Tell whether a change in a weekly number is real or noise

Tell me whether the recent change in {{metric}} is a real shift or ordinary week-to-week variation, for {{audience}}, so that we react to signals and stop explaining noise. Use a process behavior chart (an XmR chart, also called an individuals chart) built from the weeks before the change.

Weekly values, oldest first, with the week each covers:
"""
{{weekly_values}}
"""

The week or weeks in question: {{weeks_in_question}}
Known events that could affect the number (holidays, launches, price changes, tracking or definition changes), with dates: {{known_events}}
Spreadsheet app, for the formulas: {{app}}

Method:
- Baseline: at least 12 weeks before the weeks in question, never including the weeks in question. Leave out weeks hit by a known event only if I listed it, and say which you left out.
- Center line: the average of the baseline. Moving ranges: the absolute difference between each baseline week and the week before it. Average moving range: their average.
- Limits: the center line plus and minus 2.66 times the average moving range.
- Signals: a week outside the limits; three of four weeks in a row closer to the same limit than to the center line; or eight weeks in a row on the same side of the center line.

Structure the answer:
1. The verdict in one or two plain sentences: signal or routine variation, and which signal rule it meets, if any.
2. The calculation with the numbers: baseline weeks used, center line, average moving range, upper and lower limits.
3. A table with the columns Week, Value, Moving range, Above or below center, Signal.
4. The {{app}} formulas to recompute the limits next week, with a helper column for the moving ranges if the app needs one.
5. If there is a signal, what to check next, up to 4 items, such as whether a definition changed.

Keep the explanation under 300 words, outside the table and formulas. A good verdict is one a manager can repeat accurately in one sentence.

Do not call a change real because it is large in percent, and do not explain what caused it; the chart only says whether something changed. Do not fill in missing weeks: if a week's value is missing or unclear, skip the moving ranges on either side of it and say so.

If there are fewer than 12 baseline weeks, still calculate, but say the limits are provisional. If the number has a strong seasonal pattern or trend, such as holiday weeks, say that these limits may mislead and suggest comparing with the same weeks last year. If a known event changed how the number is measured, ask whether to start the baseline after it.