Of course, but there is only one logical way to trade. But that's just me.
RSI based trading strategies
- Thread starter frostengine
- Start date
Of course, but there is only one logical way to trade. But that's just me.
So far I have identified a few testable edges using RSI. Down the road I will also test the methods of Cardwell/Brown to determine if what they are selling has any edge as well.
RSI is a big subject and I do not expect to exhaust all methods within a few weeks. Will be a long process.
I have now completed testing the existing methods on the short side. The two profitable methods:
1. Hidden Divergence
2. Below/Above given threshold
Both displayed a statistical edge when used for shorting. Next I will be rounding out the first round of RSI strategies by combining the long/short divergence/threshold methods into a single method using the best parameter combinations to see if the sum of the parts also exhibits a statistical edge.
After this is completed,I will take a short break from RSI and move on to another concept. The goal is to cover several indicators/concepts before circling back around to put it all together and create strategies.
If you want to download the source used on the short side or see charts displaying gains you can view it on my blog: http://thestrategictrader.com/?p=336
When I come back around to RSI, I have several AI (neural network/pattern matching) concepts to test. However, I want to cover more indicators first as they will be useful in the next round of testing.
Does anyone have any suggestions on which indicator/concepts to look into next?
1. Hidden Divergence
2. Below/Above given threshold
Both displayed a statistical edge when used for shorting. Next I will be rounding out the first round of RSI strategies by combining the long/short divergence/threshold methods into a single method using the best parameter combinations to see if the sum of the parts also exhibits a statistical edge.
After this is completed,I will take a short break from RSI and move on to another concept. The goal is to cover several indicators/concepts before circling back around to put it all together and create strategies.
If you want to download the source used on the short side or see charts displaying gains you can view it on my blog: http://thestrategictrader.com/?p=336
When I come back around to RSI, I have several AI (neural network/pattern matching) concepts to test. However, I want to cover more indicators first as they will be useful in the next round of testing.
Does anyone have any suggestions on which indicator/concepts to look into next?
Hi,
Frostengine,
I see you are using NinjaTrader like me. I was wondering if I could e-mail you a similar indicator to the RSI but a little better, and if it is possible to make a couple coding changes.
1) Indicator will display a signal on my chart when for example it displays the hidden divergence and the below / above for both short and long trades.
2) The ability to turn on auto trade for example, if I want to NT to actually enter a trade for me automatically.
My goal will be to make my current trading strategy automated. The benefit to you is you would be able to use it to trade yourself. I would then include the other indicators that I use to fine tune the strategy so all will be working in concert. Or if someone else wants to do this, let me know.
Frostengine,
I see you are using NinjaTrader like me. I was wondering if I could e-mail you a similar indicator to the RSI but a little better, and if it is possible to make a couple coding changes.
1) Indicator will display a signal on my chart when for example it displays the hidden divergence and the below / above for both short and long trades.
2) The ability to turn on auto trade for example, if I want to NT to actually enter a trade for me automatically.
My goal will be to make my current trading strategy automated. The benefit to you is you would be able to use it to trade yourself. I would then include the other indicators that I use to fine tune the strategy so all will be working in concert. Or if someone else wants to do this, let me know.
Before moving on to another indicator, I wanted to take a hard look at the way I have been evaluating these strategies. The primary qualifier I was using is predicting market direction 3 and 5 bars into the future. I think this is vital to a solid strategy. However, it also produces some strategies that we know would be hard to trade. For example, buying when oversold. Although it has a high winning percentage if you move fast, it has poor risk/reward metrics.
I decided to establish another test where the strategies are set to exist based on pre-set profit targets and stops. Using a 1:2 risk/reward ratio. The goal is to qualify if a strategy is low risk or not.
The results of this was not that surprising. Buying on oversold RSI values has poor risk/reward. The most surprising aspect was that Hidden Divergence was unable to exceed the targets. However, the results came extremely close. In one instance missing the target by only a few dollars. I view this as further re-assurance that RSI based Hidden divergence contains an edge.
