Forecasting terms
Forecasting, in plain words
Short definitions of the words you meet when you make calls on Obvious Later, each with a small example.
- Probability call
- A number from 0 to 100 that says how likely you think an event is. A call of 80 means you expect it to happen about four times out of five in situations like this one.
- Calibration
- Calibration compares what you say with what happens. Take every call where you said 70. If about 70 percent of those events happened, you are well calibrated. If only 50 percent did, you were overconfident.
- Brier score
- A score for one call. Write your probability as a number from 0 to 1, subtract the outcome (1 if it happened, 0 if it did not), and square the result. Saying 0.8 on something that happens scores 0.04. Saying 0.8 on something that does not scores 0.64. Lower is better, and a flat 50 on everything scores 0.25.
- Base rate
- How often events of this kind happen in general. Before you weigh the news, ask what usually happens. If similar launches usually slip, start low and let the specifics move your number.
- Resolution
- The moment a question is settled. Each question names its source and its date up front, so everyone knows what counts as a yes and what counts as a no.
- Crowd forecast
- The combined view of everyone who has made a call. On Obvious Later you see it only after your own call is locked, so the crowd cannot nudge you.
- Timestamped call
- Your call saved with the time you made it, so you can show later what you thought before the answer was known. Each call gets a public fingerprint on the ledger that anyone can check.
- Hindsight bias
- The feeling that you knew it all along once the answer is out. A call saved before the answer makes it harder to fool yourself.
- Overconfidence
- Giving extreme numbers like 95 or 5 more often than reality earns them. When you feel sure, ask what would have to be true for you to be wrong.
Ready to try it? Make today's call
Want the scoring details? Read the methodology