Recognises precision lost in migration
A numeric conversion that rounds or truncates rates or amounts.
Can AI do this?
AI does this partly Some of the tested AI models master this; others do not show it. 2 of 4 models in the current benchmark master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 91% | Mastered |
| gemini-3.5-flash-lite | Low cost | 9% | Not shown |
| gpt-6-luna | Low cost | 90% | Not shown |
| claude-haiku-5-5 | Low cost | 90% | Mastered |
Mastered means a probability of at least 90% that the model has this knobit, estimated from its graded answers. “Not shown” means the answers do not settle it either way. How the models are tested.
What it looks like in practice
Expected result: The defect named on the conversion rounding rates or amounts, and its fix or the question for its owner
How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: MM-002 interest rate to NUMBER(5,2): 4.125 % becomes 4.12 % Acceptable: NUMBER(7,4), or a rounding rule agreed with the owner
Also known as: Precision lost
Related knobits
Builds on
All 15 knobits of Plan and prove a data migration.
Source: Review checklist 05-data-migration, row 8 (DMBOK; 4217)