Sociologist Thomas Elliott has written a pretty good routine to predict outcomes in this show. His website is here:

http://thomaselliott.me/blog/category/rupauls-drag-race/

Best,

Chris

The probability of loss isn’t quite right–the softmax gives the correct probabilities assuming a Gumbel distributed error term, which isn’t symmetric. It might be a reasonable approximation sometimes, but can go pretty wrong.

https://discourse.mc-stan.org/t/post-on-ranked-random-coefficients-logit/7136/3?u=james_savage

]]>