It has always been of great concern to me as to why mechanical trading seems so difficult to consistently make money whereas discretionary trading with a solid set of general rules has always given me better results. Perhaps our intuition has been always telling us something that we have been unable to quantify. My search has been for that variable and as of now I will begin testing my hypothesis.
To begin, I will take a trend riding system such as Vegas 1hr tunnel and apply my hypothesis to it to see if it is possibile to increase performance efficiency.
The greatest drawback to this as any other trend riding system is the inefficiencies caused by whipsaws. Generally, to avoid these, we try to create many filters all in an attempt to uncover a holy grail of sorts. I feel the holy grail is and has always been right under our noses. If this proves positive I do not know what kind of impact it will have, but I do not care. I want to see everybody do well, myself included and if I even contribute one inkling of what I have learned from all of your great threads and shared knowledge, It will make me feel good that I have been able to contribute to all of your great efforts. So please, if any of you have any advice or suggestions, good feedback or inappropriate jokes, it is all welcome. no holds barred, lets just make sure we stick to the point.
My idea of the holy grail begins with something we have all read time and time again. 95% of all traders lose. Some make their money out of pure skill, while others reap the spreads. Either way, I believe that from this statistic alone, we can plant the seed of the tree that will grow each and every one of our mechanical systems into money making machines.
From this statistic I began to look at any arbitrary system that I had programed into my tradestation and what I had noticed was that most of them, excluding commission, were just around break even. Once commission(spread) was added to performance, it seemed as though these systems were a complete waste. I wondered why was it that most basically broke even and how can this information help me. Well, if 95% lose, than it makes sense. Most that follow the system will lose. However, most that fade the system must also lose. If any one side of the equity curve were to increase without stopping(excluding the spread), then threre would exist a group of traders that would be constantly winning. We know this cannot happen.
Intuitively, we look to characterize markets that are trending versus ranging with all of these indicators, but what I propose is quite different and can be quantified. My idea is to measure our system performance, excluding spread in terms of an oscillator. We know that if our system is making too much money, it is destined to soon fail. If it is losing too much money(excluding spread) it is destined to start performing better.
I will create bollinger bands for system performance to measure when our system is x standard deviations away from a mean equity line and these levels will better pinpoint when we should be trading the system and when we should be fading it. To say that one system will prove this is optimistic, however, once we create a portfolio of trend riding systems and one of ranging systems, it should give us a better idea as to when to ride trends or range trade.
Other factors must be taken into consideration and that is drawdown and average profit per trade. We do not want to work on such a small time frame whereas filtering out alot of bad trades will still yield us with small gains after the spread is taken into consideration. Ideally, the system will have large fluctuations around a mean equity that will allow us to ride the system for big gains in both directions. Also, Intuitively i feel that systems that are always in the market will be better for this experiment, although after testing we may try to find ways to lower drawdown although I do not know if theoretically this is possible.
I will first work with Vegas 1hr tunnel. I do not know if this is the ideal system to test, but it is a start. If any of you have an always in the market system that yields big gains as well as big losses, and you feel as though it fits my qualifications for a system, please let me know and I will include it in the test. The overall goal will be to combine a myriad of systems into an indicator that will let us know when is a good time to go with the trend and when is a good time to fade it. Remember most systems will break even excluding the spread, too big of a gain and too many people make money, too big of a loss and too many people make money. We find optimal times when to trade and when to fade these systems and we make too much money.
I believe that deep down, great discretionary traders, although they may think it is something else( gut feeling or whatever they want to call it), all have a natural perception of this idea and act on it intuitively. Let's us mechanical traders figure this out so that we may level the playing field. Then what happens next I do not know.
To begin, I will take a trend riding system such as Vegas 1hr tunnel and apply my hypothesis to it to see if it is possibile to increase performance efficiency.
The greatest drawback to this as any other trend riding system is the inefficiencies caused by whipsaws. Generally, to avoid these, we try to create many filters all in an attempt to uncover a holy grail of sorts. I feel the holy grail is and has always been right under our noses. If this proves positive I do not know what kind of impact it will have, but I do not care. I want to see everybody do well, myself included and if I even contribute one inkling of what I have learned from all of your great threads and shared knowledge, It will make me feel good that I have been able to contribute to all of your great efforts. So please, if any of you have any advice or suggestions, good feedback or inappropriate jokes, it is all welcome. no holds barred, lets just make sure we stick to the point.
My idea of the holy grail begins with something we have all read time and time again. 95% of all traders lose. Some make their money out of pure skill, while others reap the spreads. Either way, I believe that from this statistic alone, we can plant the seed of the tree that will grow each and every one of our mechanical systems into money making machines.
From this statistic I began to look at any arbitrary system that I had programed into my tradestation and what I had noticed was that most of them, excluding commission, were just around break even. Once commission(spread) was added to performance, it seemed as though these systems were a complete waste. I wondered why was it that most basically broke even and how can this information help me. Well, if 95% lose, than it makes sense. Most that follow the system will lose. However, most that fade the system must also lose. If any one side of the equity curve were to increase without stopping(excluding the spread), then threre would exist a group of traders that would be constantly winning. We know this cannot happen.
Intuitively, we look to characterize markets that are trending versus ranging with all of these indicators, but what I propose is quite different and can be quantified. My idea is to measure our system performance, excluding spread in terms of an oscillator. We know that if our system is making too much money, it is destined to soon fail. If it is losing too much money(excluding spread) it is destined to start performing better.
I will create bollinger bands for system performance to measure when our system is x standard deviations away from a mean equity line and these levels will better pinpoint when we should be trading the system and when we should be fading it. To say that one system will prove this is optimistic, however, once we create a portfolio of trend riding systems and one of ranging systems, it should give us a better idea as to when to ride trends or range trade.
Other factors must be taken into consideration and that is drawdown and average profit per trade. We do not want to work on such a small time frame whereas filtering out alot of bad trades will still yield us with small gains after the spread is taken into consideration. Ideally, the system will have large fluctuations around a mean equity that will allow us to ride the system for big gains in both directions. Also, Intuitively i feel that systems that are always in the market will be better for this experiment, although after testing we may try to find ways to lower drawdown although I do not know if theoretically this is possible.
I will first work with Vegas 1hr tunnel. I do not know if this is the ideal system to test, but it is a start. If any of you have an always in the market system that yields big gains as well as big losses, and you feel as though it fits my qualifications for a system, please let me know and I will include it in the test. The overall goal will be to combine a myriad of systems into an indicator that will let us know when is a good time to go with the trend and when is a good time to fade it. Remember most systems will break even excluding the spread, too big of a gain and too many people make money, too big of a loss and too many people make money. We find optimal times when to trade and when to fade these systems and we make too much money.
I believe that deep down, great discretionary traders, although they may think it is something else( gut feeling or whatever they want to call it), all have a natural perception of this idea and act on it intuitively. Let's us mechanical traders figure this out so that we may level the playing field. Then what happens next I do not know.
Man who scratches ass should not bite fingernails