Reading your results

Your results describe a range of possible annual losses, not a single prediction. Here's how to read them.

The loss exceedance curve (LEC)

The headline chart. For every dollar amount on the horizontal axis, the curve shows the probability your annual loss will exceed that amount. Read it as: "there's a 20% chance we lose more than $X in a year." It's the standard way FAIR communicates risk because it answers the question decision-makers actually ask — how likely is a loss this big?

The key numbers

• p95 / p99 — the loss levels you'd exceed only 5% / 1% of the time. These are your "bad year" and "very bad year" figures.
• CVaR-95 — the average loss across the worst 5% of simulated years. It captures how bad the bad years are, not just where they start.
• Mean — the simple average. Cyber losses are driven by rare large events, so the mean can sit well above the typical year; that's why the tool leads with the tail figures, not the mean.

Choosing your detail level

The Results page has a detail toggle with four levels — Admin, Management (the default), Analyst, and Explore. They show the same underlying run; they differ only in how much statistical detail is on screen. Start at Management; switch to Analyst when you want confidence intervals, the full percentile table, and convergence diagnostics, or to Admin for a board-ready summary.

A note on trust

Every reported number carries the uncertainty of the simulation itself — the tool shows confidence intervals on the figures that matter and nudges you to run more iterations when a tail number is still noisy. For the full picture of what's validated and where the model's edges are, ask your DecipherRisk contact for the model card.