Meaning
Statistical calculations derived from observed data quantify the probability and magnitude of extreme outcomes. Using empirical tail estimates allows a person to map the edges of a distribution where traditional models often fail. These calculations rely on historical maximums and minimums, omitting the assumption of a symmetrical curve.
The process requires a sufficient sample size to capture rare events that occur with low frequency but high impact.
Observation Frequency
Accuracy of these figures depends on the quality and duration of the historical record. When the observation period is too short, the empirical tail estimates often understate the true risk of extreme loss. Data points from volatile periods provide the necessary weight to the ends of the distribution.
A full dataset ensures that the model captures the reality of the environment.
Calculation Method
Projecting the probability of events beyond the highest recorded value involves fitting a power law to the existing data. This method allows empirical tail estimates to estimate the likelihood of a catastrophe that has not yet occurred. Analysts look for the slope of the tail to determine how quickly the probability of an event drops as its magnitude increases.
A shallow slope indicates a higher probability of extreme outliers. Calculating the area under the curve provides a cardinal value for the risk of ruin.
Predictive Limit
Systemic shifts invalidate these figures when the underlying environment undergoes a structural change. If the rules of the market change, the old empirical tail estimates no longer apply to the future state. Relying on old data during a regime change leads to a false sense of security.
The model provides a map of the past.