I fully agree with Dr. Van Tharps frequent statement - We dont trade the markets, we trade our beliefs about the markets. Keeping this in mind, I am jotting down my beliefs about market behavior. The motive is to develop a trading system which will assimilate and reflect my beliefs about price behavior. Following are my beliefs:
Belief 1) There are only 2 phases in any market be it equity, commodity or currency. The two phases are Trending phase and Range bound phase
Belief 2) Inspite of the fact that 70% of the time market remains Range Bound, more money can be made in trending phase. Also money can be made in both up and down trend. I do NOT believe in long-only phenomenon.
Belief 3) Identifying whether market is in trending phase or range bound phase is difficult. Finding breakout levels of the range bound market (boundaries of the range) is one of the ways to distinguish between the two phases.
Belief 4) Once prices come out of the range, they get into trending mode. Predicting the magnitude of the trend is futile. One should just keep riding the trend. However, I believe, more the time spent during the range bound phase, larger will be the trend after the range breakout.
Belief 5) Strategies for Trending Market (called as trend following strategies) and strategies for Range Bound Market (called as Mean Reversion Strategies) are almost opposite.
System building:
Based on the above beliefs, I want to make a trading system in Amibroker (AFL). Following are the steps:
1)Scan all the 207 future stock on NSE to find those which have just come out or are about to come out of the range bound phase. This is possible by identifying stocks which are running in horizontal price ranges. The ranges can be in any form like Head and shoulders, double top, double bottom, rectangle / channel, cup & handle etc. The name of the pattern is not important as long as its boundaries are flat or almost horizontal. (Hence triangles are excluded)
2)The resulting stocks should be ranked is such a way that stocks which have spent more time within the range should come first. This can be done by counting the number of times the boundaries of the range were tested. In other words, a stock in which the boundary was tested 4 times will come first than the stock in which boundary was tested thrice and so on. Other ranking parameters can also be added like high volumes, high price stocks, etc.
3)Not more than 8 stocks should be part of the portfolio out of which not more than 4 should be Longs and not more than 4 should be shorts. Following are the various possibilities:
a. If 9 long signals and 8 short signals are generated portfolio should be of 4 longs and 4 shorts with highest ranks.
b. If 5 long signals and 3 short signals are generated portfolio should be of 4 longs and 3 shorts. ( It cant be 5 longs and 3 shorts)
c. If 8 long signals and 0 short signals are generated portfolio should be of 4 longs and 0 shorts.(It cant be 8 longs)
4)None of the stocks forming part of the portfolio should belong to the same sector. This means we have one more parameter for filtering the long/short signals. To make this happen, we have to first allocate all the 207 futures stock to their respective sectors. (I doubt if its possible to make AFL recognize the sectors to which the stocks belong)
5)Buy few pips above the boundary /neckline/channel line with Stop Loss = Buy price ATR. Opposite for Short signal.
6)Keep trailing the SL as stock moves in favor by using Parabolic SAR.
7)If any of the long/short position gets closed on a day, it should be replaced by scanning the futures list of NSE maintaining the rule that not more than 4 longs and 4 shorts should be in the portfolio.
Kindly give your feedback to improvise the system. I fully understand that there is nothing new in the system and there must be several AFLs written on similar lines. I need guidance regarding AFL writing with internet resources as I come from non-IT background.
Belief 1) There are only 2 phases in any market be it equity, commodity or currency. The two phases are Trending phase and Range bound phase
Belief 2) Inspite of the fact that 70% of the time market remains Range Bound, more money can be made in trending phase. Also money can be made in both up and down trend. I do NOT believe in long-only phenomenon.
Belief 3) Identifying whether market is in trending phase or range bound phase is difficult. Finding breakout levels of the range bound market (boundaries of the range) is one of the ways to distinguish between the two phases.
Belief 4) Once prices come out of the range, they get into trending mode. Predicting the magnitude of the trend is futile. One should just keep riding the trend. However, I believe, more the time spent during the range bound phase, larger will be the trend after the range breakout.
Belief 5) Strategies for Trending Market (called as trend following strategies) and strategies for Range Bound Market (called as Mean Reversion Strategies) are almost opposite.
System building:
Based on the above beliefs, I want to make a trading system in Amibroker (AFL). Following are the steps:
1)Scan all the 207 future stock on NSE to find those which have just come out or are about to come out of the range bound phase. This is possible by identifying stocks which are running in horizontal price ranges. The ranges can be in any form like Head and shoulders, double top, double bottom, rectangle / channel, cup & handle etc. The name of the pattern is not important as long as its boundaries are flat or almost horizontal. (Hence triangles are excluded)
2)The resulting stocks should be ranked is such a way that stocks which have spent more time within the range should come first. This can be done by counting the number of times the boundaries of the range were tested. In other words, a stock in which the boundary was tested 4 times will come first than the stock in which boundary was tested thrice and so on. Other ranking parameters can also be added like high volumes, high price stocks, etc.
3)Not more than 8 stocks should be part of the portfolio out of which not more than 4 should be Longs and not more than 4 should be shorts. Following are the various possibilities:
a. If 9 long signals and 8 short signals are generated portfolio should be of 4 longs and 4 shorts with highest ranks.
b. If 5 long signals and 3 short signals are generated portfolio should be of 4 longs and 3 shorts. ( It cant be 5 longs and 3 shorts)
c. If 8 long signals and 0 short signals are generated portfolio should be of 4 longs and 0 shorts.(It cant be 8 longs)
4)None of the stocks forming part of the portfolio should belong to the same sector. This means we have one more parameter for filtering the long/short signals. To make this happen, we have to first allocate all the 207 futures stock to their respective sectors. (I doubt if its possible to make AFL recognize the sectors to which the stocks belong)
5)Buy few pips above the boundary /neckline/channel line with Stop Loss = Buy price ATR. Opposite for Short signal.
6)Keep trailing the SL as stock moves in favor by using Parabolic SAR.
7)If any of the long/short position gets closed on a day, it should be replaced by scanning the futures list of NSE maintaining the rule that not more than 4 longs and 4 shorts should be in the portfolio.
Kindly give your feedback to improvise the system. I fully understand that there is nothing new in the system and there must be several AFLs written on similar lines. I need guidance regarding AFL writing with internet resources as I come from non-IT background.
