Several sources (online and offline) that discuss the delta of a listed vanilla option, state that its delta is a (guesstimate?) of the probability of said option expiring ITM (in the BSM framework).
However, looking at the derivation of delta from the BS model (and its variants), it is not obvious (atleast to me), that the delta can be used as a proxy for the probability of ITM expiry. I want to know if there is any supporting evidence (theoretical or otherwise), that lends at least some credence to this assertion - or is it just an "old wives tale" ?