Confidence interval
How do analysts show how unsure they are about a number they have just estimated?
In plain words
A confidence interval is a range built around an estimate to show how much that number could reasonably move if the sample changed. A narrow range means the estimate is precise; a wide one means many values remain plausible. The range depends on how much the data vary and how many observations back the estimate.
Why it matters
Risk is taken on the estimate, not on the truth, and a range makes that uncertainty visible before a position is sized. Two markets with the same central estimate can carry very different risk if one rests on few, volatile observations and therefore has a much wider plausible range.
In a real week: BTC, week of 16 April 2018
For the week starting 16 April 2018, BTC's realised volatility was 62.3% a year, its weekly range was 15.46%, and the largest single-hour move was 3.78%. The week closed 5.17% above its open. Any range drawn for that week rested on only 9 weeks of history, so the interval itself stayed wide.
A common mistake
Reading a confidence interval as a guarantee that the outcome will land inside it. The range describes uncertainty in the estimate under stated assumptions; it does not cap how far prices can actually travel.