I know what I would expect to observe if my computer weren’t fast enough (even in the absence of looking at technical specs), but I don’t know what I would expect to notice if I were safe in the absence of actually looking at the precautions that are being taken.
The closest thing I can come to that is observing that nothing disastrous has happened yet, but that’s not especially well-correlated with actual safety.
So… what kind of evidence are you envisioning here?
I’m envisioning observing that nothing even very bad has happened yet, which I think is in fact pretty well correlated with actual safety.
It’s not the same thing, for sure. But it’s probably all we have, on any side of the debate, and it seems to me to support the “no need for panic” side better than the “lock down all the borders and quarantine everyone arriving from Africa” side.
There usual context for this example is a conversation about relying on historical data and not paying too much attention to “external” (or “meta” or “structural” or… etc.) factors.
I know what I would expect to observe if my computer weren’t fast enough (even in the absence of looking at technical specs), but I don’t know what I would expect to notice if I were safe in the absence of actually looking at the precautions that are being taken.
The closest thing I can come to that is observing that nothing disastrous has happened yet, but that’s not especially well-correlated with actual safety.
So… what kind of evidence are you envisioning here?
I’m envisioning observing that nothing even very bad has happened yet, which I think is in fact pretty well correlated with actual safety.
It’s not the same thing, for sure. But it’s probably all we have, on any side of the debate, and it seems to me to support the “no need for panic” side better than the “lock down all the borders and quarantine everyone arriving from Africa” side.
OK. Thanks for clarifying.
The classic counter-example involves a turkey in the middle of November. Nothing very bad has happened to it ever—must be very safe, then...
Sure. As with everything else in the universe, we have to make do with partial information and sometimes it’ll lead us astray.
There is a bit more wisdom here :-)
There usual context for this example is a conversation about relying on historical data and not paying too much attention to “external” (or “meta” or “structural” or… etc.) factors.