Interestingly,
“Um zu erklären, wie irreführend Signifikanz sei, benutzt Krämer gern ein Beispiel aus der Finanzwelt: “An Börsentagen, die geteilt durch 7 den Rest 1 ergeben – also am 1., am 8., 15., am 22. und 29. eines Monats –, gibt es signifikant höhere Renditen an den deutschen Aktienmärkten.” Und schiebt direkt nach: “So ein Schwachsinn.”"
- Krämer’s example is one where he ignores the effects of options expiration days and of monthly inflows of pension plan money – which usually flows into the market on the first of the month. There are significant monthly patterns. The weakest day is asround the 20th.
He has indeed found a replicable effect but he doesn’t know the mechanism so he thinks it is an example of statistical arbitrariness.
My trading strategy optimizations have found that effect as well. ChiefIO has confirmed that he knows about this monthly pattern.
