Hacker News Reader

Consider for your analysis: if you're walking in the same neighborhood, even for many many walks, those walks' risks will actually be quite correlated. So it's not like the probabilities combine the same way they would for independent random samples. The vast majority of children could not force an abduction to happen to them in their life even if somehow they spent all their waking hours outside. Thinking the probability approaches 1 is pretty absurd, and I suggest reflects some poor probabilistic thinking.

Approaches was probably the wrong word. Moving in the direction towards 1 was what I meant to say.

> However, the practise of letting your children walk results in many walks over time, and it takes just one abduction to make the situation really bad. In other words, over time, the probability of an abduction will (albeit slowly) go towards 1.

They will die of old age long before the chance of abduction gets anywhere close to 0.1 much less 1.

To be conservative, assume that 1% of US children are allowed to go unsupervised walks (~750k children) and that there is 0% chance of abduction by a stranger without allowing unsupervised walks. Each year about 350 children are abducted by strangers[1]. Under those assumptions, there is a 0.04% chance of abduction each year. That would be less than 1% lifetime chance assuming that they do this for all 18 years of their childhood. Even less if you make more realistic assumptions (I certainly hope that more than 1% of 17 year olds are allowed to go on walks and that parents are waiting until after their children can walk before letting them go on walks by themselves).

[1] https://www.reuters.com/article/us-wisconsin-missinggirl-dat...

Though with three children under the same policy, you get just over 2 % risk that at least one of them get abducted. The upside is hard to quantify (especially for me who view it as obvious that children should be allowed unsupervised walks) but I can totally see how the 2 % number with a very serious downside can get someone to doubt.

(I know what "conservative estimate" means, I'm just playing devil's advocate.)

> This can be interpreted to mean that the ensemble probability of an abduction -- the expected number of abductions from a single walk around the neighbourhood -- is low.

Not just low, very low, for the lifetime probability to be low enough for it to be that number.

You’re far more likely to walk your child straight into death by teaching them to drive.

Notice that demand for abducted children is inelastic in regard to number of children walking around the neighborhood. This means there are two equilibrium points of walking children market: first is letting them play as much as they wish, making probability of abduction of _your_ children as small as it can be; second is letting them out almost never, with probability of abduction for a walk being 1. Which of that do you prefer?

Haha, I can not tell if this is satire or what.

Somehow, I doubt the statistics of this so much that my first reaction is to chuckle.

I'm not sure what's funny. I'm looking for nuanced perspectives on both sides before I default to my biases, and I prefer arguments to be based on evidence. That's all.

I edited my original comment to clarify this.

I'm sorry, I did not want to sound mean, but the statistical claim made just felt so wrong to me as to be amusing. I can not formally prove it wrong, it just... Triggered the hummour detector.