Manager v luck etc

[Fox Talbot]

On the question of whether luck evens out over a season …..

I have no background in the theory of stats but intuitively I cannot see a reason why luck should even out over 46 instances - or even a thousand instances. The distribution of luck is not even.

Over the first 2065 Lottery draws 23 came out 282 times, 20 came out 215 times - a 24% difference over what you'd have to say is as sustained and controlled experiment in generating random numbers (1-49) as you can find. The average I reckon is 253 and ten balls were drawn under 240 times.

I'm also not a believer in 'the table doesn't lie' and thought we were a false third in 2017. We were also damn lucky in 2012 (Brighton away among others!) - we've more than paid for these times in the hard years since 2012.

[stealthpapes]

On the question of whether luck evens out over a season …..

I have no background in the theory of stats but intuitively I cannot see a reason why luck should even out over 46 instances - or even a thousand instances. The distribution of luck is not even.

Over the first 2065 Lottery draws 23 came out 282 times, 20 came out 215 times - a 24% difference over what you'd have to say is as sustained and controlled experiment in generating random numbers (1-49) as you can find. The average I reckon is 253 and ten balls were drawn under 240 times.

I'm also not a believer in 'the table doesn't lie' and thought we were a false third in 2017. We were also damn lucky in 2012 (Brighton away among others!) - we've more than paid for these times in the hard years since 2012.

Is that difference statistically significant?

[The Royal Forester]

The best bit of luck a manger can get is, if he can sack a manager and then appoint himself as manager.

[stealthpapes]

On the question of whether luck evens out over a season …..

I have no background in the theory of stats but intuitively I cannot see a reason why luck should even out over 46 instances - or even a thousand instances. The distribution of luck is not even.

Over the first 2065 Lottery draws 23 came out 282 times, 20 came out 215 times - a 24% difference over what you'd have to say is as sustained and controlled experiment in generating random numbers (1-49) as you can find. The average I reckon is 253 and ten balls were drawn under 240 times.

I'm also not a believer in 'the table doesn't lie' and thought we were a false third in 2017. We were also damn lucky in 2012 (Brighton away among others!) - we've more than paid for these times in the hard years since 2012.

No answer

Anyway, this is after 1240 draws, so a lot earlier, but the historgrams are the frequency a particular ball comes out and the red line is the fit.



You would expect a spread in numbers, due entirely to chance, and it should fit a bell curve, which it does.

I have no background in the theory of stats

quite.

[stealthpapes]

Well, luck of course plays a part, there’s no denying that. Goalkeeper errors, bad referee decisions, ricochets etc etc...

But I just cannot possibly see how it can correspond to explaining how well/poorly a team does over 46 games

If you can provide the data to reject that claim, then I’ll concede the point but from where I’m standing, I don’t see how it’s possible

Hi URZZZ

This topic is covered in detail in Chapter 1 of The Numbers Game (ISBN-10: 9780241963623) and there's a few academic papers on the topic. A bit busy this weekend but there's some nice examples in the chapter.

I also recalled this Economist article, Managers in football matter much less than most fans think.


image upload

Graph provides an interesting insight into it. Certainly adds a different perspective into it

Will give the chapter a read when I've got the time to

No worries.

The one bit that I think is relevant to the 'is 46 games enough' discussion is the analysis of 2006/07 in the Prem. While the worst team was very likely the worst team, the teams clumped around the drop were all very similar in quality. The difference between staying up and going down was, basically, chance.

I also like the study of intransitive triplets. The one bit you could probably explain most easily.

[Millsy]

The best bit of luck a manger can get is, if he can sack a manager and then appoint himself as manager.