Terminology Index
Glossary
1801 terms
Showing 257-288 of 1801 terms
C32
Calibrated Estimation
A skill for making numerical risk estimates whose stated confidence matches reality, so that when an estimator says they are 90% confident, they are right about 90% of the time. It produces honest, useful ranges instead of false precision or vague guesses.
In quantitative risk analysis like FAIR, experts must estimate values such as loss frequency or magnitude as ranges with confidence levels. Calibration training teaches people to avoid overconfidence and give well-tuned ranges, which makes the resulting risk numbers trustworthy enough to base decisions on.
Introduced in: Risk Quantification with FAIR
Examples
- Giving a 90% confidence range for an annual loss rather than a single number.
- Widening an estimate after calibration reveals past overconfidence.
- Stating loss magnitude as a range with an explicit confidence level.
No related terms linked yet.
