AlexSchell comments on Group Rationality Diary, August 16-31 - Less Wrong

2 Post author: therufs 18 August 2014 02:33AM

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

Comments (10)

You are viewing a single comment's thread.

Comment author: AlexSchell 18 August 2014 05:59:21AM 12 points [-]

I greatly benefited from a silly-seeming "information hazard management scheme" (suggestions for better/existing terms are welcome):

I was going to interview at my top choice med school that I applied to, School A, and knew that I would receive my first admissions decision, from School B, on the same day I would be interviewing. I was mildly confident (60-70%) that School B would accept me, and I really wanted to know their decision. If I were to be accepted, I figured I'd get a confidence boost that would improve my interview performance at School A. But finding that I hadn't been accepted would badly shake my confidence.

So I arranged to receive a noisy and biased signal of my admissions decision. I asked my sister to execute the following when prompted: Flip a coin twice; if the outcome is HH, stop there; otherwise log into my email account and send me a text message iff School B admitted me. This protocol has the effect of diluting the bad news with noise -- P(admitted | no text) ~ 0.3-0.4 -- while still being 75% likely to give me the good news if it exists.

On interview day at School A, the decision email arrived about 30 minutes before my actual interviews. I let my sister know, and waited. I didn't hear back, and as expected, I didn't think much of it. I resisted the temptation to open the email until after my interviews (the interviews went fine). It turned out that School B waitlisted me, which predictably wrecked my confidence. School A would later be the only med school to accept me.