This is a ridiculous "statistic" like "if January is down,whole year is probably down".
Supposing that the variable we are observing is completely random, it's obvious that if any initial period is down (up), the whole period has higher probabilities of being down (up).
This comes from the...