"Our estimate for stage N in the drake equation is way too high" == "The great filter is stage N in the drake equation."
So the interpretation chart goes:
- No errors: There are tons and tons of aliens and they're prime-directive-ing us like the Federation from Star Trek.
- One error: Oh no! Great filter! (unless it's in our past)
- Many errors: Not as interesting.
As for why it's compelling at all, the original point of the Drake equation was to make you think about it, and also realize that it's very, very tough to get exactly one. Almost every set of numbers you pick results in either less than one (which we know isn't true because we exist), or more than one. So the argument is, who knows how alone we are, but we're not alone.
Of course that was 60 years ago or whatever, but you are on the same page as the guy that came up with it.
Imagine creation of intelligent life requires 36 uncorrelated parameters to line up in a planet, an each of those parameters have a 10% chance of occurring in any given planet. There would be only one such planet in the universe in all likelihood. Our loneliness in the universe doesn't seem so implausible now, does it?
Plugging in and modeling the statistics based on what we know has proven to be a very effective means of scientific discovery in other areas. It helps narrow down where one can find direct evidence.
Perhaps I remember it so vividly because it was the perfect setting: I was 3+ months into hiking the Pacific Crest Trail, North California, night hiking by myself, earphones on, the wildfire smoke started to clear up, the air was crisp and the night sky got sprinkled with countless bright stars for the first time in a while.
It is well known that the product of variables with an uniform distribution has a result that has not an uniform distribution, but some distribution that has more values closer to 0. So there is no surprise that multiplying almost 10 variables with uniform distribution you get something that almost 0 almost all the time.
But it is not realistic to use an uniform distribution. All the variables depend on a lot a subparts, so a realistic distribution must be something that looks like a log normal distribution with a peak near the current best guess https://en.wikipedia.org/wiki/Log-normal_distribution
Its much more likely that our estimates are off or our calculations are incorrect than we can draw conclusions from these roundabout statistical inferences. Nothing I’ve read has convinced me otherwise. Its a great pop-sci discussion but I’m more interested in direct evidence rather than speculation for why no direct evidence exists.
Drake Equation fans please explain what makes these theories so compelling and I will be indebted to you.