What does it mean that distinguish signal patterns from noise patterns?

Not every recurring event is meaningful — some repetitions are coincidental.

Example: You notice you've had three bad meetings on Tuesdays and conclude 'Tuesdays are my worst day.' But you've had 200 Tuesday meetings total — three bad ones is a 1.5% rate, no different from any other day. You detected a pattern, but the pattern was noise. A signal pattern would be: every meeting where you skipped your 10-minute prep routine went poorly, regardless of the day. One pattern is coincidence dressed as insight. The other is a structural cause you can act on.

Try this: Pick a pattern you currently believe about yourself — a recurring observation from your notes, journal, or memory. Write it down as a claim: 'When X happens, Y follows.' Now apply three filters. (1) Sample size: how many times has this actually occurred, vs. how many times X happened without Y? (2) Alternative explanations: what else could explain Y? List at least two. (3) Prediction: if this pattern is real, what should happen next time X occurs? Write down your prediction and watch for the result. If you can't survive all three filters, the pattern is probably noise.