Conditional systems with multiple indicators?

Discussion in 'Technical Analysis' started by IronFist, May 15, 2007.

  1. Most introductions I've seen to TIs say "when the indicator does x, buy. When the indicator does y, sell." We all know that doesn't work profitably over time or everyone would only use one indicator and it would be called the Holy Grail indicator.

    I've also seen websites that say to look for confirming indicators. They say like "when indicators x, y, and z give a buy signal, buy." That sounds like a better idea, however, it also seems like it might ruin your timing sometimes. For example, "cool, indicator x just gave a buy signal... now I just have to wait for two more... sweet, there's y... one more..." and then as soon as z gives a buy signal, the moment has passed.

    I've never seen anything about using a conditional system, however (altho I'm sure something exists because I doubt I'm the first person to think of it). So something like "if indicator x gives a buy signal, and these other conditions with other indicators are met (not necessarily buy signals), then buy." So you'd use one indicator that seems to give good signals for your main indicator, like MACD. But as stated in the first paragraph, one indicator is never consistently profitable by itself, so you assign conditions to it. "Buy when MACD turns positive if these other conditions are met. If MACD turns positive and the other conditions are not met, do not buy." So it's like a filter for MACD (or whatever main indicator you are using).

  2. Technical analysis helps you analyze the price action to determine trends and market bias so you can make a more informed decision as to long or short and price entries and exits.

    In my opinion, they are simply analytical tools and the trader/invesotr still has to make the decision and that key aspect is what makes most automated systems or pure conditional TA usage lead to losses.

    Trendlines, volume, MAs, etc.. all are methods of interpreting what is happening with the price so you can make a better decision with good risk management. But I dont think they were designed to set and forget without human analysis.
  3. MGJ


    It's a pretty old idea. If(Condition1 is met) AND (Condition2 is met) then take the trade. Otherwise, don't take the trade. Sometimes people call Condition1 the "setup" and Condition2 the "trigger". You can find it in Babcock's book (1989), LeBeau's book (1992), and Krutsinger's book (1994).

    A recent example is the Donchian Trend system on page 68 of the new book Way of the Turtle. You don't enter unless a very longterm pair of moving averages indicates that the very longterm trend is up (Condition1), AND there is also a breakout above the highest high price in the intermediate term (Condition2).
  4. I figured it wasn't new.

    Intuitively I think this would be more beneficial than a strictly TI based system.

    I need some backtesting software and to learn how to code :D
  5. Worden Bros has some nice drag and drop stuff that would fit into what you are trying to learn about.

    The specific blocks you would be enhancing are called AND, THEN functions.

    I ran a couple through their block player at the end of one of their presentations.

    The software is fun to use and you can increase the annual output of a system by a couple hundred % as you improve it.

    Unfortunately the system only does EOD and will not handle intraday math.

    The best choice of what to use is leading indicators of your trades.
  6. I should look into that. I have some programming experience but I don't know EasyLanguage (or whatever script everyone here uses) nor do I know what program is goes with (Tradestation?). And I can't program my own stuff from scratch either.

    I was actually gonna make a thread asking about this. If I had some backtesting software and knew the script I could keep myself occupied for days :)
  7. You can get the coding "behind" the drag and drop stuff by clicking an icon.

    What you see is like pascal ++ more or less.

    There are several advances coming up for worden in the future (as usual).

    One is FA info. This means that at long last worden will be able to make lists of stocks that relate to making money.

    Another is intraday so a person can actually use the block player to make money.

    And the third is the sector analysis stuff. They have some of that done already. you import a fee based data system that operates on 238 sectors and subsectors.

    They intro'd some new indicators (new yourk expo) and if you change the signal points, then you get some sharp results.

    I usually make it a point to attend their presentations to keep track of how close they are to getting a replacement for the old TC2000 ver 3.0. On that I used the 7 equations I created for the scoring alternative. I used a story board to place lists and do an "add delete" on the lists analysis. Can't do that anymore though. The math capability is no longer there.

    it would be a miracle if some platform outfit got some stuff done to make it possible for people to really make money. It is still a multiplatform requirement to clean the clock on the markets.

    the day is coming where the man in the street will be able to make several hundred ercent a year doing EOD trading by spending 15 to 20 minutes in the evening to keep his trading on the ball. why the platform people and the brokers can't learn this, I'll never know.
  8. CWU


    As mentioned, for EOD, TC2000 will do all that you desire, and the coding is simple. Additionally, you can backtest it with the same basic language in TC Companion.

    It helped me to create a number of good profitable setups.