I can't believe you're having this conversation either. You're correct, obviously; you've thrown out tons of trivially correct examples to make it clear to everyone just how right you are, none of which have anything to do with this. You are of course aware that 1% out of a population has nothing to do with the concept of a 1% error on your taxes or 1% blood alcohol. Alcohol does not affect you 1% of the time if you have a 1% BAC.
Everyone here knows that if you make the population arbitrarily large, the 1% sample becomes large too. But can you really argue that being able to represent 99% with a binary spectrum isn't a pretty good approximation? What percentage would be good enough for you? Or are you going to say "99.9% isn't good enough because 7 million is a lot of people. that's more than died in X'?
> I can't believe you're having this conversation either. You're correct, obviously; you've thrown out tons of trivially correct examples to make it clear to everyone just how right you are, none of which have anything to do with this. You are of course aware that 1% out of a population has nothing to do with the concept of a 1% error on your taxes or 1% blood alcohol. Alcohol does not affect you 1% of the time if you have a 1% BAC.
So you agree then that whether 1% is negligible is based on context?
> Everyone here knows that if you make the population arbitrarily large, the 1% sample becomes large too. But can you really argue that being able to represent 99% with a binary spectrum isn't a pretty good approximation? What percentage would be good enough for you? Or are you going to say "99.9% isn't good enough because 7 million is a lot of people. that's more than died in X'?
Yes. In case you didn't notice, 7 million people is a lot of people.
>So you agree then that whether 1% is negligible is based on context?
Yeah, or as I put it: "Numbers get their significance relatively, not in themselves".
Whereas you repeatedly stated how 70 million people is a huge number in itself.
E.g. If I told you there are 70 million people that have blue eyes, is that "a huge number?" No, it's actually a small number. One would expect blue-eyed people to be in the 100s of millions or billions.
> Whereas you repeatedly stated how 70 million people is a huge number in itself.
I haven't claimed that at all. I've said over and over that context indicates whether it's important.
> Whereas you repeatedly stated how 70 million people is a huge number in itself.
No, I've said 70 million people is a lot in terms of medical and social policy. It would be very possible, for example, for 1% of people to account for 10% of medical expenses--an amount you would probably care about at tax time. I think that's a number that matters to almost anyone's political goals.
In contrast, you've been repeatedly stating how 1% is not a large number. Based on what?
> E.g. If I told you there are 70 million people that have blue eyes, is that "a huge number?" No, it's actually a small number. One would expect blue-eyed people to be in the 100s of millions or billions.
Science doesn't give a shit about your expectations. "Expectations" are entirely irrelevant to whether a number is big or little. A number is big or little depending on what effects it causes and what effects you're trying to achieve.
You're accusing me of arguing that 70 million is inherently a large number, but you're arguing that 70 million is inherently a small number, completely arbitrarily. I'm not even saying 70 million people is a big or small number inherently, I'm saying that 70 million people is a huge number when the properties of that group have medical and social implications. 70 million blue-eyed people isn't a small or large number, it's an irrelevant number, because whether or not someone's eyes are blue has almost no implication that I care about. If you understand why it's not a big or small number, but an irrelevant number, you'll understand my point.
>Science doesn't give a shit about your expectations.
Language.
Also, I didn't say it's about "MY" expectations. It's about what the expected distribution is, which is the whole context that makes something big or small.
"Expectations" are entirely irrelevant to whether a number is big or little.
Actually, it's all about that. Bringing 10,000 times 6 by throwing dice 20,000 times is too big, because the expected outcome is about 1/6 throws to be 6.
Everyone here knows that if you make the population arbitrarily large, the 1% sample becomes large too. But can you really argue that being able to represent 99% with a binary spectrum isn't a pretty good approximation? What percentage would be good enough for you? Or are you going to say "99.9% isn't good enough because 7 million is a lot of people. that's more than died in X'?