Quote from dazzwater:
You seem to be saying that all simple strategies won't work because of the if-it's-so-simple-then-why-hasn't-anyone-thought-of-it-yet principle (There's probably a real name for this principle but anyway)
I believe simple strategies can work, and also are preferable so as to minimize the risk that your "complex" strategy is simply a result of over-fitting to historical data. But there should be a basis for your choice of price direction. In keeptradin's case, the strategy always goes long. Why?
Sorry if I gave that impression; I definitely believe that simple strategies can work (I've traded them and totally agree anything with a low parameters/samples ratio is preferable). But I doubt blatantly obvious ones will yield much of an edge, especially a statistically significant one.
And, as you said, there should be in underlying basis for the market behavior you're trying to exploit. I just don't see it in something that simply buys an index at the open each day and sells at the close; That would imply that the market on average has a positive intraday drift, closing north of the open each day. My past backtesting shows this isn't the case; Over the long run, such a system results in a zero expectation (sans commish), which is exactly what one would expect. I'd be surprise if SPY and ES were so inefficient.
But heck, if the OP has done some backtesting that shows otherwise, he should post some details and humble all of us skeptics
