How long should Ron get to turn things around?

That was, in fairness, not difficult to predict

 

"Regression to the mean" has absolutely zero application here.

I will defend the use of statistics from both those who misuse them and and those who misunderstand them.

 

If people understood regression to the mean and used it everywhere in their lives they would be significantly happier. There'd be a great deal fewer new golf clubs bought after a bad run of games, cars replaced after a run of faults and fewer people shouting at their kids when things go wrong in the mistaken belief that it's better than praising them when they do well.

Vin

 

I have to disagree: It's absolutely at the core of our current problems. We're in a bad run of games. It won't last forever as we'll regress to our mean performance. It's the explanation for my favourite graph, already posted above, but posted here again to avoid having to refer back.:

please log in to view this image

Manager sacked, results improve - regression to the mean
Manager not sacked, results improve - regression to the mean.

Vin

 

What? You suggesting I'm not the soothsayer I claim to be?

 

It could absolutely last forever as you have no idea what our "mean performance" level is.

 

Have you known any football team in the entire history of the world carry on getting a point from six games forever?

No. So you know that our mean performance level is > 1/18 points. QED.

Vin

 

But if your mean performance level is relegation standard no one cares if you return to the mean or not.

You can see the problem right here in this thread. Some people are saying our dip in form is not a big deal because we are a mid table side and we've had a lot of success in this year and years past.

You are saying that we are a good side and will regress to the mean because right now we are below.

You feel this is an aberration. Others are using the same "regression to the mean" concept to point at this being a normal and expected correction.

 

I thought you were arguing, let me remind myself, that

Whereas you're actually arguing that it doesn't matter if we regress to the mean if that's relegation performance. Sounds like we're on different wavelengths, so not really worth continuing.

Vin

 

My original response was to the idea that simply because Saints have been good, being in a bad stretch does not result in "regression to the mean."

As to your point about whether sacking managers is effective... you could also argue that the reason why teams improve both if they sack or staying with their manager is that teams generally make the right decision. It's not that if you don't sack your manager you will improve, and if you do sack your manager you still improve. It's that teams don't sack their manager when they think he can turn things around (and are right) while other teams DO sack their manager when they think they have lost the team (and are right).

So from the perspective of a fan or analyst who has no control over the situation and very few of the inside facts, that whatever the club decides to do generally works out. But from the club's situation they do have the facts and control over the situation, so they probably should not be relying on simple broad-based analysis.

It's like coming to an intersection and knowing that 50% of people turn left there and 50% of the people turn right, and 98% of them reach their destination without getting lost. That doesn't mean that it doesn't matter whether you turn left or right. It means that there is a decision being made, and most people know the route and hence make the correct choice.

 

To sack the manager, there would have to be something behind the scenes for which the board blame Koeman. A bad spell is just not a good enough reason to sack him, because he is good enough to turn things round if it is just a blip.

 

This started with you bemoaning people who don't understand statistics, specifically regression to the mean. This post proves that you simply don't understand it. And I can say that with a remarkably high probability of being correct, as what you've written is pretty much a PERFECT contradiction of the theory.

Vin

 

Courtesy of TheSackRace.com...

please log in to view this image

 

Regression to the mean is principle that in small sample sizes, there Will be greater variance. As the sample size gets larger, your variance grows smaller, thus you will regress to the mean. Therefore, any extreme result will tend to be followed by a less extreme result, because what are the odds of hitting a 1 in 100 chance twice in a row?

In real life, we do not have perfectly specified models where the only variation is due to chance. We have imperfect models where we ignore the things we cannot measure in the model. But we try to get enough samples so it evens out.

There are good managers and bad managers. There are good boards and bad boards. If you measure 10,000 clubs you will get them all and the result will be what happens to the average club when it sacks the manager.

The average club also finishes with a .500 record. But that doesn't say much. That does not mean you should expect Saints and every other club to finish .500 every year and Man U's success is due to sheer crazy luck.

What that chart shows you is that sacking a manager--on average--is not automatically a better move than not sacking a manager.

It does not mean that all managers are equal, or that every team that goes through a bad streak will automatically pull out of it or if they don't it is just bad luck.

In all likelihood, the teams that do not sack their manager are aware that perhaps they are going through a particularly tough part of the schedule or maybe their best two players are injured. So, they stand pat knowing it is not the manager's fault and they will recover and being competent, they are right more often than not.

And it is also likely that teams do fire their manager when they feel he has lost the club, and only when they feel there is a better replacement. And being competent they are right more often than not.

In both situations, you will see an improvement in performance. But it isn't just luck. It is skill.

 

That's fair enough really.

 

I'm actually a little surprised that more emphasis hasn't been put on the loss of Toby. He was quietly excellent again Saturday. I described him as "the best centre back I've seen at Saints in many years," last season and he played behind the best CM I've seen here since Steve Williams....

... It's little wonder there's been a blip.

 

Well, seems my throwaway line about regression to the mean has started a maths discussion! Just to add to this, it is worth stressing that the mean relates to the particular club. Villa are probably where they should be, given their squad. Chelsea not so. The regression is to the mean, given the circumstances. If the manager has lost the dressing room at a club then the board needs to act because the mean will be below the reasonably expected result. In the case of Chelsea, top 1-3 places. In the case of Saints, given all the circumstances I think mid-table is not wildly out of the expected mean, particularly when the league is so close. Those suggesting that Koeman has only a few matches to turn things around obviously have higher expectations than I do. Whether they are realistic remains to be seen.

 

Oh, and for those who can't be bothered to look it up, regression to the mean is simply the phenomenon where is you get a series of unexpected or extreme results (Saints being 2nd in the table last season for example) you would expect the next series to go the other way, to bring the overall average in line. It will be interesting to see what happens to Leicester. They really do look like the real deal so far, but a couple of injuries, which they could have had after two games, would tell us about their squad strength. Fascinating season.

 

Saints got second in the table at one point last season? Listening to the pundits and Wenger, you'd have thought it was West Ham

There's "regression to being mean" right there!

 

Strewth alive guys this is a bloody football Forum not a course on psychoanalysis phenomenon!!
It is an interesting fact that we all agree on and that is that Saints are playing crap football at the moment. There has been a lot of interesting suggestions as to how we can correct the situation.
Most suggest it would seem that RK should be given time to correct this and in the main it seems the board will support him.
Over the years we have gone down this route so many times before.........It seems we just haven't learned the lessons of the past by not keeping our best players and possibly our better managers so that we can at least hold our heads up and give teams a decent game of football.
In football there should be no guarantee of a result, if there were the game would be so bloody boring.
On this Forum and others, it gives us ordinary folk the chance to air our views and opinions. In doing so, our football board will get the feel of the mood of the ordinary fan. (no I am not suggesting they read not606) They do keep an eye on some forums, although it is questionable as to how much they are influenced by them. They of course will realise that without our support they are not going to get very far.
I think we will all agree we do not want them to overreact and the same can be said about our manager. We as fans need to let the board know we want them to back and support RK even further. Even if it means that our owner needs to wait for her money a little bit longer!!!!!!!