ZoneStrategy based on the 3D Zone Statistics
Posted: Fri Jun 09, 2006 12:14 am
As promised, I've written a sample testing trading system based on the Zone statistics concept. The original article which describes the basic idea you can find here in the statistics forum:
http://kreslik.com/forums/viewtopic.php?t=41
This is only one of the many possible approaches to capitalize on the ZoneStats concept. The presented trading system is in no way a final, ready-to-trade strategy. And yet, it shows a profit potential accross a 3.5+ years long sample period including the last 30% out-of-sample unseen data. That means the input parameters selected were not curve-fitted and the basic model is sound.
Given the fact that there is no money management attached to the system whatsoever, the results are quite nice. You should keep in mind that this trading system torso is just a preliminary raw test of how the Zone statistics works in the market. The only thing the system does is opening a position at the close of a bar and closing the position strictly at the close of the following bar. Nothing more.
To recap, the idea behind the Zone Strat trading system is quite simple: trade in the direction of the proabability. If the statistics shows a tendency to stay in the current Zone, open a position in this direction. If the statistics shows there is a tendency to abandon the Zone, then open the position the other way round.
I have only added a little twist to the code which ensures we are not going to open a position if there is too small a profit potential despite the fact that we would trade with the probabilities.
Example of such a case:
according to the number of consecutive bars in the Zone and the Zone percentage, there may be a high probability that the next bar will close in the same Zone as did the current bar. But if the current bar in fact only showed a 5 points range, then the minimum distance from the current close to the "stay in the Zone" price level may not be worth trading after the trading costs are subtracted or even the distance may be negative. The reverse applies to the minimum distance to abandon the current Zone (see the code).
So, incorporated into code there is a measure of what is the minimum palatable distance between the current close and the price level at which a Zone change/Zone abandonment occurs to open a trade.
The best way to come to grips with this test system is to apply it to your chart by yourself and fiddle with the settings.
Attached:
- equity curve
- trading system code
You are free to experiment with the code by yourself and post your results here!
Good luck!
Michal
If you can't see the attached image and ELD file, please log in. You must first register to be able to log in.
http://kreslik.com/forums/viewtopic.php?t=41
This is only one of the many possible approaches to capitalize on the ZoneStats concept. The presented trading system is in no way a final, ready-to-trade strategy. And yet, it shows a profit potential accross a 3.5+ years long sample period including the last 30% out-of-sample unseen data. That means the input parameters selected were not curve-fitted and the basic model is sound.
Given the fact that there is no money management attached to the system whatsoever, the results are quite nice. You should keep in mind that this trading system torso is just a preliminary raw test of how the Zone statistics works in the market. The only thing the system does is opening a position at the close of a bar and closing the position strictly at the close of the following bar. Nothing more.
To recap, the idea behind the Zone Strat trading system is quite simple: trade in the direction of the proabability. If the statistics shows a tendency to stay in the current Zone, open a position in this direction. If the statistics shows there is a tendency to abandon the Zone, then open the position the other way round.
I have only added a little twist to the code which ensures we are not going to open a position if there is too small a profit potential despite the fact that we would trade with the probabilities.
Example of such a case:
according to the number of consecutive bars in the Zone and the Zone percentage, there may be a high probability that the next bar will close in the same Zone as did the current bar. But if the current bar in fact only showed a 5 points range, then the minimum distance from the current close to the "stay in the Zone" price level may not be worth trading after the trading costs are subtracted or even the distance may be negative. The reverse applies to the minimum distance to abandon the current Zone (see the code).
So, incorporated into code there is a measure of what is the minimum palatable distance between the current close and the price level at which a Zone change/Zone abandonment occurs to open a trade.
The best way to come to grips with this test system is to apply it to your chart by yourself and fiddle with the settings.
Attached:
- equity curve
- trading system code
You are free to experiment with the code by yourself and post your results here!
Good luck!
Michal
If you can't see the attached image and ELD file, please log in. You must first register to be able to log in.