false precision in bins? about covid19_expert_survey HOT 1 CLOSED

Hi afontenot,

Thank you for the thoughtful question, and for checking in on my work!
For the consensus probability distribution you plotted above, I binned the horizontal axis.
The question related to the number of deaths by the end of 2020 you potted above is from the survey issued on 2020-03-30 and was question number 4. For this question I used bins of size 500 (i.e. 0-500 deaths, 501-1000 and so on).

I can recompute the consensus distribution for all the questions for bins of size 1 (a single death) but I don't think models can really compute probabilities on a national scale at that resolution.

Please let me know if that answers your question.
best
tom

from covid19_expert_survey.