Larry,
Thanks for crunching the numbers. There are a few glitches, discussed below.
However, my thesis is that any system will fail to predict the market and that no one will beat it in the long run. For the three year period your model uses, there are any number of systems that will show a profit including the moving average system I sent you. But, over a 20 year period no system will beat the market. It is a perfect fractal.
Over a shorter period a system might show a back test success, but I think it is the equivalent to the guy in a stadium full of coin flippers who flips heads thirty times in a row and claims its because he was standing on one foot. It is unlikely to beat the market over the next 10 years.
The market should react to money flow/supply/demand. I'ld like to think there is some way to predict money flows and thus my questions about your Hal or the BCompq systems, but my gut tells me we are searching in vain. I'm just too bull headed to give up. Or maybe there is a Harry Sheldon out there.
Maybe you are on to something based on years of experience and expertise in statistics. If so I do not blame you for wanting to keep it a black box. I appreciate you sharing your market direction calls, but its hard for me to put faith in a complete black box. I would appreciate a little more information on why you think your Hal system works if you don't think it will compromise your success.
Anyway, regarding the moving average system you analized:
Other than the first two lines, the summary is not ledgible.
There appears to be a minor glitch in the sheets. You should be short on April 4 which adds about $400 profit.
The methodology is to sell or buy QQQ after hours on the second day. Your sheets appear to execute at the third day price. For example you should sell your long position on July 27,2000 for 9187 not July 28 for 8687. Due to this, my results are significantly better than your calculations.
Thanks again. Cy Esp
AI
Market Data 
Markets 
