See also: Goodhart's law
I'm also thinking of putting a potentially large bounty on solutions to it which work in practice or which help mitigate it. In particular, I've recently become more skeptical of being able to implement the canonical Bayesian adjustment in practice.
Notes: https://wikiless.org/wiki/Winner%27s_curse?lang=en
https://www.lesswrong.com/posts/5gQLrJr2yhPzMCcni/the-optimizer-s-curse-and-how-to-beat-it#tzmqXnWvTiL69H2sR
See also: Goodhart's law
I'm also thinking of putting a potentially large bounty on solutions to it which work in practice or which help mitigate it. In particular, I've recently become more skeptical of being able to implement the canonical Bayesian adjustment in practice.
Notes: https://wikiless.org/wiki/Winner%27s_curse?lang=en
https://www.lesswrong.com/posts/5gQLrJr2yhPzMCcni/the-optimizer-s-curse-and-how-to-beat-it#tzmqXnWvTiL69H2sR