The following indicator is a normalized oscillator making use of the arc tangent sigmoid function (ArcTan), this allows to "squarify" the output result, thus visually filtering out certain variations originally in the oscillator. The magnitude of this effect can be controlled by the user. The indicator contains a gradient that shows the possibility of a reversal, with red colors indicating that a reversal might occur.
Settings
Length : Period of the oscillator
Pre-Gain : Changes the amplitude of the oscillator before passing through the ArcTan function, this allows to amplify/reduce the "squarification" effect introduced by this function. In order to make it easier for the user, the setting is in a (-10,10) range, with negative values reducing the amplitude and positive one increasing it.
Src : Source input of the indicator
Usage
The oscillator can be used to determine the direction of the trend by looking at its sign, if the oscillator is positive, market is up-trending, else down-trending, based on this usage the user might not be interested to look at every variations produced by the oscillator, this is where the hyperbolic tangent function and pre-gain setting can be useful, by using an high value of pre-gain the user will be able to only focus on the sign of the oscillator.
Here pre-gain is set to 5, we can see that the oscillator is now easier to visualize. However, the use of sigmoid functions remove useful information for a trader that needs to find divergences, this is where using a negative value of the pre-gain setting will result useful.
Here pre-gain is set to -5.
The indicator makes use of a gradient to show potential reversals, this gradient is determined by the correlation between the oscillator and the price (this is a way to measure potential divergences). If the color is closer to red it means that a potential reversal might occur, it is possible to say in which direction price might go by looking at the sign of the oscillator, so if the gradient is red and the oscillator is negative price might rise. The gradient is not affected by the pre-gain setting.
Haven't seen many useful scripts using trig functions -- this is great. Curious if you would recommend any material or media for learning more about the practical uses of sigmoid/trig functions... amazed at this idea of using statistic and trig together... taking the inverse tan of z-score, applying certain pre-gain, dividing by std.dev ... then correlating the original source with this value... ?! Please tell me more, point me to a book lol