RE: https://bayes.club/@statmodeling_bot/117254268120426935
I am a big fan of highest density intervals (HDIs/HPDs) because a large fraction of the distributions we encounter in astronomy are heavily skewed and bounded below at zero. Doesn't hurt that they are also the _shortest_ intervals for a given included probability.
But yes, be aware that they aren't stable to transformations and fall down at multimodality.