Application Application: apply foundational ideas to real estimates.
Fermi Estimates

Estimate the Unknowable with Structure

Break big questions into knowable parts, then rebuild a defensible estimate.

You will build each estimate yourself. This is hands-on.

Learning Objectives

By the end, you will be able to:

  1. Break a large unknown into measurable components.
  2. Estimate each component with a defensible range.
  3. Combine components into an order-of-magnitude answer.
  4. Explain the assumptions behind your estimate.

Recall from Bayes: Base rates are your starting point. Write them down before you decompose the problem.

What is a Fermi estimate?

Fermi estimation is a structured way to estimate hard questions using simple parts.

Break the big number into smaller numbers you can reason about, then combine them.

We care about the order of magnitude, not a false sense of precision.

Transfer Examples

Use the same breakdown logic in these situations:

  • Insider threat volume: 3,000 employees × 0.2% likely attempts per year = ~6 potential exfil attempts.
  • Vulnerable asset count: 120 apps × 30% internet-facing × 15% unpatched = ~5-6 high-risk assets.
Learning Debrief

What You Just Learned

  • Break unknowns into measurable components you can defend.
  • Use ranges to show uncertainty instead of false precision.
  • Focus on the few inputs that drive most of the estimate.
  • Communicate the order of magnitude with confidence.

Applying This to Cyber Risk

Fermi estimates help you make decisions when the data is incomplete.

Insider Incident Volume

Estimate potential insider incidents with a simple chain: employee count × attempt rate × detection probability.

Exposure Surface Sizing

Estimate high-risk assets by combining asset inventory × internet exposure × patch gap.

Continue to Applied Forecasting