Do you really expect a computer program capable of generating successful trading strategies to be for sale?
I asked about it in previous threads. Did not get an enthusiastic response. My opinion is simple and robust strategies that can work across time frames should work as well as any hand designed ta system. However, ta systems are somewhat discredited nowadays. If we look back at john ehlers work, we see that ta is simply early-quant work.
So far the systems it created were no better than placing Enterlong() at open and sell at close but we will see doesnt hurt to try.
As one ET wit noted recently, you'll find successful trading strategies for sale in the shop right next to the one selling geese that lay golden eggs.
The best systems for forex will have a sharpe of around .20 and a oos dd of about 30%. For equities index, you can do much better with .30 sharpe and oos dd of about 25%. However, stick with symmetrical strategies that have low complexity initially to minimize curve fitting. I would like to hear some positives from the experienced ET folks, too.
Hey guys listen to this, I created an automated strategy that had about 100 parameters utilizing many levels of nested indicators recomputing ES derivatives many millions of times over. The in-sample and out-of-sample results looked stellar with no losses but when i ran it on another out of sample period it came back break even.
Read this. This is an academic paper than shows how the results of multiple tests must be adjusted to account for data-mining bias. In essence the data-mining bias is too large.