[wrsh07]

They actually respond to you in the FAQ:

> "What if people who lived several centuries ago did a calculation on how many births there would be?"

> This appears to be one of the most widespread misconceptions on the topic. Many scientists have fallen into this trap, such as Lee Smolin's article from 2004 . In science there is never absolute certainty, only varying degrees of confidence. We should never be 100% sure of anything. When stating the degree of confidence in a result, typically 95%, it should be in full knowledge that one time out of twenty, we will be wrong. 5% of the time we will be misled by statistical chance.

> Now if someone who lived tens of thousands of years ago estimates the total number of human births, based on how many there had already been, they will underestimate the truth. Because we now know there has been many more. But those first 5% of people who ever lived represent the 5% of the time we expect to be wrong. This is a basic premise of how science functions, how it uses statistics. We must be wrong some of the time. In reality, we are wrong much more frequently than statistical chance suggests, because of human error or misunderstanding.

[codys]

> several centuries ago

> tens of thousands of years ago

This appears to be blatantly moving the goal posts.

> those first 5% of people who ever lived represent the 5% of the time we expect to be wrong

This implies that for only 5% of human history the calculation would be wrong, and the reset of the time it would be 100% correct.

I'm not sure that is how statistics work.

[fergussimpson]

That's exactly how statistics works.

You make a statement with some degree of confidence (often 95%) and fully expect to be wrong 5% the time.

[wrsh07]

> This appears to be blatantly moving the goal posts.

It doesn't actually matter. All we care about are values by %. So we can say that someone in the first 5% of humans could have done this same calculation and been wrong. And that's okay.

If it turns out that we are in the first 5% of humans, then we too are wrong. That's a risk we're willing to take, and statistically, it doesn't matter. It allows us to make informed decisions right now.