Statistic of the Year 2017: Winners announced | Hacker News Reader

Statistic of the Year 2017: Winners announced

statslife.org.uk

25 pointscolinprince11 comments

Oof. I hate this article.

Statistics have to be interpreted. Creating a "Statistic of the Year" is about as asinine and self-congratulatory as Blake Shelton being declared "Sexiest Man Alive".

What does it even mean?

Take Mrs. Kardashian West's post. Its sufficiently political to cause a controversy and sufficiently vacuous for readers to imprint their own thoughts and feelings on it.

What is an "Islamic", "Jihadist", "Immigrant"? If a terrorist travels to America, with an intent to kill but not an intent to obtain permanent residency, are they not included in the statistic? If a terrorist kills a group of people but a "Jihad" has not been formally declared by a group associated to the terrorist, is that terrorist not included in the statistic? If a terrorist is inspired by Wahhabis propaganda but is not himself a "true" Muslim (as defined by the researchers), is he included in the statistic?

The problem with these "award winning" statistics is that they are not, in fact, statistics. They are award winning implications which themselves are not born out of the data but are instead imprinted upon it.

The implication that toddlers are more dangerous than terrorists is absurd. Armed toddlers are the result of freak accidents and careless people. Terrorist attacks are the result of intentional policy choices. It's not hard to wonder why people focus so much time, effort, and attention on one rather than the other.

Consider the following scenario:

You are locked in a room with two other people. To your left, an armed toddler. To your right, an armed, islamic, jihadist, immigrant, terrorist. Your mission is to disarm both parties before you are allowed to leave the room. Who do you disarm first?

If you're the authors of this article, you disarm the toddler. If you're sane you disarm the terrorist.

> The implication that toddlers are more dangerous than terrorists is absurd. Armed toddlers are the result of freak accidents and careless people. Terrorist attacks are the result of intentional policy choices.

Armed toddlers are also a result of intentional policy choices. You will struggle to find a statistic about deaths from armed toddlers in the UK, because gun laws mean that the chance of such a situation occurring is minuscule.

That scenario is misleading as it assumes there are equal numbers of relevant toddlers and terrorists.

Firstly, the scenario is exact in its implication. All things held equal, an "Islamic", "jihadist", "immigrant" is treated with greater caution than an "armed" "toddler".

Secondly, you can not infer the population of "armed" "toddlers" or "Islamic", "jihadist", "immigrants" from this statistic. Terrorists could be incredibly common and safe or uncommon and dangerous. Increasing their commonality might not be desirable. Which directly contradicts the implication of the tweet.

Thirdly, proportion the population of toddlers and terrorists however you wish. Under any proportionment the terrorist would be addressed first. For the simple fact that terrorists have intent to harm, whereas toddlers are more likely to harm themselves.

I think the authors would argue that the statistic tells you more about the situation where there is a gun in the room, a toddler, and an adult who looks a bit foreign. Who should you be more worried about keeping the gun away from?

If the terrorist is not posing an immediate threat, I would consider disarming the toddler first. This could more easily supply me with a weapon, which would improve my odds against the terrorist.

Not relevant to the discussion, I know, but an interesting thought exercise the more I over-think it.

#2 statistic here is also largely bullshit based on a very flawed understanding of risk. "...help us better understand the world around us", this stat does the complete opposite.

Further reading: https://twitter.com/nntaleb/status/942730083973951490

Number of deaths caused by Islamic extremist terror attacks is a fat-tailed distribution, while deaths caused by lawn mowers is not. There is no organized movement among lawnmowers to rise up and kill people - the deaths happen by accident, and the odds of deaths increasing 1000% one year are very, very low. The same logic does not hold for terrorist attacks. There is no reason to assume an attack on the scale of 9/11 or greater will never occur again. These low-probability but highly significant events are not being factored in by Kim Kardashian West, and a lot of people are not thoughtful enough to consider them on their own.

The smugness is unbearable.

Nassim Nicholas Taleb would absolutely tear those statistics apart.

> "We are delighted to announce our first ever UK Statistic of the Year and International Statistic of the Year, a new initiative that celebrates how statistics can help us better understand the world around us"

TL;DR The winner is "This is the annual number of Americans killed, on average, by lawnmowers - compared to two Americans killed annually, on average, by immigrant Jihadist terrorists.", and this was shared by Kim Kardashian. https://twitter.com/kimkardashian/status/825580660337283073?...

Reply on news.ycombinator.com