Dominant Cycle Tuned Rsi BackgroundBackground version of the Dominant Cycle Tuned Rsi Background published here
Dominantcycle
Dominant Cycle Tuned RsiIntroduction
Adaptive technical indicators are importants in a non stationary market, the ability to adapt to a situation can boost the efficiency of your strategy. A lot of methods have been proposed to make technical indicators "smarters" , from the use of variable smoothing constant for exponential smoothing to artificial intelligence.
The dominant cycle tuned rsi depend on the dominant cycle period of the market, such method allow the rsi to return accurate peaks and valleys levels. This indicator is an estimation of the cycle finder tuned rsi proposed by Lars von Thienen published in Decoding the Hidden Market Rhythm/Fine-tuning technical indicators using the dominant market vibration/2010 using the cycle measurement method described by John F.Ehlers in Cybernetic Analysis for Stocks and Futures .
The following section is for information purpose only, it can be technical so you can skip directly to the The Indicator section.
Frequency Estimation and Maximum Entropy Spectral Analysis
“Looks like rain,” said Tom precipitously.
Tom would have been a great weather forecaster, but market patterns are more complex than weather ones. The ability to measure dominant cycles in a complex signal is hard, also a method able to estimate it really fast add even more challenge to the task. First lets talk about the term dominant cycle , signals can be decomposed in a sum of various sine waves of different frequencies and amplitudes, the dominant cycle is considered to be the frequency of the sine wave with the highest amplitude. In general the highest frequencies are those who form the trend (often called fundamentals) , so detrending is used to eliminate those frequencies in order to keep only mid/mid - highs ones.
A lot of methods have been introduced but not that many target market price, Lars von Thienen proposed a method relying on the following processing chain :
Lars von Thienen Method = Input -> Filtering and Detrending -> Discrete Fourier Transform of the result -> Selection using Bartels statistical test -> Output
Thienen said that his method is better than the one proposed by Elhers. The method from Elhers called MESA was originally developed to interpret seismographic information. This method in short involve the estimation of the phase using low amount of information which divided by 360 return the frequency. At first sight there are no relations with the Maximum entropy spectral estimation proposed by Burg J.P. (1967). Maximum Entropy Spectral Analysis. Proceedings of 37th Meeting, Society of Exploration Geophysics, Oklahoma City.
You may also notice that these methods are plotted in the time domain where more classic method such as : power spectrum, spectrogram or FFT are not. The method from Elhers is the one used to tune our rsi.
The Indicator
Our indicator use the dominant cycle frequency to calculate the period of the rsi thus producing an adaptive rsi . When our adaptive rsi cross under 70, price might start a downtrend, else when our adaptive rsi crossover 30, price might start an uptrend. The alpha parameter is a parameter set to be always lower than 1 and greater than 0. Lower values of alpha minimize the number of detected peaks/valleys while higher ones increase the number of those. 0.07 for alpha seems like a great parameter but it can sometimes need to be changed.
The adaptive indicator can also detect small top/bottoms of small periods
Of course the indicator is subject to failures
At the end it is totally dependent of the dominant cycle estimation, which is still a rough method subject to uncertainty.
Conclusion
Tuning your indicator is a great way to make it adapt to the market, but its also a complex way to do so and i'm not that convinced about the complexity/result ratio. The version using chart background will be published separately.
Feel free to tune your indicators with the estimator from elhers and see if it provide a great enhancement :)
Thanks for reading !
References
for the calculation of the dominant cycle estimator originally from www.davenewberg.com
Decoding the Hidden Market Rhythm (2010) Lars von Thienen
Ehlers , J. F. 2004 . Cybernetic Analysis for Stocks and Futures: Cutting-Edge DSP Technology to Improve Your Trading . Wiley
Cosine, In-Phase & Quadrature IFM [Ehlers]Yet another method for determining the cycle of a market: this time, you have access to the two fastest and most accurate methods
as well as the option to average these methods together.
The controls are pretty straight forward:
Source lets you select the price data to perform calculations on (close, open, etc..)
Max Period is simply the cap for the algorithm when it's checking the validity of Periods.
-> If you notice your plots have a flat top, then increase this value to accept a wider range of Periods.
-> This setting has a min. value of 8 to reduce noise and a max of 100 to ignore waves from higher time frames.
Average? simply averages the two methods of calculation.
-> You may want to do this if you notice the two plots diverging a lot.
-> Cosine IFM tends to favor shorter periods; I-Q IFM tends to favor longer.
Cheers,
- DasanC
Adaptive Bandpass Filter [Ehlers]This is my latest bandpass filter - used to determine if a security is in a trend or cycle.
Now with an adaptive period setting! I use Ehlers in-phase & quadrature dominant cycle measurement (IQ IFM) method to set the period dynamically.
This method favors longer periods which tend to produce smoother, albeit laggier bandpass oscillator plots. From my quick tests, I tend to have lag between 4 and 8 bars, depending on the Timeframe.
The lower timeframes tend to have more noise and thus produce more interfering frequencies that may cause lag.
>Settings
Source: Select the data source to perform calc's on (close, open, etc...)
Period: Select the period to tune. Periods outside of this value will be attenuated (reduced)
Adaptive: Enable to have the I-Q IFM set the period for you (disables Period setting)
Bandpass Tolerance: Allow periods that are plus/minus the chosen period to pass.
Cycle Tolerance: Sensitivity of cycle mode. Lower values consider trends more frequent, higher values consider cycles more frequent.
Bandpass tolerance example: for instance, if this setting is 0.1 (10%) and Period is set to 20, then waves with a period of 18 - 22 will pass.
>How to read
Red line is the bandpass output, showing a lagged version of the dominant cycle representing the
Black lines are the upper and lower bounds for a cycle
Green Background indicates an uptrend
Red background indicates a downtrend
Adaptive Zero Lag EMA Strategy [Ehlers + Ric]Behold! A strategy that makes use of Ehlers research into the field of signal processing and wins so consistently, on multiple time frames AND on multiple currency pairs.
The Adaptive Zero Lag EMA (AZLEMA) is based on an informative report by Ehlers and Ric .
I've modified it by using Cosine IFM, a method by Ehlers on determining the dominant cycle period without using fast-Fourier transforms
Instead, we use some basic differential equations that are simplified to approximate the cycle period over a 100 bar sample size.
The settings for this strategy allow you to scalp or swing trade! High versatility!
Since this strategy is frequency based, you can run it on any timeframe (M1 is untested) and even have the option of using adaptive settings for a best-fit.
>Settings
Source : Choose the value for calculations (close, open, high + low / 2, etc...)
Period : Choose the dominant cycle for the ZLEMA (typically under 100)
Adaptive? : Allow the strategy to continuously update the Period for you (disables Period setting)
Gain Limit : Higher = faster response. Lower = smoother response. See for more information.
Threshold : Provides a bit more control over entering a trade. Lower = less selective. Higher = More selective. (range from 0 to 1)
SL Points : Stop Poss level in points (10 points = 1 pip)
TP Points : Take Profit level in points
Risk : Percent of current balance to risk on each trade (0.01 = 1%)
www.mesasoftware.com
www.jamesgoulding.com(Measuring%20Cycles).doc
Ehlers Smoothed Adaptive MomentumEhlers Smoothed Adaptive Momentum script.
This indicator was developed and described by John F. Ehlers in his book "Cybernetic Analysis for Stocks and Futures" (2004, Chapter 12: Adapting to the Trend).
