What the calendar actually does, and how much of it survives a test.
Day of the week, place in the month, expiry week, quarter end, payrolls Friday and Fed day, over ten years of sessions. Every seasonality page slices the calendar forty ways and prints the best slice. This one prints the best slice and then tells you whether shuffling the labels produces one just as good, which it usually does.
Loading…
The verdict first
Loading…
The largest difference across every cell is compared with the largest difference that shuffled labels produce, several hundred times over. Testing the winner instead of testing them all is how a seasonality table fools people, so we test them all.
Days on the calendar
Payrolls Friday, Fed day, monthly expiry, quarter end. Only days whose date follows a rule, because a study built on guessed dates is a study of the wrong days. Read the size column, not the direction one.
Loading…
The calendar, sliced
Loading…
Everything here is descriptive market-structure context. Not investment advice, not a signal, no performance promise. Past structure does not predict future prices. Trading involves substantial risk of loss.