History shows there is a propensity for the market to rally after a full moon. http://www.moonconnection.com/moon_trading.phtml
This claim could be easily refuted. He is saying that reversals take place in conjunction with the moon cycle -- whether they are bottoms or tops. Since the moon cycle is 29.5 days, and this guy says that the reversals occur 'around' the moon phase date (i.e. -2-3 to +2-3 days by his claim), we get a 5-7 day window in which a reversal should occur. Great -- so on a 29.5 day cycle, I know that a pivot will occur within a specific 5-7 days. Since he never makes any claim to the size of the pivots, his entire claim boils down to: 'quite frequently, reversals occur within this 5-7 day window'. Notice on the chart he posts, there is a 'pivot' nearly every 5-7 days anyway, so it doesn't really matter where we start counting from -- he just happened to choose moon cycles. Since this window of 5-7 days fits nicely into 30 (you would expect 4-6 cycles of pivots every 30 days), we would expect one of the pivot windows to fall near the end of the moon cycle. On his chart, I count 17 bars that could arguably be considered 'pivot' points (by any swindler trying to make a point, at least). On a total of 51 trading days, that is a pivot point nearly every 3 days. Including weekends, we can probably expect a pivot point every 4-5 days. However, the magnitude of this pivot point is completely unpredictable, making profit and stop targets near impossible. I don't feel like whipping up some software to test the statistical validity of this claim, but the smell test tells me it ranks very high on the 'spurious correlation' scale. Without any claim as to how large the pivot should be, a 5-7 day window doesn't give us enough information to trade off of.
Well, if you're going to be logical about it, then just forget it! Might market activity be influenced by sunspots/solar flares?
I think it can also be top. Anyway, this time the change did come right at full Moon. http://www.elitetrader.com/vb/showthread.php?threadid=22943