Lumifer comments on Whole genome sequencing vs SNP genotyping - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (9)
That line of argument is flawed because actions have costs.
To give a simple example (and discussing whether it's precisely correct is besides the point), you can take baby aspirin to reduce the risk of heart attacks. Aspirin is a blood thinner, it makes clots (which cause heart attacks) less likely. However decreasing coagulation is not an unmitigated blessing. If you get internal bleeding -- e.g. a blood vessel ruptured in your brain -- that aspirin you've been taking could make things significantly worse. It's a trade-off.
Given this, you want to know on which side (basically, heart attacks vs. strokes) is your personal risk the highest. It is actionable knowledge.
Looks very different to me -- you are quite cavalier with a fourfold difference in odds...
4 is a small number. It is pretty rare that a cost-benefit calculation cares about that factor of 4, that multiplying the benefits by 4 will change the decision from reject to accept.
It seems we have a different perception of smallness and tend to encounter different cost-benefit calculations.