An equity curve picture would be nice
Example Trading System
- Thread starter sidinuk
- Start date
I'm feeling charitable :
This system is probably doomed to break even over time or worse, because it has no edge, and neither has an attempt been made to find one as far as I can see. It relies on the market exhibiting the same character as it has done in the past, which will work some of the time but relies on luck.
To find an edge you will need to get back to basics. For a breakout system like this, ignore all the curve-fitting that has been done about exact size of the range, not trading on a particular day or whatever, and look at the results for trading every breakout of the opening range and closing at EOD.
Then try this : categorize each day on which a trade was made according to it's volatility (some measure of ATR, though very basic, would probably work). As long as your measure of volatility is valid you will see the system working on "high" volatility days and vice versa.
That's all obvious, right? But not, I'll bet, something to which a great deal of thought had been given during development of this system.
So, all curve-fitting aside, if your system only trades break-outs during periods (days, in this case) of high volatility then it will be successful without any "optimization" of entries and exits.
It would be possible to hypothesise, therefore, that one way to demonstrate an edge for a break-out system would be to only allow it to trade during periods that have an increased probability of high volatility - perhaps according to a dependency that you have identified.
"The Mathematics of Technical Analysis" anyone? Anyone...?
This system is probably doomed to break even over time or worse, because it has no edge, and neither has an attempt been made to find one as far as I can see. It relies on the market exhibiting the same character as it has done in the past, which will work some of the time but relies on luck.
To find an edge you will need to get back to basics. For a breakout system like this, ignore all the curve-fitting that has been done about exact size of the range, not trading on a particular day or whatever, and look at the results for trading every breakout of the opening range and closing at EOD.
Then try this : categorize each day on which a trade was made according to it's volatility (some measure of ATR, though very basic, would probably work). As long as your measure of volatility is valid you will see the system working on "high" volatility days and vice versa.
That's all obvious, right? But not, I'll bet, something to which a great deal of thought had been given during development of this system.
So, all curve-fitting aside, if your system only trades break-outs during periods (days, in this case) of high volatility then it will be successful without any "optimization" of entries and exits.
It would be possible to hypothesise, therefore, that one way to demonstrate an edge for a break-out system would be to only allow it to trade during periods that have an increased probability of high volatility - perhaps according to a dependency that you have identified.
"The Mathematics of Technical Analysis" anyone? Anyone...?
acrary?
Libertine,
Thank you for your pointer. Can you elaborate a little more, perhaps with a couple of numbers?
One of the premises of these range breakout systems is that it depends on unidirectional expansion of volatility. However, the problem is to predict when there will be such an expansion. Crabel suggested using ID-NR4, NR7. hook days etc as filters. Do you have any suggestions or ideas that we can try?
Thanks,
Maji
Thank you for your pointer. Can you elaborate a little more, perhaps with a couple of numbers?
One of the premises of these range breakout systems is that it depends on unidirectional expansion of volatility. However, the problem is to predict when there will be such an expansion. Crabel suggested using ID-NR4, NR7. hook days etc as filters. Do you have any suggestions or ideas that we can try?
Thanks,
Maji
Do these things you mention (I don't know what they mean exactly, though I have heard them mentioned before) lead to high volatility significantly more often than not? If so, they may provide an edge for a break-out system.
In order to find out, you will need a measure of volatility that you trust and a measure of dependence that you trust.
It wasn't a question, it was a suggestion.
Good luck.
Thanks for the great step-by-step process
I agree with your thoughts as all systems have stretches where they are flat or down and it can last months. I like to diversify what I trade and the types of systems I use. This one has been doing fairly well this year for me.
FYI-I trade a modified version of this system on ER2 and it has done fairly well this year:
PL= 8% (as pct of the index value)
W/L ~ 60%
I've posted my modifications earlier in this thread.
DS
Sid,My condolences to you and all inUK. Thanks for your great contribution to these boards. I have been trading YM and ER2 with your system and am well pleased with the results. Patrick