Glossary

The terms our results are reported in, explained without formulas.

P80 (and P50, P90…): What it means: The value you have that much chance of NOT exceeding. A P80 finish date of 15 March means that in 80% of the simulated futures the project finished on or before 15 March, and in 20% it finished later.; How to use it: Pick the P-value that matches how much risk you accept. Many organisations commit dates and budgets at P80; P50 is a coin toss.
Probability of meeting the target: What it means: The share of simulated futures in which the project finished by your target date (or within your budget).; How to use it: If it is low, either the target is optimistic or the plan needs changes. Compare it before and after a mitigation to see what the action buys you.
Contingency (reserve): What it means: The extra time or money between the plan’s own estimate and the P-value you choose to commit to.; How to use it: Size the reserve from the simulation, not as a flat percentage: contingency = P80 (or your chosen level) minus the deterministic plan.
Iterations: What it means: How many possible futures the simulation plays out. Each iteration draws every uncertain input once and computes the result.; How to use it: More iterations make the percentiles more stable, especially in the tail (P95 and beyond). They do not make the inputs more accurate.
Correlation: What it means: Whether uncertain inputs tend to move together. If bad weather delays one task it probably delays the others too: those durations are positively correlated.; How to use it: Ignoring correlation usually makes outcomes look safer than they are, because it assumes bad luck in one place is offset by good luck elsewhere.
Tail: What it means: The extreme end of the results: the rare but severe outcomes, such as P95, P99 or 99.9%.; How to use it: Tail figures need many iterations to be stable. Read them as orders of magnitude when the run was short.
VaR (Value at Risk): What it means: The loss that will not be exceeded with a given confidence over a given horizon. A 10-day 95% VaR of 1 million means a loss above 1 million happens in about 1 of every 20 such periods.; How to use it: Use it to set limits and compare portfolios. It does not say how bad a loss beyond the VaR can get — that is the Expected Shortfall.
Expected Shortfall (ES): What it means: The average loss in the cases that go beyond the VaR.; How to use it: It answers “when it goes badly, how badly on average?”, which VaR alone does not.
Expected loss (EL): What it means: The average loss per year (or period) across all simulated outcomes.; How to use it: It is a cost of doing business: price it in, budget for it or provision it.
Unexpected loss / economic capital: What it means: How much the loss at a high confidence level (such as 99.9%) exceeds the expected loss.; How to use it: It is the cushion of capital needed to survive a bad year. It is what capital requirements and RAROC are measured against.
OpVaR: What it means: The Value at Risk of operational losses: the yearly loss from incidents, failures and fraud that is exceeded only with a small probability (for example 1 in 1,000 years at 99.9%).; How to use it: Use it to size operational-risk capital and to see which event types drive it.
PD (probability of default): What it means: The probability that a borrower stops paying within a given period, usually one year.; How to use it: It is the first of the three building blocks of credit loss: how often loans go bad.
LGD (loss given default): What it means: The share of the exposure that is lost when a borrower defaults, after recoveries and collateral.; How to use it: It is how bad a default is. An LGD of 40% means 60% is recovered.
EAD (exposure at default): What it means: How much the borrower owes at the moment of default.; How to use it: Expected credit loss = PD × LGD × EAD.
RAROC: What it means: Risk-adjusted return on capital: the profit a portfolio earns after expected losses and costs, divided by the economic capital it consumes.; How to use it: Compare it with the hurdle rate. Above the hurdle, the segment creates value for the risk it takes; below it, it destroys value.