4 ms·

A robust Bayesian look at Covid origins

• ggm 3y ago
  "Robust" please. It's "I assign arbitrary weight to papers based on mood and vibe, and discount views I don't like about wet market vs lab leak" Nothing wrong with a Bayesian approach to summing over your weights, but the weights are still subjective, no?
  • hfgjdssaghj 3y ago
    > based on mood and vibe Can you give a solid example of the “mood and vibe” methodology in the analysis?
    • ggm 3y ago
      Explain why and how you assign weight to the furin cleavage site. At this point it’s necessary to recognize that not only the prior odds P­0(LL)/P­0(ZW) but also the likelihoods involve some subjective estimates. In order to obtain a convincing answer we need to include some range of plausible values for each likelihood ratio Your judgement of likelihood? Your range of estimates is informed by what exactly?
      • hfgjdssaghj 3y ago
        There are ten long and very detailed technical paragraphs in the article itself, and even an appendix with even more detailed methodology to explain this exact question. I really don’t know what more the writer needs to do, if this thorough and scientific analysis is characterized as “mood and vibe”.
        • ggm 3y ago
          You include self published, non peer review sources and you appear on the surface to be selecting inputs to weight for your goal seeking. This is whitewashing opinion through a statistical mixer function to try and project objectivity when it's patently not. I suggest you put it into peer review in an AMS recognised statistics journal.