Seasonality
Whether a ticker has a real calendar tendency — a month, quarter, week, or weekday that tends to run better or worse than the rest — or whether an apparent pattern is just noise.
Answers one question: does this ticker tend to do better or worse at a particular point in the calendar, and is that tendency strong or repeated enough across years to act on rather than dismiss as noise.
Weighs each period's average return against how consistently it wins and how wide its spread of outcomes has been, so a good average built on one outlier year reads differently from one that shows up every year.
Changing the calendar basis changes what claim you're testing — a monthly effect and a day-of-week effect rest on very different amounts of history per bucket, and are correspondingly easier or harder to trust.
Tests for a monthly effect — whether one or two months have historically carried the position, or dragged on it, more than the rest of the year.
Tests a coarser, sturdier version of the same question — each bucket rests on more history than a month does, at the cost of resolution.
Tests for a finer calendar effect that a monthly or quarterly view would average away — at the cost of a much thinner sample behind each week.
Tests for an intraweek effect — the sturdiest of the four bases, since every trading day of every year in the window feeds each bucket.
Search and select a ticker in the header to run the breakdown against a different symbol.
Switch between Monthly, Quarterly, Weekly, and Day of Week to change which calendar claim you're testing.
Hover a period to get its underlying numbers, including a note flagging when the sample behind it is thin.
Check whether the current period is called out — that tells you whether what you're looking at is a live setup or purely historical record.