Section 009 · Chapter 2, From Tests to Release Evidence

Risk-Based Sampling

Testing effort should follow risk. High-impact failures deserve more samples, stricter gates, and deeper review.

risk-based samplingrisk based sampling

What to do

  1. Test the single-signal cases.
  2. Test sensor disagreement.
  3. Test the user profile that points the wrong way.
  4. Test the judge that is confident but wrong.
  5. Test what happens when retrieval returns one authoritative-looking but outdated document.

Evidence to preserve

  • Preserve the inputs, versions, configurations, raw outcomes, and results for risk-based sampling, risk based sampling needed to reproduce work on Risk-Based Sampling.
  • Report results for risk-based sampling, risk based sampling by relevant slice, separate blocker failures from averages, state uncertainty and blind spots, and connect the result to a release decision.

Expert note

Expert risk sampling combines likelihood, severity, detectability, reversibility, and exposure. A rare failure with irreversible harm may deserve more testing than a frequent cosmetic issue.

Continue the conversation

Apply this to your context.

Save your product context once, then open a focused conversation that combines it with this concept.

Cite this page

Jason Arbon. "Risk-Based Sampling." Testing AI Knowledge Edition, section 9.

https://jarbon.ai/testing-ai/knowledge/ch009-risk-based-sampling.html

Shared across the Knowledge Edition

Adapt every concept to your world.

Describe your product, role, users, risks, constraints, or current quality problem. This stays in this browser until you choose to send it to ChatGPT.

Saved only in this browser.0 / 2400