Glossary
Trend Analysis
Updated February 28, 2026
Trend analysis smooths noisy data so changes are based on patterns, not single days. A 7-day rolling average removes roughly 80% of water and sodium noise from scale weight.
Rules
Minimum trend windows
| Metric | Minimum window | Minimum confidence rule |
|---|---|---|
| Scale weight | 10 to 14 days | Two-week pattern before changing calories |
| Waist or photos | 2 to 4 week comparisons | Ignore one-week cosmetic noise |
| Protein average | Weekly average x3 weeks | Adjust only after adherence confirms |
| Steps and active volume | 10 to 14 day trend | Align carb changes after output trend stabilizes |
Common noise sources and expected magnitudes
| Event | Expected weight shift | Duration | Action |
|---|---|---|---|
| High sodium meal | +0.5–2 kg | 24–72 hours | No macro changes unless trend continues |
| Menstrual cycle | +1–3 kg | 3–7 days | Compare same cycle phase month-over-month |
| Alcohol intake | +0.5–1.5 kg rebound | 24–48 hours | Treat as event, not trend |
| Creatine loading | +1–2 kg | First 1–2 weeks | Expect water retention. Track from post-load baseline |
| One-week low adherence | Variable | Logging dependent | Hold changes and run a clean week |
Plateau and rebound protocol
| Pattern | Interpretation | Next move |
|---|---|---|
| Plateau with high adherence and good sleep | Drift likely from undercounted intake | Re-evaluate portions and liquid calories first |
| Plateau with reduced adherence | Behavioral drift likely | Simplify plan before changing targets |
| Rebound after correction | Likely delayed physiology | Keep new target for 1 to 2 additional cycles |
| Plateau + fatigue and poor sleep | Recovery bottleneck | Prioritize recovery deload, delay aggressive changes |
Action plan with delay
- Hold macro changes until two trend windows agree
- Change only one macro family at a time
- Recheck 7 to 14 days after each adjustment
- Revert change if performance and satiety both deteriorate
This reduces over-correction and keeps plans stable.