Prior and baseline

Record what you currently believe and what normally happens before you look at any new evidence.

Three weeks into a language app, you feel like it's working. Words come faster, the streak counter is at 21, and you're ready to tell friends it's the best method you've tried. But you can't actually remember how bad you were three weeks ago, and you have no idea how fast anyone improves in their first month of anything.

Prior and baseline is the move of recording, before you collect new evidence, two things: what you currently believe (including how confident you are) and what normally happens anyway. The cue is that you're about to start a test, a program, or a trial period. The operation: write down your current belief as a range ("I think this app is better than my old textbook, maybe 60 percent sure"), then capture the baseline ("before starting, I could hold a 30-second exchange; typical beginners plateau around week six"). When it works, you can later tell real improvement from normal early progress and from memory rewriting itself, because the before state is on paper instead of in hindsight. Skip the formality when the outcome is instantly obvious either way, like testing whether a key opens a door.