You cut off the rest of my sentence. It's not that never having your beliefs challenged is good, it's that it's not a strong signal that your beliefs are actually wrong, and so your confidence can't actually diminish much from pointing out this fact, because everyone should already account for this in their confidence factors for beliefs they haven't actively tried challenging. Whereas if you're frequently changing your mind, well, that is a strong signal that your beliefs are frequently wrong, or that you're interpreting evidence poorly, either way that is a signal you should be even less confident in your beliefs than general. If you're frequently testing your beliefs, then it's the results of the tests, not the fact that you're testing, plus your ability to have your beliefs appropriately shifted by the results of the tests, which determine how good your beliefs are. The latter half of that equation is probably the bigger factor for what causes wrong beliefs anyway. Ideally your confidence in the belief should match its actual probability distribution, and shifts should be in correspondence with Bayes' rule. In reality many people will see a known-to-be-uniformly-independently-random binary process output a 'run' of Trues and feel more than 50% sure the next hit has to be False because it's 'due'. (https://en.wikipedia.org/wiki/Gambler%27s_fallacy)