Time Lane ranks highly
Tim Lane is considered by many to be an elite sports commentator. And he is not at all afraid of bringing some quantitative insights to his analysis of the great game of AFL. Indeed, some claim he is one of the inspiriation for Stan the Statistician. Back in July 2005, he took it upon himself to add some quantitative meat to the laments of Victorian fans that too many foreign clubs were finishing too high up the ladder. The rather remarkable article is HERE.
He decided to measure the success of m=6 foreign clubs by the total of their rankings. He was using the Wilcoxon rank sum statistic to measure the difference between two populations. And he successfully calculated the correct null mean from N=16 teams as 51 (without generalising to the formula m(N+1)/2). He calls this “par value”. I wonder whether this wouldn’t be a good informal term to use for the null mean in teaching statistics generally.
And while he does not calibrate deviations from 51 using the Wilcoxon distribution he does at least talk about the issue of robustness – pointing out that the results from 2002 to 2005 were skewed by the anomalous performances of the Lions but also pointing out that the main patterns remain when this outlier is removed.
Ian Gordon who is the director of the Statistical Consulting Center at the University of Melbourne sent Lane a letter pointing out that his rank sum statistic was a standard statistical tool and that he had got the par value correct. Tim reportedly cops a bit of ribbing from his fellow commentators about his over-intellectualisation of the game. Apparently he was quite chuffed by Ian’s letter and let his colleagues know that his calculations had been endorsed by a properly qualified boffin.
On a different tack, Lane goes on to discuss why foreign clubs might dominate. I was particular interested in the following quote from “a Victorian based coaching assistant” who observes that
West Coast and Brisbane seem better than any team at putting hard flesh on young bodies.
I wonder what he is talking about folks!? They don’t take drugs over in the West do they? Many thanks to Ian Gordon to alerting me to this article.
You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.
September 4th, 2007 at 12:49 pm
Another point which probably should be made clearer is that results are correlated between years.
PS Stan the Statistician is my hero!
September 4th, 2007 at 2:00 pm
Further to Paul Eckermann’s point, I heard some “statistics” used on ABC football coverage which claimed to support the dominant non-Vic theory in the following manner;
Probability of a team winning the premiership is 1/16.
6 non-Vic teams.
Probability of non-Vic team winning the premiership is 3/8.
Probability of non-Vic teams winning 6 consecutive premierships (which is the current state of play) just by chance is approx. 0.003.
Therefore it is safe to say interstate teams are privileged.
So sometimes the commentators get it (almost) right. But mostly they get it very wrong.
I dream of a day when football is illuminated by useful statistics and not useless and trivial data termed “statistics”.
September 4th, 2007 at 2:23 pm
I’m glad he removed the outliers at least. Absolutely critical first step in quashing any theory.
September 4th, 2007 at 3:05 pm
but fatalberton isnt the removal of outliers the statisticians way of getting the “right” answer
September 4th, 2007 at 3:29 pm
yes but 42.5 % of all statistics are made up on the spot….according to Guru Bob
September 4th, 2007 at 3:37 pm
I’m still in raptures re Fraser Gehrig’s one mark, no disposals match against the Western Bulldogs about a month ago.
I bet you a bun that has never happened before! (Or again)
September 4th, 2007 at 3:45 pm
What do you make of this article?
Study finds rock stars live fast, die young
Doors singer Jim Morrison died before his time (file photo). (Reuters: John Schults)
Rock stars - notorious for their “crash and burn” lifestyles - really are more likely than other people to die before reaching old age.
A study of more than 1,000 mainly British and North American artists, spanning the era from Elvis Presley to rapper Eminem, found they were two to three times more likely to suffer a premature death than the general population.
Between 1956 and 2005 there were 100 deaths among the 1,064 musicians examined by researchers at the Centre for Public Health at Liverpool John Moores University.
As well as Presley, the toll of those dying before their time included Doors singer Jim Morrison, guitar hero Jimi Hendrix, T Rex star Marc Bolan and Nirvana’s Kurt Cobain.
Substance abuse
More than a quarter of all the deaths were related to drugs or alcohol abuse, the study in the Journal of Epidemial Community Health said.
“The paper clearly describes a population of rock and pop stars who are at a disproportionate risk of alcohol and drug-related deaths,” said Mark Bellis, lead author of the study.
He said the study raised questions about the suitability of using rock stars for public health messages, such as anti-drug campaigns, when their own lifestyle was so dangerous.
“In the music industry, factors such as stress, changes from popularity to obscurity and exposure to environments where alcohol and drugs are easily available can all contribute to substance use, as well as other self-destructive behaviours,” the report said.
First five years
It found that musicians were most at risk in the first five years after achieving fame, with death rates more than three times higher than normal.
Hendrix, Bon Scott of AC/DC and punk rocker Sid Vicious all died within five years of hitting the big time.
UK vs US
Among British artists, the risk of dying remains high until around 25 years after their first success. Then they return to near normal life expectancy.
That bodes well for rock survivors like The Who’s 63-year-old Roger Daltrey, who famously first sang “I hope I die before I get old” in the song “My Generation” back in 1965.
But this trend was not found in the US, where ageing rockers remain almost twice as likely to suffer a premature demise, particularly from heart attack or stroke.
American stars Jerry Garcia of the Grateful Dead, Carl Wilson of the Beach Boys and Johnny Ramone of the Ramones all died in their 50s.
The report’s author suggests that the high death rate among older American musicians could be related to the continent’s greater appetite for reunion tours, exposing the artists for more years to an unhealthy “rock n roll” lifestyle.
It could also be due to the poor medical outlook for impoverished American ex-pop stars who have no health insurance, he said.
- Reuters
September 4th, 2007 at 3:49 pm
I think your article, Rushby, is an outlier on this Blog which should be removed.
September 5th, 2007 at 1:59 pm
Nah, I’ll leave it there Stan. Speaking of outliers and Rock Star death rates I’ve got two words for you: Keith Richards.
September 5th, 2007 at 5:52 pm
There is a particular age at which these rock stars choose to leave this mortal coil which may be statistically significant.
Morrison, Hendrix, Cobain and Joplin all died at 27. Jeff Buckley may also have been 27, and perhaps also Mama Cass.
There may be outliers, but they could be removed I’m sure.
September 5th, 2007 at 5:53 pm
“42.5 % of all statistics are made up on the spot.”
And the other 59.3% are innacurate.