Stats Chat

Team Ratings for the Qualifying Finals

This year the predictions have been slightly changed with the help of a student, Joshua Dale. The home ground advantage now is different when both teams are from the same country to when the teams are from different countries. 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 120 matches played, 83 of which were correctly predicted, a success rate of 69.2%.

Here are the predictions for last week’s games.

Predictions for the Qualifying Finals

Here are the predictions forthe Qualifying Finals. 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.

According to Stuff, Paula Bennett says the number on benefits is down by 10,000 since last year . Whether this is good or bad depends on where they ended up instead (as the story points out) but it is what the Government was attempting.

What the story doesn’t point out is that by MSD numbers there were drops of 10,000 or more in the number on benefits in the years ending June 2004, 2005, 2006, 2007 (and over 7000 in the year ending 2012, and over 8000 in the year ending 2003). And that’s as early as the data file goes.

Numbers on benefits have been going down for a long time, with an interruption for the recession, when they (obviously) went up a lot. Some of what we’re seeing now is just economic recovery, some is new rules, some is long-term changes in society. It’s hard to split the credit or blame, but it’s useful to know that a fall of 10,000 in a year isn’t  unusual.

From XKCD and Bayes’ Theorem

• A data scientist is skeptical of the ‘Quantified Self’ movement
• Where data comes from: a shop that is tracking customer cellphones (and admitted it)
• Adversarial Stylometry: you can be identified by your writing style. Can you hide it? (PDF)
• Luis Apiolaza on whether corn crop yields in Europe and US refute a benefit of GMO corn (with graphs and data and stuff)
• The true cause of death, and why it’s hard to define and reconstruct

Nathan Yau, at Flowing Data, shows this barchart-like object

The most obvious problem is that the bar for Chinese is about 3 times the area of the bar for English, and should be either about a third bigger, or, if you really believe there are 1 trillion Chinese speakers, 1300 times bigger. The bars for English and Spanish are obviously out of proportion as well.

I tried to find the original source of this graph, which Nathan doesn’t give, perhaps because he couldn’t find it either.  However, Google image search did show a related graph that explains some of the problems.  I found it at a range of sites, but the one at 10-most.com is where it looks most original.

This version is still ugly, but the bars match the labels.

Comparing the numbers to Wikipedia, it looks as though someone edited the labels from ‘top ten by number of native speakers’ to ‘top ten by total number of speakers’ either as a deliberate change or because they didn’t understand the original graph (and they also inflated the Chinese count by a factor of 1000). Sadly, the editor didn’t appreciate that the bars are more than just decoration; they are supposed to convey numeric information.

PS: Yes, we could also get picky about ‘Chinese’ as a single language, but they clearly mean Mandarin Chinese, the group of dialects that includes the official language of both the PRC and the ROC.

We haven’t had a good denominator post in a while, but I was struck by Stuff’s story on ‘social lending’.  By next April, an NZ company hopes to be able to start up peer-to-peer lending here, after changes in the law, and we’re told

“This will be a way for individuals to play alongside the big institutions,” Milsom said

Just how that can happen – and the scale that is possible – is shown from what has happened in the UK, where the largest peer-to-peer lender has lent more than £330 million since it launched in 2005.

That’s an average of just over £40 million per year over 2005-2012, or about 66p per capita per year in the UK.  On the same scale in New Zealand, that would come to perhaps $6 million per year in loans.  The UK lender in question appears to be Zopa, and they make their money by charging a 1% borrowing fee and 1% lending fee.  Under the same setup, $6 million per year in loans would mean $120 000 per year in income for an NZ equivalent, before paying any costs.

It’s certainty possible that peer-to-peer lending will take off, and it might not be long before it exceeds 1% of all unsecured consumer debt. But the idea makes more sense spun as ‘room for expansion’ not as ‘look at the scale in the UK’.

Each week, we would like to invite readers of Stats Chat to submit nominations for our Stat of the Week competition and be in with the chance to win an iTunes voucher.

Here’s how it works:

• Anyone may add a comment on this post to nominate their Stat of the Week candidate before midday Friday July 19 2013.
• Statistics can be bad, exemplary or fascinating.
• The statistic must be in the NZ media during the period of July 13 – 19 2013 inclusive.
• Quote the statistic, when and where it was published and tell us why it should be our Stat of the Week.

Next Monday at midday we’ll announce the winner of this week’s Stat of the Week competition, and start a new one.

(more…)

The Bechdel Test classifies movies according to whether they have two female characters, who at some point talk to each other, about something other than a man.

It’s not that all movies should pass the test — for example, a movie with a tight first-person viewpoint is unlikely to pass the test if the viewpoint character is male, and no-one’s saying such movies should not exist.  The point of the test is that surprisingly few movies pass it.

At Ten Chocolate Sundaes there’s an interesting statistical analysis of movies over time and by genre, looking at the proportion that pass the test.  The proportion seems to have gone down over time, though it’s been pretty stable in recent years.

Irene Ros has a visualisation of ridership on the Boston subway (famous from the Kingston Trio song back before I was born)

You can see how the rush-hour varies across the subway lines, with the Silver Line (to the airport) having only an evening rush hour, the Green line having a moderate morning peak,  the Red and Orange lines having about the same peak in the morning and evening, and the Blue line having more of a peak in the morning.

In what may be the least accurate risk estimate ever published in a major newspaper, the Daily Mail said last week

• Hormone replacement could cause meningioma in menopausal women
• Those using HRT for a decade have a 70% chance of developing a tumour
• Most are benign but 15% are malignant and all have damaging side effects

You don’t actually need to look up any statistics to know this is wrong, just ask yourself how many women you know who had brain surgery. Hormone replacement therapy was pretty common (until it was shown not to prevent heart disease), so if 70% of women who used it for a decade ended up with meningioma, you’d know, at a minimum, several women who had brain surgery for cancer.  Do you?

In fact, according to the British NHS, the lifetime risk of meningioma is about 0.07%. Since it’s more common in women, that might be as much as 0.1% lifetime risk for women. The research quoted by the Mail actually found a relative risk of 1.7, so the lifetime risk might be up to 0.17% in women who take a decade of hormone replacement therapy. That is, the story overestimates the risk by 69.8 percentage points, or a factor of more than 400.

While this may be a record so far, there’s still room for improvement, and I certainly wouldn’t bet on the record standing for ever.

(via @hildabast and @BoraZ on Twitter, and Paul Raeburn of the MIT science journalism program)