Rank test
A test that compares groups using only the order of the values, not the values themselves, so a few extreme values can't swamp it. The Mann-Whitney U test is the rank test for two independent groups.
In plain English
Pool all the values, rank them, and see whether one group's ranks are bunched at the top or bottom more than chance allows. It suits small, skewed samples and ratings. It answers a slightly different question from the t-test: whether one group's values tend to beat the other's, not whether the averages differ.
Uncle Gary explains
Seven pizza deliveries from each of two places, one a 95-minute disaster: the t-test gives p = 0.86, while the rank test sees one place quicker in 42 of 49 pairs, p = 0.026.
Why it matters for your money
Trading results are dominated by a few big days. A rank-based comparison shows whether a strategy is usually better, separately from whether a handful of outliers made its average look good.
Where we tested it
Related terms
All glossary terms. Educational content, not financial advice.