Point estimate and range
When data is noisy, why should a forecast be given as a range instead of one number?
In plain words
A point estimate is a single summary value, such as one week's percentage return. A range states the span of outcomes around or within that value, such as a high-to-low price spread or a band of plausible figures. The point says where you landed; the range says how far the data wandered.
Why it matters
A single number hides uncertainty. When an asset moves in wide, irregular swings, quoting one expected return or one price target understates how wrong that number can be. A range forces the size of the possible error into the open, which matters for sizing positions, setting stops and judging whether a small difference between two numbers means anything at all.
In a real week: BTC, week of 28 May 2018
In the week of 28 May 2018, BTC's point estimate was a 5.11% return, opening at 7338.99 and closing at 7714.26. But the week's range was 10.72%, with a high of 7786.69 on Sunday at 06:00 UTC and a low of 7032.95 on Tuesday at 01:00 UTC, and realised volatility reached 51.2% a year. The single weekly number sits inside a spread more than twice its size.
A common mistake
Reading a net change as the whole story of a period. A summary figure hides the path that produced it, so the swings between the extremes matter as much as the start and end points.