Day-of-week patterns that are easy to overstate

[Amanhasnoname]

Every forum has a claim: Monday is slow, Wednesday is the trend day, Friday reverses. Some of it has a basis in data, much of it does not.

Real effects exist. Monday often starts with lower volume as the week finds its direction. Tuesday through Thursday carry most of the weekly range in many samples. Friday afternoon tends to thin out as positions are squared before the weekend.

But the effects shift year to year, and they depend on the pair. Claims that hold in one sample often fail in the next.

A better practice is testing the effect on specific pairs over a few years, using ranges and the strategy's own results. If a particular weekday consistently underperforms, trading less that day is reasonable. If the pattern is weak or unstable, it is better to ignore it.

Be careful of small samples. Twelve Mondays say very little.

Also keep in mind that holidays, central bank meetings, and data schedules distort the weekday picture in any given week.

Does the journal show a weekday that consistently underperforms?