"Let us say that, in general, failure (and disconfirmation) are more informative than success and confirmation, which is why I claim that negative knowledge is just "more robust."