LECTURE HUB
Lecture 16 - Forecasting and its intervals
🚦 Status🌟 Mastered
🧠 Mastery🦅 Adept
📅 Last Review7 Dec 2025
🎯KEY TAKEAWAYS
- A confidence interval is about the average. A prediction interval is about the next single case. The customer is always asking for the second.
- No amount of data makes one vessel, one batch or one delivery predictable beyond its own scatter.
- Forecasts get worse at the edges of what you have seen. Say so on the chart.
- This is 💎 An interval is more honest than a point with the working shown.
⚡QUICK JOTS
- a forecast is a prediction plus how wrong it usually is. without the second half it is a wish
- TWO intervals and I have been mixing them up all year
- confidence interval: where the average is
- prediction interval: where the next one will land. always wider
- more data narrows the first. it does not narrow the second below the noise in the thing itself
- further from the middle of your data → both get wider. the line pivots
- finished cleaning the port data this afternoon, four days late. this lecture is about exactly what I do next
☑️HOMEWORK
- Put a prediction interval on every forecast in 📌 Statistics II - Project - What predicts vessel delay at Greywater
- Add the two-interval distinction to 💎 An interval is more honest than a point
- Check which of the two our delivery estimates at work actually are
- Redraw the trumpet chart as a card for Statistics flashcard deck