Error

Unrecognized LW server error:

Field "fmCrosspost" of type "CrosspostOutput" must have a selection of subfields. Did you mean "fmCrosspost { ... }"?

RyanCarey Jun 13, 2016, 5:48 PM
2 points
0 ∶ 0

Paul has said quite a bit about this previously:
MichaelDickens Jun 14, 2016, 2:36 PM
1 point
0 ∶ 0


Your best bet is to use your prior between the two, with a 90% chance of AMF being the best and a 10% chance of MFA being the best, for a weighted average of 0.9$3000 + 0.1$30 = $2703 per life saved, making the information worth ~$24.3M.

I believe you’re doing this upside down. You should be calculating lives saved per dollar, not dollars per life saved, so your weighted average is 0.9*(1/$3000) + 0.1*(1/$30) = 0.003633 --> $275 per life saved, making the information worth ~$2.4M. So you’re overpaying the data scientist by an order of magnitude.

Also, it would be nice if your footnotes were hyperlinks.
Tom_Ash Jun 13, 2016, 7:32 PM
1 point
0 ∶ 0

I’m interested as to different people’s views on whether they can beat the wisdom of the EA crowds on this.* Those who think they can’t might theoretically want to give to a portfolio of charities based on a particular crowd’s pick. We’ve been talking about collaborating with Michael Page to make something like this happen, though my purely personal estimate is that it won’t happen any time soon, and I’m not sure how many people would donate to certain sorts of portfolios—that’d also be interesting to hear!
- Gleb_T Jun 19, 2016, 4:02 AM
  1 point
  0 ∶ 0
  Parent
  
  I think the wisdom of EA crowds is kind of hard to estimate, as we don’t have a true and definitive evaluation of the actual value of the investments in a portfolio. Moreover, a complicating factor is that people frequently donate based on what they personally value, and have different trade-offs (for example, the value of a current human life vs. current animal life vs. future human/animal life).
- number42 Jun 17, 2016, 7:19 PM
  0 points
  0 ∶ 0
  Parent
  
  I might give to a portfolio in a particular cause area like meta. I’d only be giving a small portion of my pledge to that though, perhaps about £100 a month—would that even be worth it for the admin hassle for you?
  - Tom_Ash Jun 17, 2016, 7:52 PM
    0 points
    0 ∶ 0
    Parent
    
    Probably not! We plan to use an off-the-shelf service from somewhere like Vanguard in the US or the UK’s Charities Aid Foundation if we do this at all, but even so.
- Tom_Ash Jun 13, 2016, 7:32 PM
  0 points
  0 ∶ 0
  Parent
  
  (*For my part, I’m unsure of my credence that I can, but if forced to pick one I’d put it at below 50%, though that’s for various reasons which are quite idiosyncratic to me.)