Vasco Grilo🔸 comments on My current best guess on how to aggregate forecasts

[Vasco Grilo🔸]

Hi Jaime,

How would you aggregate forecasts for extinction risk, spanning many orders of magnitude, which are based on fitting the same data to different probability distributions? I suspect your diagram would suggest using the mean, as each distribution relies on mutually exclusive assumptions, but I am not confident. I am asking in the context of this and this analyses, where I used the median (see discussion). Feel free to share your thoughts there.

[Jaime Sevilla]

Interesting case. I can see the intuitive case for the median.

I think the mean is more appropriate—in this case, what this is telling you is that your uncertainty is dominated by the possibility of a fat tail, and the priority is ruling it out.

I’d still report both for completeness sake, and to illustrate the low resilience of the guess.

Very much enjoyed the posts btw

[Vasco Grilo🔸]

Thanks for the feedback! I am still standing by the median, but it would be nice to empirically investigate similar cases, and see which method performs better!

[Ryan Greenblatt]

(I think it seems in principle promising to try to setup toy models of tail prediction, but I’m unsure what the right toy model is here.)