Originally posted by nkhoi
If some TS guy can optimize on RSI indicator and post winning/losing % please do. My favorite is 5m MACD diverg signal(didn't see one today) and some basic flag, wedge, H&S pattern.
You provide the rules, and we'll see - at a minimum, you'd need:
1. Entry condition(s) - must be specific: lookback period (or range) for RSI, specific levels (crossover?, if so at what level?), any additional filter (e.g., time of day exclusions).
2. Exit condition(s)
3. Data Resolution (time frame)
4. Tradable(s)
Frequently, when system-testing, you'll just take the simplest conditions (no special filters), a very basic profit-taking method (trailing stop, profit target, etc.) and a simple stop-loss, just to get a basic idea of what you're working with. In order to avoid curve-fitting, some system-traders will refuse to do any optimizing of indicator parameters at all, though others consider that an overly radical position.
I'm always open to looking at stuff, but I gave up on the whole system-trading idea, especially for day-trading. I have yet to see a computer program or system that come close to handling all of the varied "fuzzy" considerations that an experienced trader takes into account, often by reflex. Nor, on the subject of this thread, have I seen a study of chart patterns and other trading tools that comes close to addressing how a good trader REALLY uses them.
The TASC article on head-and-shoulders patterns, mentioned earlier, was typical of how useless such efforts can be, but I don't want to go into the numerous methodological problems with the specific article. I'll just offer my opinion: The question isn't whether the appearance of, e.g., an h&s pattern, as defined by the tester, turns out to tend to predict a certain kind of price movement in tradables across a large sample of identified instances, and according to some simplistic measurement. The question is whether or not the pattern, used intelligently, in conjunction with other indications and filters and good trading techniques, can be
traded profitably. Getting a meaningful answer to this question is a much more complex project than researchers ever seem to acknowledge.
