Not be confused with non-parametric statistics, i define a "non-parametric" indicator as an indicator who does not have any parameter input. Such indicators can be useful since they don't need to go through parameter optimization. I present here a non parametric based on exponential averaging using a modified ratio of open-close to high-low range indicator as smoothing variable.
The ratio of open-close to high-low range is a measurement involving calculating the ratio between the absolute close/open price difference and the range (high - low), now the relationship between high/low and open/close price has been studied in econometrics for some time but there are no reason that the ohlc range ratio may be an indicator of , however we can make the hypothesis that trending markets contain less indecision than ranging market and that indecision is measured by the high/low movements, this is an idea that i've heard various time.
Since the range is always greater than the absolute close/open difference we have a scaled smoothing variable in a range of 0/1, this allow to perform exponential averaging. The ratio of open-close to high-low range is calculated using the of the close/high/low/open price in order to increase the smoothing effect. The tend to smooth more with low time frames than higher ones, since the indicator use for the calculation of its smoothing variable, smoothing may differ depending on the time frame you are in.
1 minute tf
1 hour tf
Making non parametric indicators is quite efficient, but they wont necessarily outperform classical parametric indicators. I also presented a modified version of the ratio of open-close to high-low range who can provide a smoothing variable for exponential averaging. I hope the indicator can help you in any way.
Thanks for reading !
« Je prendrai l’essor et je m’envolerai vers la mer.
« Je connaîtrai le goût des brises du large. J’entendrai les grands cris de la tempête.