Concepts
This index lists canonical reference concepts for analytical reliability. Use these definitions to classify failures that look operationally "healthy" but produce incorrect, misleading, or untrusted analytical outcomes.
Start here
Domains
-
Data Movement Reliability
How data is ingested, transformed, and delivered as intended.
-
Semantic Reliability
Whether analytical meaning is preserved in models, metrics, and relationships.
-
Execution Reliability
How analytical workloads behave at runtime, including performance and drift.
-
Change Reliability
Whether code/config/model changes introduce unintended analytical failures.
Base Concepts
How to use these concepts
- Classify an incident by domain first, then identify the base concept that best matches the failure mode.
- Link incident writeups to the relevant concept pages to keep language consistent over time.
- Treat concept pages as stable references; update definitions carefully and consistently.
- Use the domains map to explain why "green" systems can still produce wrong outcomes.
- Add specialized pages on other sites by referencing these base concepts rather than duplicating them.