Posts filed under General (3152)

Team Ratings for Round 4

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 24 matches played, 15 of which were correctly predicted, a success rate of 62.5%.
Here are the predictions for last week’s games.

Predictions for Round 4

Here are the predictions for Round 4. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

Team Ratings for Round 3

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 16 matches played, 12 of which were correctly predicted, a success rate of 75%.
Here are the predictions for last week’s games.

Predictions for Round 3

Here are the predictions for Round 3. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

• How many people are killed by police in the USA? Comes with a description of the statistical technique of mark-recapture estimation, from Granta magazine.

• Another example of cheating being exposed by surveillance and data collection — this time in contract bridge. From the New Yorker.

• If you want views on the `Kiwimeter’ survey, I suggest Tze Ming Mok’s piece at Public Address.

• The Ombudsman has released guidelines on Official Information Act requests through social media (PDF). Summary: it’s still a question, so it still gets answered.

• From NiemanLabs, how some news publishers are doing interactive graphics for mobile devices

And from XKCD: how much of various fluids does the US consume, using pipeline diameters to illustrate

(Update: Yes, I realise this is the sort of bubble plot we usually say mean things about. Not the point, here).

Usually when there are complaints about cancer screening in New Zealand it’s people complaining there isn’t enough.  The Herald has an interesting example of the opposite, under the headline, “Cervical test switch ‘wrong’“.  The ‘switch’ is from looking for actual abnormal cells on screening to just looking for high-risk strains of the virus HPV, which are responsible for nearly all cervical cancer.

Everyone agrees that viral testing is important, and that, all things being equal it is more sensitive.  The specialists who are complaining say that the initial screen should look for abnormal cells as well, and only proceed further if they are found. The problem with screening based just on the virus is that it will lead to a bigger increase in repeat screening, biopsy, and treatment, with added inconvenience, risk and expense.

Also, as they say in the NZ Medical Journal

The detection of a sexually transmitted infection rather than a significant cytological abnormality is a major change in the aim of screening. This may reduce screening participation. Any reduction in screening coverage will reduce protection from cervical cancer.

I’m not an expert on cervical screening, so I don’t know the answers, but the issues being raised are the right sort of questions to ask about a change in a successful population screening program.

Some headlines:

Herald: “Dead people on Facebook could outnumber the living”

Stuff: “There will be more dead people than living on Facebook”

But not any time soon:

Daily Mail: “Facebook will become the world’s biggest virtual graveyard with more profiles of dead people than living users by the end of the century, say experts”

So who are these experts? They are a statistician, Hachem Saddiki. The original idea came from Fusion, where Kristen V. Brown raised the question of when declining growth and lack of automatic deletion after death would lead to a majority of dead accounts, and looked for someone to work it out.

The Fusion story talks about some of the uncertainties — how fast will Facebook grow, what will happen to death rates across the world.  It doesn’t consider death rates of companies and technologies, though.

A 2012 report said

According to the report, the 61-year tenure for the average firm in 1958 narrowed to 25 years in 1980—to 18 years in 2011. At the current churn rate, 75% of the S&P 500 will be replaced by 2027.

You might expect Facebook to last longer than average, but there must be some chance it doesn’t make it to four times the average.

Much more importantly, will Facebook survive the technology transitions in anything like its current form — the equivalent of moving from player piano to Spotify in the music world?

Mark Twain noted in Life on the Mississippi

“In the space of one hundred and seventy six years the Lower Mississippi has shortened itself two hundred and forty-two miles. That is an average of a trifle over a mile and a third per year.”

His conclusion might well apply to the Facebook prediction: “One gets such wholesale returns of conjecture out of such a trifling investment of fact.”

• What Every Journalist Should Know About Science,  from the Neiman Foundation at Harvard — and they mean ‘know about how science works’ not ‘know about squid or neutrinos’ (via Harkanwal Singh)

• “Uber officials suggested that if an email address or rider/driver last name contains the word “rape” like “Jason Rape” or “Don Draper” it will be included when queried. … misspellings of the word “rate” and expressions like “you raped my wallet” accounted for false positives in the search results seen in the obtained screenshots.” So maybe we have a return of the ‘Scunthorpe’ problem, but maybe that’s as bogus an explanation as it sounds. From Buzzfeed.

• In Stuff’s list of the 17 most dangerous foods, the entry on the Jamaican vegetable ‘ackee’ says “In 2011 there were 35 poisoning cases” and also “1 in 1000 people develop ackee fruit poisoning each year in the Caribbean”.  This fails simple arithmetic: there are more than 35000 people in the Caribbean. A 1991 analysis by the CDC found a rate of 1 in 100,000 restricted to Jamaica, which seems much more plausible.  The list is also wrong about absinthe, and is unusual in considering rhubarb leaves to be a food.

• A visualisation of the ages people get married (in the US), from Flowing Data

• “If Bernie Sanders were to defeat Hillary Clinton in Michigan’s Democratic primary, it would be “among the greatest polling errors in primary history,”” He did. It was. Fivethirtyeight.com tries to explain how.

Auckland commuters know the frustration well: You’re waiting for a bus, and the electronic board shows it’s three minutes away. Ten minutes later, you’re still standing there, wondering what’s going on. Or the board may say a bus is 10 minutes away – but it suddenly turns up when you’re not paying attention.

Department of Statistics doctoral student Tom Elliott, pictured, also knows that irritation well – he was once a regular bus user – and his PhD research aims to come up with a model that leads to better predictions. Read the full story here.

Team Ratings for Round 3

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 17 matches played, 10 of which were correctly predicted, a success rate of 58.8%.
Here are the predictions for last week’s games.

Predictions for Round 3

Here are the predictions for Round 3. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

Team Ratings for Round 2

The basic method is described on my Department home page.

Here are the team ratings prior to this week’s games, along with the ratings at the start of the season.

Performance So Far

So far there have been 8 matches played, 6 of which were correctly predicted, a success rate of 75%.
Here are the predictions for last week’s games.

Predictions for Round 2

Here are the predictions for Round 2. The prediction is my estimated expected points difference with a positive margin being a win to the home team, and a negative margin a win to the away team.

My first Listener column, a couple of years ago was on the US gender pay gap and what it does and doesn’t mean. It’s now open-access.