-eight months these things tend to even themselves out anyway.

 

[Kendall]

To be fair, I did mention in the OP that it should be league only. And I definitely don't have Sherwood bias. However, it does make the basis for a far fairer comparison (obviously).

It's clearly the case. How can you compare anyone pre-Jol to those who get a bunch of nobodies to play against at the start of each season in the EL? Skew anyone?

 

[taricco]

Isn't points per game a better way to judge as it takes into account wins and losses? Ultimately the point of a league is to amass the most points.

No idea how our last 4 managers compare though.

I agree with this. Might make for a more interesting comparison.

 

[taricco]

Based on this theory @taricco can you post a scatter plot with time series for the last few managers? Would be interesting to see how well this theory works for each of the managers and whether the 2 season 'cut off' is the right benchmark to use in terms of time / success.

Woah! Some of us have jobs here. I can only spend 4-5 hours of the working day on SC!

 

[TheGreenLily]

According to Wikipedia (the mother of all factually accurate stats), it's 54.17%. I think this includes cup games though, as it seems very high! In the league, he's won 6/15, so around 40%.

Well, this is where it all becomes bullshit.

Some people like to complain that cup games are include, but when those cup games are included those same people than claim that they are against pub teams... Then when you don't include the pub teams. those same people claim that you are manipulating figures....

So I say it is all a load of bollox and waste of time debating it... You might as well just piss into the wind, at least you would have learned something.

 

[parklane1]

Well, this is where it all becomes bullshit.

Some people like to complain that cup games are include, but when those cup games are included those same people than claim that they are against pub teams... Then when you don't include the pub teams. those same people claim that you are manipulating figures....

So I say it is all a load of bollox and waste of time debating it... You might as well just piss into the wind, at least you would have learned something.

That stats can be used to push whatever opinion the user is trying to push, they very rarely tell the same story.

 

[taricco]

Well, this is where it all becomes bullshit.

Some people like to complain that cup games are include, but when those cup games are included those same people than claim that they are against pub teams... Then when you don't include the pub teams. those same people claim that you are manipulating figures....

So I say it is all a load of bollox and waste of time debating it... You might as well just piss into the wind, at least you would have learned something.

How unhelpful and ineloquently put! For the purposes of this particular thread, I've decided against including cup games. The reason being that it can skew the results in a not so conducive way. E.g. a manager in the Europa League will probably have better stats than someone who is not due to a number of matches against quite easy teams. However, a manager in the Champions League may face much more difficult teams and thus get a lower win percentage, despite being in a better competition.

It makes much more sense to compare like for like, which is why league only works well.

 

[SlickMongoose]

Isn't points per game a better way to judge as it takes into account wins and losses? Ultimately the point of a league is to amass the most points..

Yes! Yes!

For the love of god, what's the fucking point in "win %" or "loss %"? Where did these fucking moronic stats come from? Which arse on TV thought it would be a good idea? For fucks sake. Win % pretends that it doesn't matter whether you lose or draw, and loss percentage is twice as bad - It pretends that it doesn't matter whether you draw or win!

Points per game is what matters, the stat that best reflects how a team has done.

And for what it's worth, I would take out cup games too, because of the difference in opposition ability. Comparing just the league, if a manager has been there more than half a season, you can get a fairly decent idea. If you compare a manager who has managed Europa League qualifyers and group games to a manager who hasn't, of course that's going to disrupt the stats.

 

[mattstev2000]

Groan. Hardly the same thing.

With the league you're more in line with comparing apples with apples. With cup games you can be comparing apples with piss covered shit.

Ok, so just to confirm, Sherwoods good results against the might of Fulham, West Brom, Cardiff and Stoke are the apples that are worth including in a statistical assessment of a managers performance but AVB's good results against Basel, Inter Milan, Panathanaikos and Lyon are the 'piss covered shit' that should be discounted? Ok. No creative accounting going on here.

 

[Kendall]

Ok, so just to confirm, Sherwoods good results against the might of Fulham, West Brom, Cardiff and Stoke are the apples that are worth including in a statistical assessment of a managers performance but AVB's good results against Basel, Inter Milan, Panathanaikos and Lyon are the 'piss covered shit' that should be discounted? Ok. No creative accounting going on here.

You're really missing the point. We can't compare the cup teams because they differ so greatly. We can better compare the Premier League standard.

Creative accounting? Do you understand that one of the fundamental characteristics of the IASB accounting framework is comparability?

 

[mattstev2000]

You're really missing the point. We can't compare the cup teams because they differ so greatly. We can better compare the Premier League standard.

Creative accounting? Do you understand that one of the fundamental characteristics of the IASB accounting framework is comparability?

Of course I understand the fundamental characteristics of the IASB accounting framework, who doesn't?

Personally I think a large percentage of the games in the Europa league are against teams that would comfortably compete in the premier league and hence could be considered apples. Including those apples just increases your sample size when attempting to judge a managers performance statistically and hence is perfectly valid.

Alternatively, If you really want to toe the consistency line then really you have to only include the subset of premier league teams that all managers played games against. So a lot of the relegation fodder, like Cardiff, would be excluded because not all managers being compared played games against Cardiff every season.

You could take it even further, if it's consistency you're worried about and you want a view solely of how the manager is performing then really you have to compare games against only the same opposition players and with only the same squad of players available to chose from. If a player was missing through injury for a game for manager 1 then you have to exclude manager 2's game against the same opposition unless the same player was also unavailable through injury.

Basically, what I'm saying is that picking a consistent point to compare with is somewhat arbitrary and therefore I've decided to pick the point that makes Sherwood look more like the doofus he was

 

[striebs]

Is this a roundabout way of saying that Tim was better than Poch ?

 

[eddiebailey]

Loss % should be considered too, as I alluded to in my previous post. Winning loads and loosing loads isn't a recipe for success.

WIth three points for a win and one for a draw, win percentage tells you more than loss percentage. PPG though is the real measure.

 

[CowInAComa]

Sherwood?

Who gives a fuck. Seriously. I took down his poster in the summer.