Mean, median, mode (modal means 'typical').

Usually it's the median or the mode that's more informative than the mean. To be intellectually honest (maybe related: Responsible Research), journalists could just adopt the practice of reporting all three, or at least never using the word "average" in favour of specifying mean, median or mode. Even better, show the shape of the distribution anytime an average is mentioned, so we're reassured it's not multimodal or something else weird. Actually upon reflection, I find it odd and disturbing that it isn't already the norm in newspapers & blog articles to show the distribution.

Q: Can mode be calculated for continuous data? How? Interpolate a kernel density curve and find its tallest peak?

Your original motivation for something can be forgotten as you focus on a solution that grows ever more complex – and you act as if the solution is what you fundamentally want, because you forgot how you ended up working so hard on it.

The refrain "don't lose the root for the tree" reminds you to think of what you originally wanted and ask if there are different paths towards achieving it.

I guess the idea overlaps with the sunk cost fallacy, but it uses an active phrase (easier to apply) and is more general, covering sunk costs you forgot about. Also may remind you to do a "breadth-first search" of the solution space.