Rohin Shah comments on Reducing collective rationality to individual optimization in common-payoff games using MCMC

[Rohin Shah]

Two typos in the proof that confused me:

Since in general x/​1+x≤1−1/​x for positive x

Sign should be flipped.

A non-optimal state has probability at least umin

Should be “A non-optimal state has utility at least umin”.

[jessicata]

Thanks, fixed.