I am being reminded that there’s no real way to caveat “scenario X seems like a possibility with a 20% chance of being true” sufficiently that people won’t read it as “X is true.”
Which is unfortunate, because it’s usually a good exercise to make yourself consider a range of explanations rather than just planting a flag in the one you find most likely. Gives you a place to slot new information & update your probabilities.
It’s easy to miss the significance of new data points if you don’t have lower probability models loaded in the back of your head somewhere. If you do you’re more apt to notice when new info fits one of the initially disfavored models better than your operative one.
