Trading threshold by John Ehlers
In radar, signal-to-noise ratio is used to measure the quality of target detection. In trading, this simple concept can be used to hit our profit targets better. We can improve our trading profitability if trading decisions are deferred until the signal-to-noise ratio is high. The charts can be viewed as noisy channels in which the daily range and very short-term day-to-day variations are the "noise." The longer-term variations mark the channel envelope.
The trader's profitability can be enhanced if he waits for conditions in which the peak-to-peak variation of the channel envelope exceeds four times the width of the "noise band." Noise as classically defined is energy (in our case, price activity) that carries no information. Complete randomness has no information and, therefore, random events are noisy. When noise is completely random, all frequencies are present.
This is called "white noise" because the picture of the uniform frequency distribution resembles snow. White noise has a Gaussian, or normal, amplitude probability distribution, which is the familiar bell-shaped curve that describes many statistical cases, such as a given population's IQ distribution. In the same way very few people have exceptionally high IQs, the amplitude of white noise can be very large but occur very seldom.