Posts from May 2018 (28)

Team Ratings for Round 16

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 94 matches played, 66 of which were correctly predicted, a success rate of 70.2%.
Here are the predictions for last week’s games.

Predictions for Round 16

Here are the predictions for Round 16. 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 13

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 96 matches played, 57 of which were correctly predicted, a success rate of 59.4%.
Here are the predictions for last week’s games.

Predictions for Round 13

Here are the predictions for Round 13. 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.

• “It is important that guidelines for mitigation measures are proportionate to the risk posed, and that remediation strategies should be informed by a risk-based approach.” That’s not the money quote from Sir Peter Gluckman’s report on methamphetamine testing for houses (PDF), but it’s the generic StatsChat message.  For more, see Russell Brown’s story; he has been a consistent journalistic voice against panic-based testing.
• From an NY Times oped “Knowing a person’s political leanings should not affect your assessment of how good a doctor she is — or whether she is likely to be a good accountant or a talented architect. But in practice, does it?  Recently we conducted an experiment to answer that question”.  As Andrew Gelman explains, they totally didn’t.
• A new UK Parliament report “Algorithms in Decision Making”
• “Why Government needs sustainable [statistical] software, too”

Q: Did you see using a mobile phone after 10pm leads to depression and loneliness, depression, bipolar disorder and neuroticism?

A: Where?

Q: The Independent, the Daily Mail, news.com.au, the Geelong Advertiser, Student Problems, …

A: So, what are we supposed to check first?

Q: Mice. It isn’t mice, it’s people.  “However, now a study of more than 91,000 people has found that scrolling through your Instagram and Twitter feeds from the comfort of your pillow in the wee hours could increase the likelihood of developing a number of psychological problems such as depression, bipolar disorder and neuroticism.”

A: Ok, ok.  Do any of them link?

Q: The Independent does. But it’s behind a paywall

A: <sighs> Ok. Here’s the press release. 

Q: But that doesn’t mention mobile phones. Or Twitter or Instagram.

A: No, it doesn’t.

Q: It looks like they used fitbits, though

A: Yes, or near offer.

Q: Could they tell from those when someone was using their phone?

A: I don’t know if they could, but they didn’t. They just looked at how much people’s physical activity differed between night and day.

Q: What’s that got to do with mobile phones?

A: If using your phone late at night stops you sleeping, then you might have less difference in activity between night and day.

Q: I suppose. Did they look at actual sleep?

A: Not in this study

Q: So, the people with less day-night difference in physical activity ended up with more mental health problems?

A: No, they started off with more mental health problems.  As the comment at the bottom of the press release says “The study population (median age at baseline of 62 years, IQR 54-68 years) is not ideal to examine the causes of mental health, given that 75% of disorders start before the age of 24 years.”

Q: These were 60-year olds?

A: Yes. In 2013-14.

Q: Did British 60-year-olds even use Twitter and Instagram in 2013-14? Instagram was only on iPhones back then, wasn’t it?

A: About a third of Brits between 55-64 had a smartphone then, and about 10% of older people.

No.

This isn’t even supposed to be new research: it’s just a new set of guidelines based on all the same existing research. Since it’s a new set of public guidelines, you’d think a link would be appropriate: here it is.

The story says “”No level of intake” of processed meats will reduce cancer risks.”  and the quote from the report is The data show that no level of intake can confidently be associated with a lack of risk.  I don’t think that will surprise many people, and it’s what we’ve been told for a long time. There isn’t a magic threshold where bacon switches from being a health food to being bad for you. If you want something more quantitative, the figures we had last bacon panic haven’t changed: eating an extra serving of bacon every day is estimated to increase your lifetime bowel cancer risk by a factor of 1.2, or a bit under two extra cases per hundred people.

For alcohol, the focus on cancer is a bit misleading.  Low levels of alcohol consumption increase cancer risk but reduce heart disease risk, and there’s a range where it’s pretty much a wash — there isn’t a ‘safe level’ from a cancer viewpoint, but there probably is from a not-dying viewpoint. Still, there are lots of people who’d be healthier if they drank less alcohol — and that’s probably not the first time they’ve heard the message.

From Toby Manhire at The Spinoff, quoting Reuters

“New Zealand’s dairy-fuelled economy has for several years been the envy of the rich world, yet despite the rise in prosperity tens of thousands of residents are sleeping in cars, shop entrances and alleyways.”

There was something similar in the Guardian, too. As Toby says

The juxtaposition is compelling and well made. The number is compelling and nonsense.

I’ve posted about this issue before. The OECD report that people use (directly or indirectly) as source, says

Australia, the Czech Republic and New Zealand report a relatively large incidence of homelessness, and this is partly explained by the fact that these countries adopt a broad definition of homelessness…..In New Zealand homelessness is defined as “living situations where people with no other options to acquire safe and secure housing: are without shelter, in temporary accommodation, sharing accommodation with a household or living in uninhabitable housing.

That’s much broader than ‘sleeping in cars, shop entrances and alleyways.” One of the researchers behind the NZ figure said, in a Herald interview in 2016

“If the homeless population were a hundred people, 70 are staying with extended family or friends in severely crowded houses, 20 are in a motel, boarding house or camping ground, and 10 are living on the street, in cars, or in other improvised dwellings.”

Homelessness is a real problem in New Zealand. Because it’s a real problem, it’s important to focus on what the problem actually is, and not to make up a different problem.

Reuters has corrected the figure but hasn’t otherwise changed the story.  The fact that reducing the figure by a factor of ten doesn’t otherwise change the story might tell you something about the story.

(update: ok, now I’ve actually read all of Toby’s post, not just the first few paragraphs, he basically says all this already)

From the Herald (via @aw_nz on Twitter)

One of the features of pie charts is that it’s relatively hard to judge angles and compare segments. Still, if you get them wrong enough, people can tell.   For example, the taxes — the grey and orange wedges — are clearly more than half the circle, but the numbers add to only 43%.  Less dramatically, the 13% wedge for GST is larger than the 18% wedge for importer margin, and the 30% wedge for fuel excise is larger than the 35% wedge for refined fuel.  You don’t have to be very cynical to wonder whether it’s a coincidence that the tax components are being exaggerated. [update: you don’t, but you’d probably be wrong — see comments]

Here’s an accurate piechart, assuming the numbers are correct:

Team Ratings for Round 15

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 87 matches played, 60 of which were correctly predicted, a success rate of 69%.
Here are the predictions for last week’s games.

Predictions for Round 15

Here are the predictions for Round 15. 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 12

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 88 matches played, 49 of which were correctly predicted, a success rate of 55.7%.
Here are the predictions for last week’s games.

Predictions for Round 12

Here are the predictions for Round 12. 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 the Final

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 134 matches played, 95 of which were correctly predicted, a success rate of 70.9%.
Here are the predictions for last week’s games.

Predictions for the Final

Here are the predictions for the Final. 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.