After many weeks of backtesting the systems that came with MS11 and then buying the Performance Plus plug-in and testing those systems, I am amazed to find that none of these systems seem to perform as well as the very basic logic of buying on, say, a 30-day high and then placing, say, a 20% stop loss. Is that possible, or must I be doing something completely wrong in my backtesting?

At first I tested on major market indexes like SPX, RUT, etc. Then someone suggested that technical systems do better on more volatile vehicles, so I moved my testing to the 100 stocks in the Nasdaq 100. Still the same. The included systems in MS11 & the PS Plus group seem to generate many more trades and net a lower return. Is that possible, or am I right to question these results? Does anyone out there have system logic they would suggest I compare to the very basic going long on H>Ref(HHV(H,30),-1) mentioned above?

Looking for insight.

/kim

      You are probably on the right track. Most signals won't make money right out of the
  box. If you study  some of the indicators you will find that each works best in specific
 markets (bull/bear/sideways) then fail badly in the others. So the trick becomes to
 isolate those positive signals while ditching the others, skewing the probably of a successful
 trade in your favor.

 

Thank you for the feedback, John. I have heard people speak of certain systems working during certain market conditions. That said, it seems as if the true power of a system like MS is in its ability to make non-emotional, objective entry and exit decisions. It further seems to follow that the only way to trust and follow a system completely is to include all of one's criteria into the model and then thoroughly backtest that "full" model. If one "overrides" the system with additional overlay criteria it would seem to invalidate all backtesting and certainly aggravate the already very difficult marching orders of all great traders to "follow your system".

Does one add an "OR" to the Buy signal tab that allows your system to know what type of market is prevailing and therefore isolating any final buy signals until that type of market exists? It would seem that this would be the only way to objectively define such a filter, test it and then know without any emotion when those conditions are deemed in place at a given moment in time.

Are there indicators one can use to formally define this filter and subsequently test it?

/kim

kardenmiller:

If one "overrides" the system with additional overlay criteria it would seem to invalidate all backtesting and certainly aggravate the already very difficult marching orders of all great traders to "follow your system".

There are many aspects to developing a trading system. To expand on your quoted point, you must have some assurance that you a going to stay with a system. Quoting one of my recent blog posts, "Imagine that you’re using a system in live trading and the last five trades have been losers. Your trading capital is depleted and your confidence shaken. Are you going to make the next trade without question?
"Consider the case of a huge drawdown during an open trade. Your profit has turned into a loss. Will you stay with the current rules or jump ship? Can you hold on for a vague promise of a profitable outcome in a few days or weeks? That’s unlikely."

---

Charles Brockman
President, Security Genetics Inc.