Writing template. This is a guide for a future experiment log. It does not report an experiment or measured result.
Question and expectation
State the question and explain why the approach seemed plausible. Write the expected observation before introducing the actual outcome.
Describe the conditions under which the hypothesis would be supported, weakened, or left unresolved.
Setup and evidence
Record the task, baseline, controlled variables, evaluation procedure, and the code revision. Separate what was measured from what you think it means.
| Part of the note | What to record |
|---|---|
| Hypothesis | The claim being tested |
| Intervention | What changed from the baseline |
| Observation | Measurements and relevant artifacts |
| Interpretation | Possible explanations and uncertainty |
| Next step | A discriminating follow-up experiment |
Include traces or qualitative examples when they help explain what the summary metric misses.
What the failure does and does not establish
An unsuccessful run does not by itself explain the failure. Distinguish a problem with the idea from a problem with the implementation, evaluation, or experimental setup.
Keep alternative explanations visible. If they haven’t been tested, describe them as possibilities.
The next useful question
End with the smallest follow-up that would resolve an uncertainty. A negative result becomes more useful when the reader can see exactly what it changes about the next decision.
The notebook continues.
Explore more notes →