How to Train Your Dragon

The Dragon filter is a direct dirrivative of the Voss Predictive Filter. Dr. Voss describes this filter as “A filter for universal real-time prediction of band- limited signals” This algorithm was developed to provide greater resolution and insight into a wide class of signals generated by deterministic or stochastic systems. It attempts to remove group and phase delays from the Weighted Moving Average output. This is done by extracting the first harmonic of the output using a bandpass filter and then applying a "negative-delay" formula to it.

Forecasting financial time series is regarded as one of the most challenging applications of time series prediction due to their dynamic nature. However, it is the fundamental element of most investment activities thus attracting the attention of practitioners and researchers for many decades.

Forecasting Financial Time Series using Linear Predictive Filters; by Bin Li; A thesis submitted for the degree of Doctor of Philosophy in Electrical and Electronic Engineering, Imperial College London, Imperial College London


I called the indicator the Dragon because, not unlike the movie and book; “How to train your Dragon, you must adjust this indicator so that is coincides with all of your other indicators. This requires you to change the Dragon indicator settings for each and every new stock symbol or equity.


The ultimate advantage of the ALMA component of this filter is the ability it provides to traders, to adjust the data window allowing the primary entry/exit signals to be dialed in exactly at the optimal spot between the Early Warning Indicator dot and the Cross/Over dot.  Because this system was initially designed for swing traders who’s time frame is normally a few days to a few weeks, the look back periods are short.  ALMA greatly improves Swing•Genies ability to dial in these short time frame filters with as few whipsaw trade indications as possible.  These same attributes also enable the day traders to alter trading time frames to very effectively utilize this indicator. 

All stocks produce parabolic swings using price and time to calculate and draw the arc's that lead to the pinnacle of the peaks and the absolute bottom of there troughs. This is regardless of the time frame you choose to trade in.  When similar patterns recur at progressively smaller scales, they are identified as fractals.  This fractal property of price and time in the stock market is what allows Swing•Genie to excel at identifying optimal opportunities within the Trend, Swing or Day trading time frames.  Swing•Genie signals, projected on your main trading chart are generated in tandem within the Early Warning System and the Cycle Indicators.


      Slides can be advanced by clicking on the large slide, using thumbnails above or                the   "<" ">"   controls provided at the left and right edge of each slide.

Forward Projection of Current Price Trends

Any technician attempting to forecast what future prices will be, is engaging in technical analysis, hocus-pocus. Current methods being utilized to accomplish some level of success in forecasting price are detailed in the TASC article describing the Swing•Genie indicator.  It is, however, possible to simply project the current price action forward a few bars with the expectation that the current trend will continue on the approximate arc that is being fellowed by the equity. This provides to the trader a visual projection of the potential price action, but in no way is it any guarantee that prices will actually fellow this tract.  The fellowing slides provide a detailed explanation of how Swing•Genie price forecast is calculated so that the trader can understand exactly what credence and confidence to give to them.  Hurst Laguerre channels are also touched on in the following slides.  For a complete Hurts cycle explanation, scroll down this page to a dedicated slide.  


Laguerre Hurst Cycles.png

JM Hurst developed a trading system (Hurst’s Cyclic Theory) to clearly display the cyclic harmonics that exist within multiple time frames. Hurst theories attempt to deconstruct the fractal hierarchy of an equity and determine alignments and convergences.  He produced two books on this system in the 1970's and several books have been authored by individuals who were advocates of Hurst's published theories. Many trading message boards have threads on Hurst cycles with multiple descriptions of various methods to calculate, code and execute trades based on signals generated by these indicators.  In addition, many institutions and hedge funds incorporate some version of Hurst cycles into their trading protocols and algorithms. 

Three images taken from "cycles course," J.M. Hurst; Traders press inc. GREENVILLE SC copyright 1973

The primary principals of Hurst cycles are price moves in waves (See illustration above) that exhibit cyclic attributes based on their time scales. Although the concept may seem simple, the execution utilizing the lessons and incorporating the finer points from Hurst's second book, "Cycles Course," can be very difficult to correctly and profitably apply.  Strong signals are generated by Hurst cycles when cycles of different lengths have peaks or troughs that converge. 

There have been many different attempts at executing Hurst cycles as a Technical Indicator.  Most of those attempts have been centered around displacing moving averages backward to center or align the different lengths so that the troughs and peaks align with the highs and lows of the current stock cycles or swings.  This displacement is usually 1/2 of each of the moving average lengths utilized.  This produces a setback of the active signal causing it to end many bars prior to the most current bar.  

Because this ends the moving average indicators signals as many bars back as the displacement, no current or real time indication is provided.  With prior cycles highs and lows as a guide, combined with the current trend, an artificial signal can be produced, projecting the indicator forward. This determines the current trend based on the last X number of price points in the chart starting where the indicators setback ends. 

• Laguerre Hurst Cycles •

Swing•Genie Cycles incorporates Hurst cycles theories but utilizes only two nested Laguerre moving filters.  Laguerre moving filters have significantly less lag and the setback required is only three days/bars. These moving filters take as there inputs the highest high and the lowest lows for the two adjustable periods. The point of the indicator is to determine when a short term swing cycle harmonizes or aligns with a long term cycle,  i.e.,  determining when the tops and bottoms of these cycles align. The default intervals utilized in Swing•Genie Cycles are short period of 5 days and the long period of 20 days.  Each equity can have unique cycle lengths and the Swing•Genie Cycle Indicator allows you to change these parameters in the formatting/input pallets.   

Because the Swing•Genie Laguerre cycles utilize the highest highs and lowest lows of two time periods, the nested channels produce natural bounding boxes.  This dramatically highlights likely support and resistance levels as they often occur at prior highs or lows that this indicator is drawing.  

This provides the trader with a high probability gage of the profit that can be expected from your entry into a trade executed at these demarcation lines.

If the entry point is at the shorter interval, then the upper limit of this trade in likely to be the upper band of the short term channel.  If the entry is at the convergence of both the short term channel's lower band and the long term channel's lower band then your trade potential is the high band of the long term channel. This inflection point is obviously stronger and results in a much higher probability of success and more sustained and profitable entry positions.

This obviously produces another indication that your trade may be approaching an inflection point and turn against you as you approach the upper demarcation lines. 

Every time the high or low of the bar touches either bands, top or bottom, an intersection dot is produced.  These dots are blue/white if they are touching the inside 5 bars band and red/white/blue or green/white/blue if they are touching the upper or lower, 20-day bar, longer length time period channel.  

These double dots produce entry and exit signals with the greatest potential. The exception to this statement is when stocks are very strong trends and continually bounce off of the upper or lower limits of the long term channel.   


Closing Score Sentiment allows you to turn on the color coded background you see below.   Red will coincide with a negative trend and green indicates a positive trend. Each is indicating what the prevailing market/trader sentiment currently is.  You can read a complete explanation of Closing Score Sentiment here

Screen Shot 2017-05-25 at 2.30.51 PM.png

The control panel allow you to control the intensity of these sentiment background flags.  

The next page in this series is: