Null hypothesis
What do we assume when a market looks ordinary, and how do we decide that something unusual has actually happened?
In plain words
The null hypothesis is the starting assumption that a pattern is just normal randomness, with no special cause behind it. Statistical tests try to rule out that assumption before anyone claims a result is meaningful. In markets it works the same way: the default reading of a price move is that nothing unusual is happening.
Why it matters
Without a default, every move looks like news and every strategy looks like skill. A statistical test asks how likely the observed result would be if the null hypothesis were true; only when that likelihood is low do analysts treat the move as a genuine signal rather than ordinary variation.
In a real week: ETH, week of 8 December 2025
In the week of 8 December 2025, ETH closed at 3063.5, up 0.12% from 3059.86, with a return percentile of 49.0 against 408 weeks of history, so the week's close-to-close result sits almost exactly at the historical middle. Under the null hypothesis this is an ordinary week, and nothing in the return demands an explanation. The weekly range of 13.99% and a largest one-hour move of -4.95% on Friday 12 December show that the path inside the week was far wider than the flat close suggests.
A common mistake
Treating rejection of the null hypothesis as proof of a specific cause. Ruling out randomness says the move was unusual, not what produced it.