Thanks for more clarity. I understand your reluctance to give away the fundamentals of your system.
A friend of mine, Joe Rios, founder of Rios Quantitative, developed an automated trading robot he named Einstein. It's quite effective when utilized optimally. Einstein is activated by the trader making a discretionary decision, and it's premise is based on bull and bear trading zones. Once activated, it takes it from there in determining the entry and exit based on your pre-set parameters (time frame to engage a trade/profit objective/stop/quantity). One version (presently for Trade Station only) even allows a trailing stop capability.
Some general things Joe found in his development of Einstein:
1) Successful equity curves varied for different instruments depending on the time frames traded for each specific instrument (i.e. Generally you would not to trade Crude or Nat Gas during the Tokyo session, after 2:30 PM EST)
2) The best profit target odds put the profit target objective at $190-$200/contract, with a $300-$400 stop depending on the instrument traded.
3) A risk-on environment strongly favored an instrument like Crude long entries since it is a risk commodity, and conversely with a risk-off environment (BTW, NG is what I consider an independent commodity, so it is not influenced by risk on/risk-off scenarios).
Anyway good luck.
The following user says Thank You to TraderJim618 for this post:
Unfortunately this paper pretty much concludes there's no way to come up with a programmatic trading system, at least not using any historical data. And using future data is impossible. So lets all just give up ?
When I think that I've found the automated "Holy Grail" , I simply run it real-time, on my simulated IB account for a few months..(using IQfeed for data), That will show very quickly, if it really works or not...