Speaking of missed facts...
The reason your observation is useless for trading is that your window is wider than the median run length for upside moves. A *lot* wider. This leads to the (surprising to some) observation that any day of the year - chosen by any means you like, including randomly - has an 80% chance of being within a 1/2 month window of a short-term top. In other words, the reason you observe what you observe on days in early January is because the same thing happens on every single day of the year.
You've got a long row of "facts" to hoe, mate.
PS. At some point you're going to have to deal with the "complexity" of normalizing data: 5% isn't a correction when markets are trading at 40% stat-vol, its noise. Until you learn how (and what) to normalize you're going to keep getting these screwy analyses.
Market Data 
Markets 
AI
