Posts from May 2018 (28)

• David Fisher in the Herald “Garrett appeared unaware both Edgeler and Lumley had since withdrawn their comments, having found they were based on bad data from the Ministry of Justice.“
• “Why Is It So Hard to Figure Out When the Bus Is Coming?” From Citylab. At least Auckland Transport doesn’t seem to have the problem of inaccurate maps that the story focuses on.  Prediction is still hard, especially when it comes to the future.
• There was a story in the more-excitable media about alien influences on Earth species. Some species of lizards in New Guinea have toxic, bright-green blood, and they aren’t even each others closest relatives. That wasn’t the alien story, though. The alien story was about octopuses. Here’s the story the Herald republished, and here’s a more sensible one from Buzzfeed. Not only is there no evidence for it, there’s evidence against it. New Zealand has plenty of scientists who know enough evolutionary biology to say the story is crap; it wouldn’t have been hard to ask one of them.
• A tweet about toothpaste with an unusual ingredient that “increases the defenses of teeth and gums…cells are loaded with new life energy, the destroying effect of bacteria is hindered… it gently polishes the dental enamel and turns it white and shiny.” Fairly typical non-quite-therapeutic claims, but this toothpaste is from the early 20th century and the unusual ingredient is radium
• A Guardian story with a nice animated chart showing the uncertainty in global warming predictions.
• From Quartz, “Satellite images reveal which countries cheat on their economic statistics”

From Science

Brooks then said that erosion plays a significant role in sea-level rise, which is not an idea embraced by mainstream climate researchers. He said the California coastline and the White Cliffs of Dover tumble into the sea every year, and that contributes to sea-level rise. He also said that silt washing into the ocean from the world’s major rivers, including the Mississippi, the Amazon and the Nile, is contributing to sea-level rise.

It’s true that this idea is not embraced by mainstream researchers. But there’s more than that. Unlike some of the more complex and subtle uncertainties in the rate and causes of sea level rise, this argument is not embraced by people who can use Google and a calculator.

The main contribution of the land to the sea floor is the 20 billion tons or so of silt from the world’s rivers.  Soil weighs in at between 1 and 2 tons per cubic meter; rock mostly between 2 and 3. Let’s be generous and call it 20 billion cubic metres. Not all of that ends up in the ocean — lots of it goes to coastal wetlands and deltas — but we’re messing them up too, so again let’s be generous and say it all ends up underwater.

The ocean area is 360 billion square kilometres, so about 360 million billion square metres. It would take 360 million billion cubic metres of added stuff to raise sea levels 1 metre — actually more, since the ocean area would expand, which is actually the main issue with sea level rise, but again we’ll be generous and say only 360 million billion. Dividing by 20 billion gives 0.000055 metres or 0.05 millimetres.  Actual sea level rise over the whole 20th century averaged 1.8mm/year, about 30 times more.

• “The Home Office seems to have used some sort of voice identification test in an effort to find international students who got someone else to take their mandatory English exam. And this went about as well as you might have expected if you read this blog.” from the Yorkshire Ranter
• NZ supermarkets have apparently been using computerised facial recognition systems to identify shoplifters.  Here’s the Privacy Commissioner on Radio NZ.  Of course, supermarkets have used non-computerised facial recognition systems for a long time, and there have definitely been concerns about the bias of these, too.  And while the current systems seem to have high error rates, that’s temporary: they are going to get better.
• From the Washington Post, a video on how opinion polls work
• “To work for society, data scientists need a Hippocratic oath with teeth” Cathy O’Neil in Wired.  I agree on the basic ethics and design principles, but statistics, unlike medicine or law or engineering, has never had a wall to keep people out: anyone can practise statistics. That makes professional regulation hard.
• “Across the city, black people were arrested on low-level marijuana charges at eight times the rate of white, non-Hispanic people over the past three years, The New York Times found.” This won’t actually surprise most people, but the NYT has done a good job of showing that the usual excuses don’t explain the bias.
• Beautiful infographic/visualisation of different types of alcoholic drinks by Rego Sen
• The birth cohort study Growing Up In New Zealand has been refunded
• The story of the HP-35 calculator, the first scientific pocket calculator. I’ve used the next model, the HP-45. Calculators were much better engineered when they cost the equivalent of what a Macbook does now.
• “Baking the most average chocolate chip cookie” from pudding.cool
• Why not to use two y-axes on a graph, from Datawrapper

Team Ratings for Round 14

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 80 matches played, 56 of which were correctly predicted, a success rate of 70%.
Here are the predictions for last week’s games.

