Doing intra-day dollar-cost averaging is bad.
When stocks tank, they usually tank for a few days. Hence the same tactic for spotting intra-day lows should be applied to a more "macro" picture of using days. Doing that over "days" allows time for the good/bad news to be widely reported, caught on by other investors, acted upon by everyone. Perhaps a rough estimate of 4 days is necessary for a stock price to fully adjust for a piece of news.