Strategy Gaussian Anomaly DerivativeConcept behind this Strategy :
Considering a normal "buy/sell" situation, an asset would be bought in average at the median price following a Gaussian like concept. A higher or lower average trend would significate that the current perceived value is respectively higher or lower than the current median price, which mean that the buyers are evaluating the price underpriced or overpriced.
This behaviour would be even more relevent depending on its derivative evolution.
Therefore, this Strategy setup is based on this Gaussian like concept anomaly of average close positionning compare to high-low average derivative, such as the derivative of the following ploted basic signal : 1-(high+low)/(2*close).
This Strategy can actually be used like a trend change and continuation strength indicator aswell.
In the Setup Signal part :
You can define the filtering of the basis signal "1-(high+low)/(2*close)" on EMA or SMA as you wish.
You can define the corresponding period and the threathold as a mutiply of the average 1/3 of all time value of the basis signal.
You can define the SMA filtering period of the Derivative signal and the corresponding threathold on the same mutiply of the average 1/3 of all time value of the derivative.
In the Setup Strategy part :
You can set up your strategy assesment based on Long and/or Short. You can also define the considered period.
The most successful tuned strategies I did were based on the derivative indicator with periods on the basis signal and the derivative under 30, can be 1 to 3 of te derivative and 7 to 21 for the basis signal. The threathold depends on the asset volatility aswell, 1 is usually the most efficient but 0 to 10 can be relevent depending on the situation I met. You can find an example of tuning for this strategy based on Kering's case hereafter.
I hoping that you will enjoy using this Strategy, don't hesitate to comment, to question, to correct or complete it ! I would be very curious about similar famous approaches that would have already been made.
Thank to you !
Gaussian
Gaussian MFI BFThe signals are based on the Gaussian Channel calculated by Donovan Wall. Thanks also to Kiasaki for Rate of Change code.
Simply going long and short based on Gaussian channel was not consistent enough so I also included an MFI filter. We only go long if Money Flow Index is greater than the last candle (more money is flowing in than out). The opposite for short. Also employed Rate of Change function (white background) to avoid sideways market.
A 2% stop loss seems to be the most balanced for this strategy but can can altered in the settngs.
Also included a take-profit function but it seems this is best not implemented for this strategy, at least in the case of XBT/USD.
A derivative of The MFI is also plotted on the chart simply to compare the current candle with the last candle visually.
Background color:
Green = Long
Red = Short
White = No Trade
