Plain-language answer
Pre-registration means writing down a specific, precise prediction before an event, so that the criteria for “the prediction came true” are fixed before anyone knows the outcome. Researchers use it to guard against HARKing — quietly fitting a hypothesis to already-known results.(Kerr, 1998) The same discipline scales down to ordinary life: writing “I predict my next dream will feature a departure, defined as any scene of leaving a place” before you go to sleep is a genuinely different, stronger kind of evidence than deciding afterward that a dream about “packing a bag” counts as a match.
Why it matters
Almost every “I predicted this” or “I sensed this coming” story is reconstructed after the fact, when the specific match being claimed is already known — and once the match is known, human memory and pattern perception are very good at finding a way to make an earlier, vaguer feeling fit it (see hindsight bias and availability and memory). Pre-registering a prediction removes this entire problem at the source: there is no “after” for the matching criteria to have been quietly adjusted during.
How to pre-register an everyday prediction
- Write the prediction down, in a form you can’t edit later (a paper note, a dated message to yourself, a fixed record) — before the event whose outcome you’re predicting.
- Specify the match criteria precisely, the same way step 5 of how to evaluate a coincidence claim asks for a precise match definition: what exactly would count as the prediction coming true, and what would count as a miss?
- Commit to counting misses, not just hits. If you predict something every night for a month, record every night, not only the nights it seemed to work — the miss rate is what turns an anecdote into usable data.
- Resist widening the criteria after the fact. If the outcome is close but doesn’t quite meet your stated criteria, record it as a miss (or a partial, explicitly noted near-miss) rather than expanding the definition to make it count.
- Review the full record, hits and misses together, before drawing any conclusion about how often your predictions actually succeed.
Common misconception
“I don’t need to write it down, I remember what I predicted” undermines the entire point. Unrecorded predictions are reconstructed from memory after the outcome is known, which is exactly the condition pre-registration exists to rule out — memory for a prediction, once you know how things turned out, is not a neutral record of what you actually expected beforehand.
Limits and open questions
Pre-registration controls for the “was the match defined before or after the outcome” problem specifically — it does not, by itself, control for other sources of bias, such as vague prediction wording that leaves room for generous interpretation even when written down in advance, or selectively remembering to check some pre-registered predictions and not others. Precise wording (step 2 above) and consistent record-keeping (step 3) both remain necessary alongside pre-registration itself.
Related
- Post-hoc probability covers the general problem — a target painted after the arrow lands — that pre-registration is designed to prevent.
- How to evaluate a coincidence claim — this page develops step 4 (“ask when the matching rule was chosen”) into a practical personal habit.
Key takeaways
- Writing a precise prediction down before an event, not reconstructing it afterward, is the single clearest way to distinguish real predictive success from after-the-fact pattern-matching.
- A pre-registered record is only useful if it includes the misses, not just the apparent hits.
- Vague criteria, even if written down in advance, still leave room for the same after-the-fact flexibility pre-registration is meant to remove — precision in the prediction’s wording matters as much as its timing.