Predictions for Round 14

Here are the predictions for Round 14. 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 11

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 80 matches played, 43 of which were correctly predicted, a success rate of 53.8%.
Here are the predictions for last week’s games.

Predictions for Round 11

Here are the predictions for Round 11. 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.

There’s a story in Stuff about a potential new cure for baldness

A cure for baldness could be on the horizon after British scientists discovered that an osteoporosis drug stimulates hair growth three times quicker than other drugs.

The first thing to notice about the story is the name of the new drug, WAY-316606. That’s not a drug name. That’s a drug-company internal name of a research compound. This isn’t a story about repurposing an already-approved drug. The name of the cancer drug “CsA” they supposedly started with also looks a bit weird.

The next thing to note is that we aren’t given much source information, apart from the Daily Telegraph. The story there has the headline Baldness cure could come from side-effect of cancer drug. You’d need to sign up there to actually read it; I wouldn’t bother if I were you.  Searching for keywords like “baldness” and “Manchester” finds the press release.

The press release tells us some interesting new facts. First, “CsA” isn’t a cancer drug at all. It’s cyclosporine A, the immune-suppressant drug that made organ transplants really feasible in the 1980s.  It looks as thought the Telegraph invented the connection with cancer — maybe by misunderstanding the fact cyclosporine increases cancer risk. Second, the research paper is in the journal PLoS Biology — unlike the Daily Telegraph story, it’s open-access.   The press release is also more measured about the medical potential, saying “Clearly though, a clinical trial is required next to tell us whether this drug or similar compounds are both effective and safe in hair loss patients.” As far as I can tell, WAY-316606 hasn’t ever been tested even for safety in humans.

The research paper is harder going. It does have this graph (Figure 3A) comparing hair growth in hair follicles treated with WAY-316606 and untreated controls. It’s true that the treated hair grew faster, and it’s true it was 2mm after 6 days, but the difference from untreated hair was a lot smaller than that.

The research paper links to another paper about WAY-316606, which is not open-access.  That paper gives a hint as to why WAY-316606 isn’t a drug already: it doesn’t last long enough in the body (at least, the bodies of rats). It might work better applied to the scalp. Or not.

So, this is interesting biochemical research that might turn out to have practical applications for hair growth, but it’s not really a way to use an existing drug to cure baldness.

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 73 matches played, 51 of which were correctly predicted, a success rate of 69.9%.
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.

Team Ratings for Round 10

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 72 matches played, 39 of which were correctly predicted, a success rate of 54.2%.
Here are the predictions for last week’s games.

Predictions for Round 10

Here are the predictions for Round 10. 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 Semi-finals

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 132 matches played, 93 of which were correctly predicted, a success rate of 70.5%.
Here are the predictions for last week’s games.

Predictions for the Semi-finals

Here are the predictions for the Semi-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.

• Creative arithmetic in some official statistics “an incident in the Rio Grande Valley Sector on February 14, 2017, involved seven U.S. Border Patrol Agents assaulted by six subjects utilizing three different types of projectiles (rocks, bottles, and tree branches), totaling 126 assaults.” 
• Interesting, slightly-nerdy post on fake news, network structure, and Cambridge Analytica
• Also on Cambridge Analytica: “The psychology ethics committee refused permission, and when he appealed to the [Cambridge] University Ethics Committee  this refusal was upheld”
• “Seven Visualization Talks That Terrified Me At CHI”: Michael Correll on this year’s Conference on Human Factors in Computing Systems
• Abstract art to a human; representational art to a neural network
• The success of statistical models in horse-racing (Bloomberg: limited free articles)
• The UK’s opt-out internet porn filter has lots of completely stupid false positives because of course it does (BoingBoing). There’s a site where you can search or browse them.  I think my favourite so far is Utopia Palms and Cycads, a nice ecologically-responsible tropical-plant supplier in Queensland, Australia.
• And in other UK computerised false-positives news “However, when ETS’s automated voice analysis was checked against human analysis, its computer programme was found to be wrong in 20 per cent of cases, meaning that more than 7,000 students were likely to have been wrongly accused of cheating.”  Sometimes 80% is a fail, not  an A-