FTA: “If either fault were to rupture, it would be the single-worst natural disaster in American history. It would kill anything from 5,000 to 100,000 people, cause over $500 billion in property damage, and destroy some of the country’s most important infrastructure. Combined, these faults pose one of the biggest threats to American well-being and prosperity in the twenty-first century. And yet, efforts to mitigate this risk remain alarmingly limited.”

and

“genuine pathways to the ‘holy grail’ breakthrough have been limited. This grail would be minutes or hours of advance warning, since even just stepping outside would prevent the majority of fatalities and injuries.”

I think that’s the problem: except for limiting the number of deaths and injuries, better earthquake predictions at “hours of warning” do fairly little to mitigate the risks of earthquakes.

Also, the number of deaths isn’t that high. 100,000 is a lot, but with these big ones happening less than once a century, statistically it is less than 1,000 a year, I guess predominantly of the less well-off that cannot afford to live in a house that follows modern building codes.

And that’s assuming predictions are perfect. If they’re imperfect, will people spend hours outside of their airconditioned homes after having done so a few times after a warning that didn’t turn out to be necessary? Even if they do, how many deaths will that cause?