Cyber Risk Lab

Defensive, educational estimators. Nothing here touches a real system.

Entropy H = L·log₂N-
Keyspace NL-
Mean time to crack-
Verdict-
Assumes the password is random. Human-chosen passwords are far weaker than their length suggests; use a password manager and MFA.
Direct + regulatory-
Reputation-
Total-
Cost per record (all-in)-
Illustrative linear model with made-up default rates, not a forecast; real per-record costs fall as breach size grows.
Total timeline-
Time to respond (detect → recovered)-
72 h notification rule-
Longest phase-
The 72-hour window (e.g. GDPR Art. 33) runs from becoming aware of the breach, taken here as the end of detection.

Self-check

running…
The science

Password entropy. For L characters drawn uniformly from N symbols, H = L·log₂(N) bits and the keyspace is NL. An attacker who tries guesses in random order finds the secret after NL/2 guesses on average, so t = NL / (2·r) at r guesses per second. Each extra character multiplies the time by N.

Breach cost. Total = records × cost × m + months × loss, with m the regulatory multiplier. Pure arithmetic over assumed inputs.

Response timeline. Phases run back to back: detect, contain, eradicate, notify. Time-to-respond is the span from the end of detection to the end of eradication. Shorter detection is the biggest lever because every later phase starts only once it ends.

Simplifications. Attackers use dictionaries and rules rather than pure brute force; hashing, salting and rate limits change r by orders of magnitude; the cost and timeline inputs are illustrative defaults, not statistics.