Datasheet · A prompt line
Write tests first
Before you write the function, list a test case for every rule and edge case, then write code that passes all of them.
Hurts on Coding, Haiku 5.5 with thinking off (−2.1 points).
Coding, Haiku 5.5 with thinking off
Pooled across 2 notebooks: −2.1 points, 95% interval −3.5 to −0.6; measured once.
- INHIBITSNotebook #057−2.8 points (−5.6 to −0.3), 432 pairs, cost ×1.90signed verdict 361
- INERTNotebook #058−1.2 points (−4.2 to +1.6), 432 pairs, cost ×1.90signed verdict 367
Where it acts
Its effect sits in AMM swap output (−25.3); no other family moved clearly.
- AMM swap output−25.3−39.9 to −13.379 tasks
The other 10 families, none moved clearly
- Moving averages+7.6−3.2 to +18.479 tasks
- Business days+0.0+0.0 to +0.078 tasks
- Fees in whole cents+0.0+0.0 to +0.078 tasks
- Merging intervals+0.0+0.0 to +0.078 tasks
- Run-length encoding+0.0+0.0 to +0.079 tasks
- Comparing versions+0.0+0.0 to +0.079 tasks
- URL slugs+0.0+0.0 to +0.079 tasks
- Top k by score+0.0+0.0 to +0.079 tasks
- Parsing durations−1.9−7.7 to +3.978 tasks
- Trade profit−1.9−12.2 to +7.078 tasks
Read from 864 paired tasks in 2 notebooks, each task against its own notebook’s two controls, with every interval widened for the 11 families read. No notebook preregisters a family: this says where to look, and a notebook says whether it holds. The runs behind it are on each notebook’s page.
Put this datasheet in a README
The badge shows what the signed record says about this part, in the setting measured most, and links here. It follows the record as new notebooks publish. Add ?suite= and a suite’s id to show another setting.
[](https://opsinlab.io/parts/2d8d463b81907e26)Measure it in a setting it has not been measured in: propose it. As JSON for an agent: /v1/parts/2d8d463b81907e26. Every part: the atlas.