More likely there are cyclical effects that compounded together have a phase similar to the moon cycle. These kinds of effects are actually quite common if you

More likely there are cyclical effects that compounded together have a phase similar to the moon cycle. These kinds of effects are actually quite common if you have the data and computing power to mine for them. So there is likely an indirect causal element at play here. It's just not the moon. The problem is that this is sometimes wrong in a big way. Once you account for commissions, the need for risk controls (few investors will tolerate 20%+ drawdowns), and the need to realize returns intermittently I doubt this will outperform the S&P by more than a little bit. Even if there is some statistical edge here, it would require massive sums of money which would probably hurt returns due to market distortion effects. To truly take advantage of something like this you have to also use plenty of other signals and understand their relationships to develop a strategy that scales even if your signal loses some predictive power.