You're looking at Less Wrong's discussion board. This includes all posts, including those that haven't been promoted to the front page yet. For more information, see About Less Wrong.

Lumifer comments on Whole genome sequencing vs SNP genotyping - Less Wrong Discussion

5 Post author: harcisis 11 June 2015 10:09PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (9)

You are viewing a single comment's thread. Show more comments above.

Comment author: Lumifer 12 June 2015 03:34:45PM *  4 points [-]

If the baseline is a 30% chance of dying of heart disease and it told you that you had a 70% chance, what would you do differently? (but it won't.) Probably you should already be doing it because 30% is a big number!

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.

an 80% chance of breast cancer, which is really not that different from a 20% chance

Looks very different to me -- you are quite cavalier with a fourfold difference in odds...

Comment author: Douglas_Knight 12 June 2015 04:46:51PM 1 point [-]

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.

Comment author: Lumifer 12 June 2015 04:49:53PM 4 points [-]

It seems we have a different perception of smallness and tend to encounter different cost-benefit calculations.