A plant is 2 cm tall after week 1, 4 cm after week 2 and 6 cm after week 3. It has grown 2 cm each week, so your best estimate for week 4 is 8 cm. That is a prediction: a value you have not measured, worked out from a pattern in values you have. The honest version is '8 cm, give or take about 1 cm', because a cool week or a dry pot could slow the plant down.
Two things make predictions less certain. Messy data is one: if the plant had grown 1, 3 and 2 cm in three weeks, the pattern is less clear and the range should be wider. Distance from the data is the other. Predicting week 4 from weeks 1 to 3 is a short step, but the same rule for week 52 gives 2×52=104 cm, and a small potted plant is unlikely to keep growing steadily for a whole year. Patterns hold where you have evidence and can break beyond it.
Not every number about data is a prediction. The plant's average growth so far, 2 cm a week, describes weeks you already measured; it is a summary. A strong correlation is not a prediction either: knowing that study time and test scores rise together does not tell you exactly what one student will score, only roughly what to expect.
Think of it like this. Making a prediction is like throwing a ball to a friend who is running. You aim not where they are but where their speed says they will be. A short throw usually lands close; a long throw gives them time to slow down or turn, so you are less sure where they will be. A good prediction comes with that sense of how sure you are.
Core idea. A prediction is an estimate of an unknown value that always carries how sure you are about it.