**Comment of the Day: Robert Waldmann**: *Cognitive Science, Behavioral Economics, and Finance: Some Fairly-Recent Must- and Should-Reads*: "On Cosma. I must read Judea Pearl. My reply to Shalizi is that we can and do have priors to within some number in the total variation norm...

...the probability is between p and p+10%, say. That's enough that Bayes's formula is useful. That's not so much that we must claim to have absolute confidence in very precise subjective probabilities. There is positive breakdown point sort of Bayesian reasoning closely related to positive breakdown point parameter estimation (the breakdown point is the proportion A of data such that I can trick you into believing anything if I can replace fraction A of observations with something I myscheviously make up)...

```
#commentoftheday
```