One of the hardest parts of a trend exit is keeping it close enough to protect profits while letting the trend work. The typical parabolic exit is nice for protection, but it fails the latter stipulation for a good exit. One exit I think works well is a derivation of the parabolic that has a cyclical component. The cycle needs to be matched to the market it is used in for best results. The idea came to me while looking at some non-trading material with the logistic growth curve. It dawned on me that this concept (though changed in major way) needs to be applied to typical trend exits. Without getting too detailed about specifics, try using a cyclical coefficient to the main trailing process in the parabolic (highest high/low in trade - current price), while increasing the lower limit of the cycle with time. What results is a stop that trails slowly, but tightens with market retracements and time. There are many ways one could do this. Get creative and see how it works for you.
To know when to exit (or to enter because the problem is the same) you need to measure the strength of the trend. Trends can be very strong, less strong , weak ,very weak, non-existingâ¦.. For each kind of strength the size of the move is different. So it is evident that you have to diversify your entry- or exit points according to the strength of the trend. You cannot use 1 system for all kinds of trends, because the system will be optimal for some kind of strengths but lousy for other kinds of strengths. I will give a very stupid but clear example: if the trend is weak you can go short (or exit a long) in almost any overbought area from any indicator. But if you do this when the trend is very strong you will be way to soon. You will have to adapt your exit to the very strong trend. So to me it is not logic to talk about 1 system, the fact that the indicator might be self-adapting to the move of the trend is not sufficient. A parabolic indicator is self-adapting but not very efficient if you want to maximize your profit; heâs too slow. Therefore I made different entry and exit methods depending on the strength of the trend. For each type of strength I have build a scenario that tries to predict the evolution of the complete trade from entering to exiting. Each scenario is adapted to the strength is was designed for. I donât use the same indicators in each scenario; they vary depending on the scenario because some indicators are better for strong trends and some are better for weak trends. I know it may sound a little complicated, but good systems are always complicated; because if it could be done in a simple way every one would be an excellent trader. This is how I see it, but I donât pretend to have found the holy grail or to be an expert. I just invest a lot of time in studying and experimenting. Trading is a job, not a hobby.
interesting approach spike... using multiple strategies together. Question is how to define such type of trends. I am unable to do that as of yet...
How do you measure the strength of the trend before it has occured? I would like to know the current strength of trend in the following markets: US Bonds, US 5yrs, Bobl, NY Coffee, and Japanese Gov't Bonds (10yr, JGBs). I am eagerly awaiting your response. Thanks 5yr If you don't want to give out this info, you could just PM, please use the qualifiers you used below and the direction of the trend: very strong, less strong , weak ,very weak, non-existing and long or short Oh and I think most trading systems that work are amazingly simple, but to each his own.
Iâm daytrader in the ES. I cannot give the details of how I calculate the strength because I donât want to disclose my system, but I will try to give all the information I can give. Perhaps it can give you an idea about the logic behind it. I define the actual trend and its actual strength on hourly charts. So I donât measure the trend before it has occurred. There are 8 different degrees of strength. For each degree I have a separate scenario that has a high probability to happen. The strenght and the trend are recalculated constantly, so they are dynamical, not statical, and can change during the day. The probability that the scenario fulfills itself is high and even if there is a change to another scenario losses are limited. The best way to show how it works is to give an example with the chart added. The chart will be a 5 minute chart. At the opening the trend is short and the strength is moderate. The corresponding scenario says to go short and stay short until point 2 at least (Chicago time/ exchange time). This is the start of the second scenario. Here we have to go long because the downwards trend should make a bottom and reverse. At point 3 we have to close the long because the strength is weakening and the reverse of the trend did not happen. Because the strength weakens the reverse will not happen. The system indicates a downwards correction. Our position is flat now. End of the second scenario. We cannot take any position until the system gives a clear signal; the system is confused so we stay aside. At point 4 we get a signal. Start of the third scenario. The trend is clearly downwards again and the strength is moderate and strengthening. We go short but we have to be careful because of the weakness of the trend. At point 5 it is clear that we should stay short ; the strength of the trend has become strong, this means normally stay in till the close of the day. At point 5 we went from the third scenario into the forth scenario. The forth scenario is a continuation of the third but in a more trending form. It confirms that the move will be much bigger than expected at the start of the third scenario. Thatâs the advantage of a dynamical system; it adapts itself to the ever changing trend. Between point 3 and 4 we couldnât trade due to the confusion in the system; you can see that there was no direction in the market and we did more or less ârangetradeâ. The direction was clear again before the big move down happened. You said: âOh and I think most trading systems that work are amazingly simpleâ This is correct till a certain degree. But if we talk about returns of over 100% a month your statement isnât valid anymore. And if it would indeed be amazingly simple almost everybody would make money, which isnât the case in reality.
Sounds good Spike. I find it hard to believe any trend indicator can not be lagging and cause whipsaws though. I use coincident indicators, like trendlines, fibs, gann, stochastics.
... Good exit = Perfect entry. Perfect exit = Holy Grail. Think in matter of choice. But... considering that you clearly understanding the underlying edge you're trading under, exits are effortless. More clarity = Better exits. For example, let's say you have a Moving Average system. There's an issue on which parameter to use for the BarsBack. Without any optimization, would you know which # to use? Can you quantify and know which and why the specific parameter is optimal for that period???