Moving forward I will be testing all strategies based on both tests to determine high percentage winners as well as good risk profile. In my mind to qualify as a tradeable edge a strategy does not have to exceed statistically valid targets in both tests. I think exceeding in one test and performing well in the other is good enough. These thresholds are not "magic" numbers. They are just numbers required to be "statistically significant". Failure to pass doesn't mean its a "bad" strategy. Therefore, passing in one is all that is needed to warrant further consideration down the line as something that "works" based on your needs. For instance, if you need a high winning percentage and very short holding period the buying on oversold RSI could still be useful to you, but if your building a well rounded strategy RSI hidden divergence is a better bet based on performing well in both.
For more info on the risk/reward testing procedure and thresholds used view this article on my blog: <a href="http://thestrategictrader.com/?p=363">Risk Reward Testing</a>
I decided to establish another test where the strategies are set to exist based on pre-set profit targets and stops. Using a 1:2 risk/reward ratio. The goal is to qualify if a strategy is low risk or not.
The results of this was not that surprising. Buying on oversold RSI values has poor risk/reward. The most surprising aspect was that Hidden Divergence was unable to exceed the targets. However, the results came extremely close. In one instance missing the target by only a few dollars. I view this as further re-assurance that RSI based Hidden divergence contains an edge.
Moving forward I will be testing all strategies based on both tests to determine high percentage winners as well as good risk profile. In my mind to qualify as a tradeable edge a strategy does not have to exceed statistically valid targets in both tests. I think exceeding in one test and performing well in the other is good enough. These thresholds are not "magic" numbers. They are just numbers required to be "statistically significant". Failure to pass doesn't mean its a "bad" strategy. Therefore, passing in one is all that is needed to warrant further consideration down the line as something that "works" based on your needs. For instance, if you need a high winning percentage and very short holding period the buying on oversold RSI could still be useful to you, but if your building a well rounded strategy RSI hidden divergence is a better bet based on performing well in both.
For more info on the risk/reward testing procedure and thresholds used view this article on my blog: <a href="http://thestrategictrader.com/?p=363">Risk Reward Testing</a>
Well, maybe they don't deserve it.
I don't believe that concepts in themselves contain secrets, thus I allow myself to talk about them.
However you'll never see me talk about implementing the concepts.
Ninna
I do understand the situation.But one thing that always gets me is how greedy we all have become.Most of us can make money,but can we share our wealth or the knowledge that created the wealth.You think the whole world is going to jump and start using your strategy and its implementation.The answer is no.You will just share that witha a few people and if they start sharing in your good luck,what is wrong with that.
Most of you will not understand the concept of sharing yet,but I hope one day you do,because when you do,you will find real happinesss Good luck to you all
I don't believe that concepts in themselves contain secrets, thus I allow myself to talk about them.
However you'll never see me talk about implementing the concepts.
Ninna
I do understand the situation.But one thing that always gets me is how greedy we all have become.Most of us can make money,but can we share our wealth or the knowledge that created the wealth.You think the whole world is going to jump and start using your strategy and its implementation.The answer is no.You will just share that witha a few people and if they start sharing in your good luck,what is wrong with that.
Most of you will not understand the concept of sharing yet,but I hope one day you do,because when you do,you will find real happinesss Good luck to you all
Hello RCG,
I was in the same position as you with the so called "Composite Index" and the books where it is written about it, but checked out the matter when I heard that Mr. Cardwell was the creator and owner of the code.
He does not allow her to continue and maybe he is about to sue her. I am not in position to say what Mr. Cardwell will finally deside to do, but I know there is something happening about the matter.
The indicator formula was copyrighted by Mr. Cardwell back in 1992. Take a look at this page: http://www.fxcodebase.com/code/viewtopic.php?f=17&t=3007&start=0&hilit=cfg
As told in the text the different platform versions at the bottom of the page has been posted with Mr. Cardwell's permission.
Since I have benefited from the indicator without previously having permission from the acctual owner, I think it is fair to pay something back by informing about the matter when I see threads or posts about the "Composite Index".
- Laurus