Actually, there is a fair number of reliable statistical studies you can do to detect short-date opportunities (not really intra-day, but e.g. pre-expiration or pre-event). The key difference is not to get hung up on a small subset of names, but scan across a large universe of stocks.